diff --git a/docs/ops/ttnn/ttnn.add.md b/docs/ops/ttnn/ttnn.add.md
new file mode 100644
index 00000000..3a1dcc20
--- /dev/null
+++ b/docs/ops/ttnn/ttnn.add.md
@@ -0,0 +1,66 @@
+# ttnn.add
+
+| Name | Input Shapes | Input Layouts | Attributes | Output Shapes | Output Layouts |
+|------|--------------|---------------|------------|---------------|----------------|
+| ttnn.add | tensor<[1,9,3072,bf16]>
tensor<[1,9,3072,bf16]>
tensor<[1,9,3072,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 96, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 96, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 96, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,9,3072,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 96, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,1024,640,bf16]>
tensor<[1,1024,640,bf16]>
tensor<[1,1024,640,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1024 + d1', 'd2'), memory_config: (32, 20, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1024 + d1', 'd2'), memory_config: (32, 20, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1024 + d1', 'd2'), memory_config: (32, 20, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,1024,640,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1024 + d1', 'd2'), memory_config: (32, 20, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,1200,320,bf16]>
tensor<[1,1200,320,bf16]>
tensor<[1,1200,320,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1200 + d1', 'd2'), memory_config: (38, 10, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1200 + d1', 'd2'), memory_config: (38, 10, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1200 + d1', 'd2'), memory_config: (38, 10, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,1200,320,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1200 + d1', 'd2'), memory_config: (38, 10, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,64,256,256,bf16]>
tensor<[1,64,256,256,bf16]>
tensor<[1,64,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,64,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,12,27,27,f32]>
tensor<[1,12,27,27,f32]>
tensor<[1,12,27,27,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 324 + d1 * 27 + d2', 'd3'), memory_config: (11, 1, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 324 + d1 * 27 + d2', 'd3'), memory_config: (11, 1, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 324 + d1 * 27 + d2', 'd3'), memory_config: (11, 1, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[1,12,27,27,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 324 + d1 * 27 + d2', 'd3'), memory_config: (11, 1, 'tile<32x32, f32>', 'dram') |
+| ttnn.add | tensor<[1,300,512,bf16]>
tensor<[1,300,512,bf16]>
tensor<[1,300,512,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 300 + d1', 'd2'), memory_config: (10, 16, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 300 + d1', 'd2'), memory_config: (10, 16, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 300 + d1', 'd2'), memory_config: (10, 16, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,300,512,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 300 + d1', 'd2'), memory_config: (10, 16, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,9,2048,bf16]>
tensor<[1,9,2048,bf16]>
tensor<[1,9,2048,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 64, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 64, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 64, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,9,2048,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 64, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,10,768,bf16]>
tensor<[1,10,768,bf16]>
tensor<[1,10,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 10 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 10 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 10 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,10,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 10 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,64,240,320,f32]>
tensor<[1,64,240,320,f32]>
tensor<[1,64,240,320,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 15360 + d1 * 240 + d2', 'd3'), memory_config: (480, 10, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 15360 + d1 * 240 + d2', 'd3'), memory_config: (480, 10, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 15360 + d1 * 240 + d2', 'd3'), memory_config: (480, 10, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[1,64,240,320,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 15360 + d1 * 240 + d2', 'd3'), memory_config: (480, 10, 'tile<32x32, f32>', 'dram') |
+| ttnn.add | tensor<[1,64,30,40,f32]>
tensor<[1,64,30,40,f32]>
tensor<[1,64,30,40,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1920 + d1 * 30 + d2', 'd3'), memory_config: (60, 2, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1920 + d1 * 30 + d2', 'd3'), memory_config: (60, 2, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1920 + d1 * 30 + d2', 'd3'), memory_config: (60, 2, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[1,64,30,40,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1920 + d1 * 30 + d2', 'd3'), memory_config: (60, 2, 'tile<32x32, f32>', 'dram') |
+| ttnn.add | tensor<[1,9,768,bf16]>
tensor<[1,9,768,bf16]>
tensor<[1,9,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,9,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,12,3072,bf16]>
tensor<[1,12,3072,bf16]>
tensor<[1,12,3072,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 96, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 96, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 96, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,12,3072,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 96, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,2048,768,bf16]>
tensor<[1,2048,768,bf16]>
tensor<[1,2048,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 2048 + d1', 'd2'), memory_config: (64, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 2048 + d1', 'd2'), memory_config: (64, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 2048 + d1', 'd2'), memory_config: (64, 24, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,2048,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 2048 + d1', 'd2'), memory_config: (64, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,1024,16,16,bf16]>
tensor<[1,1024,16,16,bf16]>
tensor<[1,1024,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 16 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 16 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 16 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,1024,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 16 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,320,64,64,bf16]>
tensor<[1,320,64,64,bf16]>
tensor<[1,320,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 64 + d2', 'd3'), memory_config: (640, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 64 + d2', 'd3'), memory_config: (640, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 64 + d2', 'd3'), memory_config: (640, 2, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,320,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 64 + d2', 'd3'), memory_config: (640, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,1280,8,8,bf16]>
tensor<[1,1280,8,8,bf16]>
tensor<[1,1280,8,8,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 10240 + d1 * 8 + d2', 'd3'), memory_config: (320, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 10240 + d1 * 8 + d2', 'd3'), memory_config: (320, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 10240 + d1 * 8 + d2', 'd3'), memory_config: (320, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,1280,8,8,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 10240 + d1 * 8 + d2', 'd3'), memory_config: (320, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,256,1280,bf16]>
tensor<[1,256,1280,bf16]>
tensor<[1,256,1280,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 256 + d1', 'd2'), memory_config: (8, 40, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 256 + d1', 'd2'), memory_config: (8, 40, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 256 + d1', 'd2'), memory_config: (8, 40, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,256,1280,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 256 + d1', 'd2'), memory_config: (8, 40, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,64,60,80,f32]>
tensor<[1,64,60,80,f32]>
tensor<[1,64,60,80,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3840 + d1 * 60 + d2', 'd3'), memory_config: (120, 3, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3840 + d1 * 60 + d2', 'd3'), memory_config: (120, 3, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3840 + d1 * 60 + d2', 'd3'), memory_config: (120, 3, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[1,64,60,80,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3840 + d1 * 60 + d2', 'd3'), memory_config: (120, 3, 'tile<32x32, f32>', 'dram') |
+| ttnn.add | tensor<[1,16384,32,bf16]>
tensor<[1,16384,32,bf16]>
tensor<[1,16384,32,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 16384 + d1', 'd2'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 16384 + d1', 'd2'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 16384 + d1', 'd2'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,16384,32,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 16384 + d1', 'd2'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,1024,160,bf16]>
tensor<[1,1024,160,bf16]>
tensor<[1,1024,160,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1024 + d1', 'd2'), memory_config: (32, 5, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1024 + d1', 'd2'), memory_config: (32, 5, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1024 + d1', 'd2'), memory_config: (32, 5, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,1024,160,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1024 + d1', 'd2'), memory_config: (32, 5, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,512,32,32,bf16]>
tensor<[1,512,32,32,bf16]>
tensor<[1,512,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 32 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 32 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 32 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,512,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 32 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,9,128,bf16]>
tensor<[1,9,128,bf16]>
tensor<[1,9,128,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,9,128,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,10,i32]>
tensor<[1,10,i32]>
tensor<[1,10,i32]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 1, 'tile<32x32, u32>', 'dram')
mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 1, 'tile<32x32, u32>', 'dram')
mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 1, 'tile<32x32, u32>', 'dram') | operandSegmentSizes: array | tensor<[1,10,i32]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 1, 'tile<32x32, u32>', 'dram') |
+| ttnn.add | tensor<[1,64,56,56,bf16]>
tensor<[1,64,56,56,bf16]>
tensor<[1,64,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 56 + d2', 'd3'), memory_config: (112, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 56 + d2', 'd3'), memory_config: (112, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 56 + d2', 'd3'), memory_config: (112, 2, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,64,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 56 + d2', 'd3'), memory_config: (112, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,24,56,56,bf16]>
tensor<[1,24,56,56,bf16]>
tensor<[1,24,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 56 + d2', 'd3'), memory_config: (42, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 56 + d2', 'd3'), memory_config: (42, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 56 + d2', 'd3'), memory_config: (42, 2, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,24,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 56 + d2', 'd3'), memory_config: (42, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,201,768,bf16]>
tensor<[1,201,768,bf16]>
tensor<[1,201,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 201 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 201 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 201 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,201,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 201 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,160,7,7,bf16]>
tensor<[1,160,7,7,bf16]>
tensor<[1,160,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1120 + d1 * 7 + d2', 'd3'), memory_config: (35, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1120 + d1 * 7 + d2', 'd3'), memory_config: (35, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1120 + d1 * 7 + d2', 'd3'), memory_config: (35, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,160,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1120 + d1 * 7 + d2', 'd3'), memory_config: (35, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,1445,192,bf16]>
tensor<[1,1445,192,bf16]>
tensor<[1,1445,192,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1445 + d1', 'd2'), memory_config: (46, 6, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1445 + d1', 'd2'), memory_config: (46, 6, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1445 + d1', 'd2'), memory_config: (46, 6, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,1445,192,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1445 + d1', 'd2'), memory_config: (46, 6, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,1280,16,16,bf16]>
tensor<[1,1280,16,16,bf16]>
tensor<[1,1280,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 16 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 16 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 16 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,1280,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 16 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,64,480,640,f32]>
tensor<[1,64,480,640,f32]>
tensor<[1,64,480,640,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 30720 + d1 * 480 + d2', 'd3'), memory_config: (960, 20, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 30720 + d1 * 480 + d2', 'd3'), memory_config: (960, 20, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 30720 + d1 * 480 + d2', 'd3'), memory_config: (960, 20, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[1,64,480,640,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 30720 + d1 * 480 + d2', 'd3'), memory_config: (960, 20, 'tile<32x32, f32>', 'dram') |
+| ttnn.add | tensor<[920,1,256,bf16]>
tensor<[920,1,256,bf16]>
tensor<[920,1,256,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (29, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (29, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (29, 8, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[920,1,256,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (29, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,256,64,64,bf16]>
tensor<[1,256,64,64,bf16]>
tensor<[1,256,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,256,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,32,32,128,bf16]>
tensor<[1,32,32,128,bf16]>
tensor<[1,32,32,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 * 32 + d2', 'd3'), memory_config: (32, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 * 32 + d2', 'd3'), memory_config: (32, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 * 32 + d2', 'd3'), memory_config: (32, 4, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,32,32,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 * 32 + d2', 'd3'), memory_config: (32, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,640,32,32,bf16]>
tensor<[1,640,32,32,bf16]>
tensor<[1,640,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 32 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 32 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 32 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,640,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 32 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,9,4096,bf16]>
tensor<[1,9,4096,bf16]>
tensor<[1,9,4096,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 128, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 128, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 128, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,9,4096,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 128, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,64,14,14,bf16]>
tensor<[1,64,14,14,bf16]>
tensor<[1,64,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 14 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 14 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 14 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,64,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 14 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,96,14,14,bf16]>
tensor<[1,96,14,14,bf16]>
tensor<[1,96,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 14 + d2', 'd3'), memory_config: (42, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 14 + d2', 'd3'), memory_config: (42, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 14 + d2', 'd3'), memory_config: (42, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,96,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 14 + d2', 'd3'), memory_config: (42, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,8,768,bf16]>
tensor<[1,8,768,bf16]>
tensor<[1,8,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 8 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 8 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 8 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,8,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 8 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,64,120,160,f32]>
tensor<[1,64,120,160,f32]>
tensor<[1,64,120,160,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 7680 + d1 * 120 + d2', 'd3'), memory_config: (240, 5, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 7680 + d1 * 120 + d2', 'd3'), memory_config: (240, 5, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 7680 + d1 * 120 + d2', 'd3'), memory_config: (240, 5, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[1,64,120,160,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 7680 + d1 * 120 + d2', 'd3'), memory_config: (240, 5, 'tile<32x32, f32>', 'dram') |
+| ttnn.add | tensor<[1,9,1024,bf16]>
tensor<[1,9,1024,bf16]>
tensor<[1,9,1024,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 32, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 32, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 32, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,9,1024,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 32, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,14,128,bf16]>
tensor<[1,14,128,bf16]>
tensor<[1,14,128,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 14 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 14 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 14 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,14,128,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 14 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,256,256,bf16]>
tensor<[1,256,256,bf16]>
tensor<[1,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 256 + d1', 'd2'), memory_config: (8, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 256 + d1', 'd2'), memory_config: (8, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 256 + d1', 'd2'), memory_config: (8, 8, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 256 + d1', 'd2'), memory_config: (8, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,768,8,bf16]>
tensor<[1,768,8,bf16]>
tensor<[1,768,8,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 768 + d1', 'd2'), memory_config: (24, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 768 + d1', 'd2'), memory_config: (24, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 768 + d1', 'd2'), memory_config: (24, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,768,8,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 768 + d1', 'd2'), memory_config: (24, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,9,8192,bf16]>
tensor<[1,9,8192,bf16]>
tensor<[1,9,8192,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 256, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 256, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 256, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,9,8192,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 256, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,19200,64,bf16]>
tensor<[1,19200,64,bf16]>
tensor<[1,19200,64,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 19200 + d1', 'd2'), memory_config: (600, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 19200 + d1', 'd2'), memory_config: (600, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 19200 + d1', 'd2'), memory_config: (600, 2, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,19200,64,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 19200 + d1', 'd2'), memory_config: (600, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,197,1024,bf16]>
tensor<[1,197,1024,bf16]>
tensor<[1,197,1024,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 197 + d1', 'd2'), memory_config: (7, 32, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 197 + d1', 'd2'), memory_config: (7, 32, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 197 + d1', 'd2'), memory_config: (7, 32, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,197,1024,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 197 + d1', 'd2'), memory_config: (7, 32, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,12,197,197,bf16]>
tensor<[1,12,197,197,bf16]>
tensor<[1,12,197,197,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 2364 + d1 * 197 + d2', 'd3'), memory_config: (74, 7, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 2364 + d1 * 197 + d2', 'd3'), memory_config: (74, 7, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 2364 + d1 * 197 + d2', 'd3'), memory_config: (74, 7, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,12,197,197,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 2364 + d1 * 197 + d2', 'd3'), memory_config: (74, 7, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,193,768,bf16]>
tensor<[1,193,768,bf16]>
tensor<[1,193,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 193 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 193 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 193 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,193,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 193 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,128,128,128,bf16]>
tensor<[1,128,128,128,bf16]>
tensor<[1,128,128,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (512, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (512, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (512, 4, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,128,128,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (512, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,256,14,14,bf16]>
tensor<[1,256,14,14,bf16]>
tensor<[1,256,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,256,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[100,1,256,bf16]>
tensor<[100,1,256,bf16]>
tensor<[100,1,256,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (4, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (4, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (4, 8, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[100,1,256,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (4, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,512,7,7,bf16]>
tensor<[1,512,7,7,bf16]>
tensor<[1,512,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,512,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,4096,320,bf16]>
tensor<[1,4096,320,bf16]>
tensor<[1,4096,320,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 4096 + d1', 'd2'), memory_config: (128, 10, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 4096 + d1', 'd2'), memory_config: (128, 10, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 4096 + d1', 'd2'), memory_config: (128, 10, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,4096,320,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 4096 + d1', 'd2'), memory_config: (128, 10, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,14,768,bf16]>
tensor<[1,14,768,bf16]>
tensor<[1,14,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 14 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 14 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 14 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,14,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 14 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,12,128,bf16]>
tensor<[1,12,128,bf16]>
tensor<[1,12,128,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,12,128,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,12,768,bf16]>
tensor<[1,12,768,bf16]>
tensor<[1,12,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,12,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,4800,128,bf16]>
tensor<[1,4800,128,bf16]>
tensor<[1,4800,128,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 4800 + d1', 'd2'), memory_config: (150, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 4800 + d1', 'd2'), memory_config: (150, 4, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 4800 + d1', 'd2'), memory_config: (150, 4, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,4800,128,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 4800 + d1', 'd2'), memory_config: (150, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,256,180,320,bf16]>
tensor<[1,256,180,320,bf16]>
tensor<[1,256,180,320,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 180 + d2', 'd3'), memory_config: (1440, 10, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 180 + d2', 'd3'), memory_config: (1440, 10, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 180 + d2', 'd3'), memory_config: (1440, 10, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,256,180,320,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 180 + d2', 'd3'), memory_config: (1440, 10, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,32,4096,bf16]>
tensor<[1,32,4096,bf16]>
tensor<[1,32,4096,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 32 + d1', 'd2'), memory_config: (1, 128, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 32 + d1', 'd2'), memory_config: (1, 128, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 32 + d1', 'd2'), memory_config: (1, 128, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,32,4096,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 32 + d1', 'd2'), memory_config: (1, 128, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,16,197,197,bf16]>
tensor<[1,16,197,197,bf16]>
tensor<[1,16,197,197,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3152 + d1 * 197 + d2', 'd3'), memory_config: (99, 7, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3152 + d1 * 197 + d2', 'd3'), memory_config: (99, 7, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3152 + d1 * 197 + d2', 'd3'), memory_config: (99, 7, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,16,197,197,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3152 + d1 * 197 + d2', 'd3'), memory_config: (99, 7, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,32,28,28,bf16]>
tensor<[1,32,28,28,bf16]>
tensor<[1,32,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 28 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 28 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 28 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,32,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 28 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.add | tensor<[1,19,1024,bf16]>
tensor<[1,19,1024,bf16]>
tensor<[1,19,1024,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 19 + d1', 'd2'), memory_config: (1, 32, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 19 + d1', 'd2'), memory_config: (1, 32, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 19 + d1', 'd2'), memory_config: (1, 32, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,19,1024,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 19 + d1', 'd2'), memory_config: (1, 32, 'tile<32x32, bf16>', 'dram') |
diff --git a/docs/ops/ttnn/ttnn.concat.md b/docs/ops/ttnn/ttnn.concat.md
new file mode 100644
index 00000000..124e6e69
--- /dev/null
+++ b/docs/ops/ttnn/ttnn.concat.md
@@ -0,0 +1,22 @@
+# ttnn.concat
+
+| Name | Input Shapes | Input Layouts | Attributes | Output Shapes | Output Layouts |
+|------|--------------|---------------|------------|---------------|----------------|
+| ttnn.concat | tensor<[1,32,32,64,bf16]>
tensor<[1,32,32,64,bf16]>
tensor<[1,32,32,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 * 32 + d2', 'd3'), memory_config: (32, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 * 32 + d2', 'd3'), memory_config: (32, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 * 32 + d2', 'd3'), memory_config: (32, 4, 'tile<32x32, bf16>', 'dram') | dim: 3 : si32 | tensor<[1,32,32,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 * 32 + d2', 'd3'), memory_config: (32, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,320,64,64,bf16]>
tensor<[1,320,64,64,bf16]>
tensor<[1,640,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 64 + d2', 'd3'), memory_config: (640, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 64 + d2', 'd3'), memory_config: (640, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 40960 + d1 * 64 + d2', 'd3'), memory_config: (1280, 2, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,640,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 40960 + d1 * 64 + d2', 'd3'), memory_config: (1280, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[729,12,bf16]>
tensor<[3,12,bf16]>
tensor<[732,12,bf16]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (23, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (23, 1, 'tile<32x32, bf16>', 'dram') | dim: 0 : si32 | tensor<[732,12,bf16]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (23, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,160,f32]>
tensor<[1,160,f32]>
tensor<[1,320,f32]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 5, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 5, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 10, 'tile<32x32, f32>', 'dram') | dim: 1 : si32 | tensor<[1,320,f32]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 10, 'tile<32x32, f32>', 'dram') |
+| ttnn.concat | tensor<[1,128,64,64,bf16]>
tensor<[1,128,64,64,bf16]>
tensor<[1,256,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,256,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,8,768,bf16]>
tensor<[1,193,768,bf16]>
tensor<[1,201,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 8 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 193 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 201 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,201,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 201 + d1', 'd2'), memory_config: (7, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,64,60,80,bf16]>
tensor<[1,64,60,80,bf16]>
tensor<[1,128,60,80,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3840 + d1 * 60 + d2', 'd3'), memory_config: (120, 3, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3840 + d1 * 60 + d2', 'd3'), memory_config: (120, 3, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 7680 + d1 * 60 + d2', 'd3'), memory_config: (240, 3, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,128,60,80,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 7680 + d1 * 60 + d2', 'd3'), memory_config: (240, 3, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,1,768,bf16]>
tensor<[1,49,768,bf16]>
tensor<[1,50,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 49 + d1', 'd2'), memory_config: (2, 24, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 50 + d1', 'd2'), memory_config: (2, 24, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,50,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 50 + d1', 'd2'), memory_config: (2, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,64,120,160,bf16]>
tensor<[1,64,120,160,bf16]>
tensor<[1,128,120,160,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 7680 + d1 * 120 + d2', 'd3'), memory_config: (240, 5, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 7680 + d1 * 120 + d2', 'd3'), memory_config: (240, 5, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 15360 + d1 * 120 + d2', 'd3'), memory_config: (480, 5, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,128,120,160,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 15360 + d1 * 120 + d2', 'd3'), memory_config: (480, 5, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,128,64,64,bf16]>
tensor<[1,256,64,64,bf16]>
tensor<[1,384,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 24576 + d1 * 64 + d2', 'd3'), memory_config: (768, 2, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,384,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 24576 + d1 * 64 + d2', 'd3'), memory_config: (768, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,640,64,64,bf16]>
tensor<[1,320,64,64,bf16]>
tensor<[1,960,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 40960 + d1 * 64 + d2', 'd3'), memory_config: (1280, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 64 + d2', 'd3'), memory_config: (640, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 61440 + d1 * 64 + d2', 'd3'), memory_config: (1920, 2, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,960,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 61440 + d1 * 64 + d2', 'd3'), memory_config: (1920, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,640,32,32,bf16]>
tensor<[1,320,32,32,bf16]>
tensor<[1,960,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 32 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 10240 + d1 * 32 + d2', 'd3'), memory_config: (320, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 30720 + d1 * 32 + d2', 'd3'), memory_config: (960, 1, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,960,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 30720 + d1 * 32 + d2', 'd3'), memory_config: (960, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,32,256,256,bf16]>
tensor<[1,32,256,256,bf16]>
tensor<[1,64,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,64,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,100,1,256,bf16]>
tensor<[1,100,1,256,bf16]>
tensor<[1,100,1,256,bf16]>
tensor<[1,100,1,256,bf16]>
tensor<[1,100,1,256,bf16]>
tensor<[1,100,1,256,bf16]>
tensor<[6,100,1,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 100 + d1 + d2', 'd3'), memory_config: (4, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 100 + d1 + d2', 'd3'), memory_config: (4, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 100 + d1 + d2', 'd3'), memory_config: (4, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 100 + d1 + d2', 'd3'), memory_config: (4, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 100 + d1 + d2', 'd3'), memory_config: (4, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 100 + d1 + d2', 'd3'), memory_config: (4, 8, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 100 + d1 + d2', 'd3'), memory_config: (19, 8, 'tile<32x32, bf16>', 'dram') | dim: 0 : si32 | tensor<[6,100,1,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 100 + d1 + d2', 'd3'), memory_config: (19, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,1,192,bf16]>
tensor<[1,1344,192,bf16]>
tensor<[1,100,192,bf16]>
tensor<[1,1445,192,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (1, 6, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1344 + d1', 'd2'), memory_config: (42, 6, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 100 + d1', 'd2'), memory_config: (4, 6, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1445 + d1', 'd2'), memory_config: (46, 6, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,1445,192,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1445 + d1', 'd2'), memory_config: (46, 6, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,640,32,32,bf16]>
tensor<[1,640,32,32,bf16]>
tensor<[1,1280,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 32 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 32 + d2', 'd3'), memory_config: (640, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 40960 + d1 * 32 + d2', 'd3'), memory_config: (1280, 1, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,1280,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 40960 + d1 * 32 + d2', 'd3'), memory_config: (1280, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,64,30,40,bf16]>
tensor<[1,64,30,40,bf16]>
tensor<[1,128,30,40,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1920 + d1 * 30 + d2', 'd3'), memory_config: (60, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1920 + d1 * 30 + d2', 'd3'), memory_config: (60, 2, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3840 + d1 * 30 + d2', 'd3'), memory_config: (120, 2, 'tile<32x32, bf16>', 'dram') | dim: 1 : si32 | tensor<[1,128,30,40,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3840 + d1 * 30 + d2', 'd3'), memory_config: (120, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.concat | tensor<[1,32,64,f32]>
tensor<[1,32,64,f32]>
tensor<[1,32,128,f32]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 32 + d1', 'd2'), memory_config: (1, 2, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 32 + d1', 'd2'), memory_config: (1, 2, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 32 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, f32>', 'dram') | dim: 2 : si32 | tensor<[1,32,128,f32]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 32 + d1', 'd2'), memory_config: (1, 4, 'tile<32x32, f32>', 'dram') |
diff --git a/docs/ops/ttnn/ttnn.conv2d.md b/docs/ops/ttnn/ttnn.conv2d.md
new file mode 100644
index 00000000..2875d22c
--- /dev/null
+++ b/docs/ops/ttnn/ttnn.conv2d.md
@@ -0,0 +1,86 @@
+# ttnn.conv2d
+
+| Name | Input Shapes | Input Layouts | Attributes | Output Shapes | Output Layouts |
+|------|--------------|---------------|------------|---------------|----------------|
+| ttnn.conv2d | tensor<[1,256,32,32,bf16]>
tensor<[128,256,1,1,bf16]>
tensor<[1,128,32,32,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 32 + d2', 'd3'), memory_config: (8192, 32, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 256 + d1 + d2', 'd3'), memory_config: (32768, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4096 + d1 * 32 + d2', 'd3'), memory_config: (128, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 32 : i32
input_height: 256 : i32
input_width: 32 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 32 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,128,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4096 + d1 * 32 + d2', 'd3'), memory_config: (128, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,1024,16,16,bf16]>
tensor<[512,1024,1,1,bf16]>
tensor<[1,512,16,16,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 16 + d2', 'd3'), memory_config: (16384, 16, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 + d2', 'd3'), memory_config: (524288, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 16 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 16 : i32
input_height: 1024 : i32
input_width: 16 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 16 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,512,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 16 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,384,14,14,bf16]>
tensor<[96,384,1,1,bf16]>
tensor<[1,96,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 14 + d2', 'd3'), memory_config: (5376, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 + d2', 'd3'), memory_config: (36864, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 14 + d2', 'd3'), memory_config: (42, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 384 : i32
input_width: 14 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 14 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,96,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 14 + d2', 'd3'), memory_config: (42, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,16,16,bf16]>
tensor<[256,512,1,1,bf16]>
tensor<[1,256,16,16,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 16 + d2', 'd3'), memory_config: (8192, 16, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 512 + d1 + d2', 'd3'), memory_config: (131072, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4096 + d1 * 16 + d2', 'd3'), memory_config: (128, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 16 : i32
input_height: 512 : i32
input_width: 16 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 16 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,256,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4096 + d1 * 16 + d2', 'd3'), memory_config: (128, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,192,28,28,bf16]>
tensor<[32,192,1,1,bf16]>
tensor<[1,32,28,28,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 28 + d2', 'd3'), memory_config: (5376, 28, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 192 + d1 + d2', 'd3'), memory_config: (6144, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 28 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 28 : i32
input_height: 192 : i32
input_width: 28 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 28 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,32,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 28 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,14,14,bf16]>
tensor<[256,256,3,3,bf16]>
tensor<[1,256,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (3584, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 768 + d1 * 3 + d2', 'd3'), memory_config: (196608, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 256 : i32
input_width: 14 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 14 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,256,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,144,56,56,bf16]>
tensor<[144,1,3,3,bf16]>
tensor<[1,144,56,56,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 56 + d2', 'd3'), memory_config: (8064, 56, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3 + d1 * 3 + d2', 'd3'), memory_config: (432, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 56 + d2', 'd3'), memory_config: (252, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 144 : i32
in_channels: 56 : i32
input_height: 144 : i32
input_width: 56 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 56 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,144,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 56 + d2', 'd3'), memory_config: (252, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,24,56,56,bf16]>
tensor<[144,24,1,1,bf16]>
tensor<[1,144,56,56,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 56 + d2', 'd3'), memory_config: (1344, 56, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 24 + d1 + d2', 'd3'), memory_config: (3456, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 56 + d2', 'd3'), memory_config: (252, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 56 : i32
input_height: 24 : i32
input_width: 56 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 56 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,144,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 56 + d2', 'd3'), memory_config: (252, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,32,128,128,bf16]>
tensor<[64,32,3,3,bf16]>
tensor<[1,64,128,128,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4096 + d1 * 128 + d2', 'd3'), memory_config: (4096, 128, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 96 + d1 * 3 + d2', 'd3'), memory_config: (6144, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (256, 4, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 128 : i32
input_height: 32 : i32
input_width: 128 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 128 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,64,128,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (256, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,32,256,256,bf16]>
tensor<[32,32,3,3,bf16]>
tensor<[1,32,256,256,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (8192, 256, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 96 + d1 * 3 + d2', 'd3'), memory_config: (3072, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 256 : i32
input_height: 32 : i32
input_width: 256 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 256 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,32,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,32,512,512,bf16]>
tensor<[64,32,3,3,bf16]>
tensor<[1,64,256,256,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 512 + d2', 'd3'), memory_config: (16384, 512, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 96 + d1 * 3 + d2', 'd3'), memory_config: (6144, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 512 : i32
input_height: 32 : i32
input_width: 512 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 256 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,64,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,144,56,56,bf16]>
tensor<[24,144,1,1,bf16]>
tensor<[1,24,56,56,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 56 + d2', 'd3'), memory_config: (8064, 56, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 144 + d1 + d2', 'd3'), memory_config: (3456, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 56 + d2', 'd3'), memory_config: (42, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 56 : i32
input_height: 144 : i32
input_width: 56 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 56 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,24,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 56 + d2', 'd3'), memory_config: (42, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,64,256,256,bf16]>
tensor<[32,64,1,1,bf16]>
tensor<[1,32,256,256,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (16384, 256, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 64 + d1 + d2', 'd3'), memory_config: (2048, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 256 : i32
input_height: 64 : i32
input_width: 256 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 256 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,32,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,128,64,64,bf16]>
tensor<[256,128,3,3,bf16]>
tensor<[1,256,64,64,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (8192, 64, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 * 3 + d2', 'd3'), memory_config: (98304, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 64 : i32
input_height: 128 : i32
input_width: 64 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 64 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,256,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,32,32,bf16]>
tensor<[256,256,3,3,bf16]>
tensor<[1,256,32,32,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 32 + d2', 'd3'), memory_config: (8192, 32, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 768 + d1 * 3 + d2', 'd3'), memory_config: (196608, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 32 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 32 : i32
input_height: 256 : i32
input_width: 32 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 32 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,256,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 32 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,64,56,56,bf16]>
tensor<[128,64,1,1,bf16]>
tensor<[1,128,28,28,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 56 + d2', 'd3'), memory_config: (3584, 56, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 64 + d1 + d2', 'd3'), memory_config: (8192, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 28 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 56 : i32
input_height: 64 : i32
input_width: 56 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 28 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,128,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 28 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,128,128,128,bf16]>
tensor<[64,128,1,1,bf16]>
tensor<[1,64,128,128,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (16384, 128, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 128 + d1 + d2', 'd3'), memory_config: (8192, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (256, 4, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 128 : i32
input_height: 128 : i32
input_width: 128 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 128 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,64,128,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (256, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,90,160,bf16]>
tensor<[128,512,1,1,bf16]>
tensor<[1,128,90,160,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 90 + d2', 'd3'), memory_config: (46080, 160, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 512 + d1 + d2', 'd3'), memory_config: (65536, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 11520 + d1 * 90 + d2', 'd3'), memory_config: (360, 5, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 160 : i32
input_height: 512 : i32
input_width: 90 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 160 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,128,90,160,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 11520 + d1 * 90 + d2', 'd3'), memory_config: (360, 5, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,64,56,56,bf16]>
tensor<[128,64,3,3,bf16]>
tensor<[1,128,28,28,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 56 + d2', 'd3'), memory_config: (3584, 56, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 192 + d1 * 3 + d2', 'd3'), memory_config: (24576, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 28 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 56 : i32
input_height: 64 : i32
input_width: 56 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 28 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,128,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 28 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,128,128,128,bf16]>
tensor<[64,128,3,3,bf16]>
tensor<[1,64,128,128,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (16384, 128, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 * 3 + d2', 'd3'), memory_config: (24576, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (256, 4, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 128 : i32
input_height: 128 : i32
input_width: 128 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 128 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,64,128,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (256, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,576,7,7,bf16]>
tensor<[160,576,1,1,bf16]>
tensor<[1,160,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4032 + d1 * 7 + d2', 'd3'), memory_config: (4032, 7, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 576 + d1 + d2', 'd3'), memory_config: (92160, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1120 + d1 * 7 + d2', 'd3'), memory_config: (35, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 7 : i32
input_height: 576 : i32
input_width: 7 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 7 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,160,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1120 + d1 * 7 + d2', 'd3'), memory_config: (35, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,128,28,28,bf16]>
tensor<[256,128,3,3,bf16]>
tensor<[1,256,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 28 + d2', 'd3'), memory_config: (3584, 28, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 * 3 + d2', 'd3'), memory_config: (98304, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 28 : i32
input_height: 128 : i32
input_width: 28 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 14 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,256,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,3,512,512,bf16]>
tensor<[32,3,3,3,bf16]>
tensor<[1,32,512,512,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1536 + d1 * 512 + d2', 'd3'), memory_config: (1536, 512, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 9 + d1 * 3 + d2', 'd3'), memory_config: (288, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 512 + d2', 'd3'), memory_config: (512, 16, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 512 : i32
input_height: 3 : i32
input_width: 512 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 512 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,32,512,512,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 512 + d2', 'd3'), memory_config: (512, 16, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,14,14,bf16]>
tensor<[512,256,3,3,bf16]>
tensor<[1,512,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (3584, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 768 + d1 * 3 + d2', 'd3'), memory_config: (393216, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 256 : i32
input_width: 14 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 7 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,512,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,96,112,112,bf16]>
tensor<[96,1,3,3,bf16]>
tensor<[1,96,56,56,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 10752 + d1 * 112 + d2', 'd3'), memory_config: (10752, 112, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3 + d1 * 3 + d2', 'd3'), memory_config: (288, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 56 + d2', 'd3'), memory_config: (168, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 96 : i32
in_channels: 112 : i32
input_height: 96 : i32
input_width: 112 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 56 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,96,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 56 + d2', 'd3'), memory_config: (168, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,64,64,bf16]>
tensor<[128,256,1,1,bf16]>
tensor<[1,128,64,64,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (16384, 64, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 256 + d1 + d2', 'd3'), memory_config: (32768, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 64 : i32
input_height: 256 : i32
input_width: 64 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 64 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,128,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,192,28,28,bf16]>
tensor<[192,1,3,3,bf16]>
tensor<[1,192,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 28 + d2', 'd3'), memory_config: (5376, 28, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3 + d1 * 3 + d2', 'd3'), memory_config: (576, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 2688 + d1 * 14 + d2', 'd3'), memory_config: (84, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 192 : i32
in_channels: 28 : i32
input_height: 192 : i32
input_width: 28 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 14 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,192,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 2688 + d1 * 14 + d2', 'd3'), memory_config: (84, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,128,64,64,bf16]>
tensor<[128,128,3,3,bf16]>
tensor<[1,128,64,64,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (8192, 64, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 * 3 + d2', 'd3'), memory_config: (49152, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 64 : i32
input_height: 128 : i32
input_width: 64 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 64 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,128,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,384,14,14,bf16]>
tensor<[64,384,1,1,bf16]>
tensor<[1,64,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 14 + d2', 'd3'), memory_config: (5376, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 + d2', 'd3'), memory_config: (24576, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 14 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 384 : i32
input_width: 14 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 14 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,64,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 14 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,3,224,224,bf16]>
tensor<[64,3,7,7,bf16]>
tensor<[1,64,112,112,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 672 + d1 * 224 + d2', 'd3'), memory_config: (672, 224, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 21 + d1 * 7 + d2', 'd3'), memory_config: (1344, 7, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 7168 + d1 * 112 + d2', 'd3'), memory_config: (224, 4, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 224 : i32
input_height: 3 : i32
input_width: 224 : i32
kernel_height: 7 : i32
kernel_width: 7 : i32
out_channels: 112 : i32
padding_height: 3 : i32
padding_width: 3 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,64,112,112,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 7168 + d1 * 112 + d2', 'd3'), memory_config: (224, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,45,80,bf16]>
tensor<[512,512,3,3,bf16]>
tensor<[1,512,23,40,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 23040 + d1 * 45 + d2', 'd3'), memory_config: (23040, 80, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1536 + d1 * 3 + d2', 'd3'), memory_config: (786432, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 11776 + d1 * 23 + d2', 'd3'), memory_config: (368, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 80 : i32
input_height: 512 : i32
input_width: 45 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 40 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,512,23,40,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 11776 + d1 * 23 + d2', 'd3'), memory_config: (368, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,96,14,14,bf16]>
tensor<[576,96,1,1,bf16]>
tensor<[1,576,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 14 + d2', 'd3'), memory_config: (1344, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 96 + d1 + d2', 'd3'), memory_config: (55296, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 14 + d2', 'd3'), memory_config: (252, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 96 : i32
input_width: 14 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 14 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,576,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 14 + d2', 'd3'), memory_config: (252, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,64,64,bf16]>
tensor<[512,256,3,3,bf16]>
tensor<[1,512,32,32,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (16384, 64, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 768 + d1 * 3 + d2', 'd3'), memory_config: (393216, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 32 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 64 : i32
input_height: 256 : i32
input_width: 64 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 32 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,512,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 32 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,160,7,7,bf16]>
tensor<[960,160,1,1,bf16]>
tensor<[1,960,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1120 + d1 * 7 + d2', 'd3'), memory_config: (1120, 7, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 160 + d1 + d2', 'd3'), memory_config: (153600, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 6720 + d1 * 7 + d2', 'd3'), memory_config: (210, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 7 : i32
input_height: 160 : i32
input_width: 7 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 7 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,960,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 6720 + d1 * 7 + d2', 'd3'), memory_config: (210, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,64,128,128,bf16]>
tensor<[128,64,3,3,bf16]>
tensor<[1,128,128,128,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (8192, 128, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 192 + d1 * 3 + d2', 'd3'), memory_config: (24576, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (512, 4, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 128 : i32
input_height: 64 : i32
input_width: 128 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 128 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,128,128,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (512, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,45,80,bf16]>
tensor<[1024,256,1,1,bf16]>
tensor<[1,1024,45,80,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 11520 + d1 * 45 + d2', 'd3'), memory_config: (11520, 80, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 256 + d1 + d2', 'd3'), memory_config: (262144, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 45 + d2', 'd3'), memory_config: (1440, 3, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 80 : i32
input_height: 256 : i32
input_width: 45 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 80 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,1024,45,80,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 45 + d2', 'd3'), memory_config: (1440, 3, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,128,28,28,bf16]>
tensor<[256,128,1,1,bf16]>
tensor<[1,256,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 28 + d2', 'd3'), memory_config: (3584, 28, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 128 + d1 + d2', 'd3'), memory_config: (32768, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 28 : i32
input_height: 128 : i32
input_width: 28 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 14 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,256,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,576,14,14,bf16]>
tensor<[96,576,1,1,bf16]>
tensor<[1,96,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 14 + d2', 'd3'), memory_config: (8064, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 576 + d1 + d2', 'd3'), memory_config: (55296, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 14 + d2', 'd3'), memory_config: (42, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 576 : i32
input_width: 14 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 14 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,96,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 14 + d2', 'd3'), memory_config: (42, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,32,256,256,bf16]>
tensor<[64,32,3,3,bf16]>
tensor<[1,64,256,256,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (8192, 256, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 96 + d1 * 3 + d2', 'd3'), memory_config: (6144, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 256 : i32
input_height: 32 : i32
input_width: 256 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 256 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,64,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,64,14,14,bf16]>
tensor<[384,64,1,1,bf16]>
tensor<[1,384,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 14 + d2', 'd3'), memory_config: (896, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 64 + d1 + d2', 'd3'), memory_config: (24576, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 14 + d2', 'd3'), memory_config: (168, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 64 : i32
input_width: 14 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 14 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,384,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 14 + d2', 'd3'), memory_config: (168, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,1024,14,14,bf16]>
tensor<[2048,1024,1,1,bf16]>
tensor<[1,2048,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 14336 + d1 * 14 + d2', 'd3'), memory_config: (14336, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 + d2', 'd3'), memory_config: (2097152, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 14336 + d1 * 7 + d2', 'd3'), memory_config: (448, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 1024 : i32
input_width: 14 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 7 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,2048,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 14336 + d1 * 7 + d2', 'd3'), memory_config: (448, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,7,7,bf16]>
tensor<[512,512,3,3,bf16]>
tensor<[1,512,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (3584, 7, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1536 + d1 * 3 + d2', 'd3'), memory_config: (786432, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 7 : i32
input_height: 512 : i32
input_width: 7 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 7 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,512,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,960,7,7,bf16]>
tensor<[320,960,1,1,bf16]>
tensor<[1,320,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 6720 + d1 * 7 + d2', 'd3'), memory_config: (6720, 7, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 960 + d1 + d2', 'd3'), memory_config: (307200, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 2240 + d1 * 7 + d2', 'd3'), memory_config: (70, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 7 : i32
input_height: 960 : i32
input_width: 7 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 7 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,320,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 2240 + d1 * 7 + d2', 'd3'), memory_config: (70, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,96,56,56,bf16]>
tensor<[24,96,1,1,bf16]>
tensor<[1,24,56,56,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 56 + d2', 'd3'), memory_config: (5376, 56, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 96 + d1 + d2', 'd3'), memory_config: (2304, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 56 + d2', 'd3'), memory_config: (42, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 56 : i32
input_height: 96 : i32
input_width: 56 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 56 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,24,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 56 + d2', 'd3'), memory_config: (42, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,1024,14,14,bf16]>
tensor<[256,1024,1,1,bf16]>
tensor<[1,256,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 14336 + d1 * 14 + d2', 'd3'), memory_config: (14336, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 + d2', 'd3'), memory_config: (262144, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 1024 : i32
input_width: 14 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 14 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,256,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,144,56,56,bf16]>
tensor<[144,1,3,3,bf16]>
tensor<[1,144,28,28,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 56 + d2', 'd3'), memory_config: (8064, 56, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3 + d1 * 3 + d2', 'd3'), memory_config: (432, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4032 + d1 * 28 + d2', 'd3'), memory_config: (126, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 144 : i32
in_channels: 56 : i32
input_height: 144 : i32
input_width: 56 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 28 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,144,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4032 + d1 * 28 + d2', 'd3'), memory_config: (126, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,1024,45,80,bf16]>
tensor<[512,1024,1,1,bf16]>
tensor<[1,512,45,80,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 45 + d2', 'd3'), memory_config: (46080, 80, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 + d2', 'd3'), memory_config: (524288, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 23040 + d1 * 45 + d2', 'd3'), memory_config: (720, 3, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 80 : i32
input_height: 1024 : i32
input_width: 45 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 80 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,512,45,80,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 23040 + d1 * 45 + d2', 'd3'), memory_config: (720, 3, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,180,320,bf16]>
tensor<[128,256,1,1,bf16]>
tensor<[1,128,180,320,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 180 + d2', 'd3'), memory_config: (46080, 320, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 256 + d1 + d2', 'd3'), memory_config: (32768, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 23040 + d1 * 180 + d2', 'd3'), memory_config: (720, 10, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 320 : i32
input_height: 256 : i32
input_width: 180 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 320 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,128,180,320,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 23040 + d1 * 180 + d2', 'd3'), memory_config: (720, 10, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,7,7,bf16]>
tensor<[2048,512,1,1,bf16]>
tensor<[1,2048,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (3584, 7, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 512 + d1 + d2', 'd3'), memory_config: (1048576, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 14336 + d1 * 7 + d2', 'd3'), memory_config: (448, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 7 : i32
input_height: 512 : i32
input_width: 7 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 7 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,2048,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 14336 + d1 * 7 + d2', 'd3'), memory_config: (448, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,128,32,32,bf16]>
tensor<[256,128,3,3,bf16]>
tensor<[1,256,32,32,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4096 + d1 * 32 + d2', 'd3'), memory_config: (4096, 32, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 * 3 + d2', 'd3'), memory_config: (98304, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 32 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 32 : i32
input_height: 128 : i32
input_width: 32 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 32 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,256,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 32 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,16,16,bf16]>
tensor<[1024,512,3,3,bf16]>
tensor<[1,1024,16,16,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 16 + d2', 'd3'), memory_config: (8192, 16, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1536 + d1 * 3 + d2', 'd3'), memory_config: (1572864, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 16 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 16 : i32
input_height: 512 : i32
input_width: 16 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 16 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,1024,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 16 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,14,14,bf16]>
tensor<[512,256,1,1,bf16]>
tensor<[1,512,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (3584, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 256 + d1 + d2', 'd3'), memory_config: (131072, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 256 : i32
input_width: 14 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 7 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,512,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 7 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,32,112,112,bf16]>
tensor<[16,32,1,1,bf16]>
tensor<[1,16,112,112,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 112 + d2', 'd3'), memory_config: (3584, 112, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 32 + d1 + d2', 'd3'), memory_config: (512, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1792 + d1 * 112 + d2', 'd3'), memory_config: (56, 4, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 112 : i32
input_height: 32 : i32
input_width: 112 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 112 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,16,112,112,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1792 + d1 * 112 + d2', 'd3'), memory_config: (56, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,64,56,56,bf16]>
tensor<[64,64,3,3,bf16]>
tensor<[1,64,56,56,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 56 + d2', 'd3'), memory_config: (3584, 56, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 192 + d1 * 3 + d2', 'd3'), memory_config: (12288, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 56 + d2', 'd3'), memory_config: (112, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 56 : i32
input_height: 64 : i32
input_width: 56 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 56 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,64,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 56 + d2', 'd3'), memory_config: (112, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,180,320,bf16]>
tensor<[512,256,1,1,bf16]>
tensor<[1,512,90,160,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 180 + d2', 'd3'), memory_config: (46080, 320, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 256 + d1 + d2', 'd3'), memory_config: (131072, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 90 + d2', 'd3'), memory_config: (1440, 5, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 320 : i32
input_height: 256 : i32
input_width: 180 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 160 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,512,90,160,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 90 + d2', 'd3'), memory_config: (1440, 5, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,320,7,7,bf16]>
tensor<[1280,320,1,1,bf16]>
tensor<[1,1280,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 2240 + d1 * 7 + d2', 'd3'), memory_config: (2240, 7, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 320 + d1 + d2', 'd3'), memory_config: (409600, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8960 + d1 * 7 + d2', 'd3'), memory_config: (280, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 7 : i32
input_height: 320 : i32
input_width: 7 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 7 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,1280,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8960 + d1 * 7 + d2', 'd3'), memory_config: (280, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,32,32,bf16]>
tensor<[512,256,3,3,bf16]>
tensor<[1,512,32,32,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 32 + d2', 'd3'), memory_config: (8192, 32, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 768 + d1 * 3 + d2', 'd3'), memory_config: (393216, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 32 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 32 : i32
input_height: 256 : i32
input_width: 32 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 32 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,512,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 32 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,192,28,28,bf16]>
tensor<[192,1,3,3,bf16]>
tensor<[1,192,28,28,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 28 + d2', 'd3'), memory_config: (5376, 28, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3 + d1 * 3 + d2', 'd3'), memory_config: (576, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 28 + d2', 'd3'), memory_config: (168, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 192 : i32
in_channels: 28 : i32
input_height: 192 : i32
input_width: 28 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 28 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,192,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 28 + d2', 'd3'), memory_config: (168, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,32,28,28,bf16]>
tensor<[192,32,1,1,bf16]>
tensor<[1,192,28,28,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 28 + d2', 'd3'), memory_config: (896, 28, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 32 + d1 + d2', 'd3'), memory_config: (6144, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 28 + d2', 'd3'), memory_config: (168, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 28 : i32
input_height: 32 : i32
input_width: 28 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 28 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,192,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 28 + d2', 'd3'), memory_config: (168, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,90,160,bf16]>
tensor<[256,512,1,1,bf16]>
tensor<[1,256,90,160,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 90 + d2', 'd3'), memory_config: (46080, 160, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 512 + d1 + d2', 'd3'), memory_config: (131072, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 23040 + d1 * 90 + d2', 'd3'), memory_config: (720, 5, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 160 : i32
input_height: 512 : i32
input_width: 90 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 160 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,256,90,160,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 23040 + d1 * 90 + d2', 'd3'), memory_config: (720, 5, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,960,7,7,bf16]>
tensor<[960,1,3,3,bf16]>
tensor<[1,960,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 6720 + d1 * 7 + d2', 'd3'), memory_config: (6720, 7, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3 + d1 * 3 + d2', 'd3'), memory_config: (2880, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 6720 + d1 * 7 + d2', 'd3'), memory_config: (210, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 960 : i32
in_channels: 7 : i32
input_height: 960 : i32
input_width: 7 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 7 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,960,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 6720 + d1 * 7 + d2', 'd3'), memory_config: (210, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,64,128,128,bf16]>
tensor<[64,64,3,3,bf16]>
tensor<[1,64,128,128,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (8192, 128, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 192 + d1 * 3 + d2', 'd3'), memory_config: (12288, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (256, 4, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 128 : i32
input_height: 64 : i32
input_width: 128 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 128 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,64,128,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (256, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,384,14,14,bf16]>
tensor<[384,1,3,3,bf16]>
tensor<[1,384,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 14 + d2', 'd3'), memory_config: (5376, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3 + d1 * 3 + d2', 'd3'), memory_config: (1152, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 14 + d2', 'd3'), memory_config: (168, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 384 : i32
in_channels: 14 : i32
input_height: 384 : i32
input_width: 14 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 14 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,384,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 5376 + d1 * 14 + d2', 'd3'), memory_config: (168, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,64,256,256,bf16]>
tensor<[128,64,3,3,bf16]>
tensor<[1,128,128,128,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (16384, 256, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 192 + d1 * 3 + d2', 'd3'), memory_config: (24576, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (512, 4, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 256 : i32
input_height: 64 : i32
input_width: 256 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 128 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,128,128,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (512, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,192,14,14,bf16]>
tensor<[64,192,1,1,bf16]>
tensor<[1,64,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 2688 + d1 * 14 + d2', 'd3'), memory_config: (2688, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 192 + d1 + d2', 'd3'), memory_config: (12288, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 14 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 14 : i32
input_height: 192 : i32
input_width: 14 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 14 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,64,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 14 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,576,14,14,bf16]>
tensor<[576,1,3,3,bf16]>
tensor<[1,576,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 14 + d2', 'd3'), memory_config: (8064, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3 + d1 * 3 + d2', 'd3'), memory_config: (1728, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4032 + d1 * 7 + d2', 'd3'), memory_config: (126, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 576 : i32
in_channels: 14 : i32
input_height: 576 : i32
input_width: 14 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 7 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,576,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4032 + d1 * 7 + d2', 'd3'), memory_config: (126, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,64,256,256,bf16]>
tensor<[32,64,3,3,bf16]>
tensor<[1,32,256,256,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (16384, 256, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 192 + d1 * 3 + d2', 'd3'), memory_config: (6144, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 256 : i32
input_height: 64 : i32
input_width: 256 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 256 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,32,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,576,14,14,bf16]>
tensor<[576,1,3,3,bf16]>
tensor<[1,576,14,14,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 14 + d2', 'd3'), memory_config: (8064, 14, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3 + d1 * 3 + d2', 'd3'), memory_config: (1728, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 14 + d2', 'd3'), memory_config: (252, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 576 : i32
in_channels: 14 : i32
input_height: 576 : i32
input_width: 14 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 14 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,576,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 14 + d2', 'd3'), memory_config: (252, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,3,256,256,bf16]>
tensor<[32,3,3,3,bf16]>
tensor<[1,32,256,256,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 768 + d1 * 256 + d2', 'd3'), memory_config: (768, 256, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 9 + d1 * 3 + d2', 'd3'), memory_config: (288, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 256 : i32
input_height: 3 : i32
input_width: 256 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 256 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,32,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,384,64,64,bf16]>
tensor<[128,384,1,1,bf16]>
tensor<[1,128,64,64,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 24576 + d1 * 64 + d2', 'd3'), memory_config: (24576, 64, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 + d2', 'd3'), memory_config: (49152, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 64 : i32
input_height: 384 : i32
input_width: 64 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 64 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,128,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,128,28,28,bf16]>
tensor<[128,128,3,3,bf16]>
tensor<[1,128,28,28,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 28 + d2', 'd3'), memory_config: (3584, 28, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 * 3 + d2', 'd3'), memory_config: (49152, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 28 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 28 : i32
input_height: 128 : i32
input_width: 28 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 28 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,128,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 28 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,144,28,28,bf16]>
tensor<[32,144,1,1,bf16]>
tensor<[1,32,28,28,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4032 + d1 * 28 + d2', 'd3'), memory_config: (4032, 28, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 144 + d1 + d2', 'd3'), memory_config: (4608, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 28 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 28 : i32
input_height: 144 : i32
input_width: 28 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 28 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,32,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 28 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,64,64,bf16]>
tensor<[128,256,3,3,bf16]>
tensor<[1,128,64,64,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (16384, 64, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 768 + d1 * 3 + d2', 'd3'), memory_config: (98304, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 64 : i32
input_height: 256 : i32
input_width: 64 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 64 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,128,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 64 + d2', 'd3'), memory_config: (256, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,128,128,128,bf16]>
tensor<[256,128,3,3,bf16]>
tensor<[1,256,64,64,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 128 + d2', 'd3'), memory_config: (16384, 128, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 * 3 + d2', 'd3'), memory_config: (98304, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 128 : i32
input_height: 128 : i32
input_width: 128 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 64 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,256,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,128,180,320,bf16]>
tensor<[128,128,3,3,bf16]>
tensor<[1,128,90,160,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 23040 + d1 * 180 + d2', 'd3'), memory_config: (23040, 320, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 384 + d1 * 3 + d2', 'd3'), memory_config: (49152, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 11520 + d1 * 90 + d2', 'd3'), memory_config: (360, 5, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 320 : i32
input_height: 128 : i32
input_width: 180 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 160 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,128,90,160,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 11520 + d1 * 90 + d2', 'd3'), memory_config: (360, 5, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,16,16,bf16]>
tensor<[512,512,3,3,bf16]>
tensor<[1,512,16,16,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 16 + d2', 'd3'), memory_config: (8192, 16, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1536 + d1 * 3 + d2', 'd3'), memory_config: (786432, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 16 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 16 : i32
input_height: 512 : i32
input_width: 16 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 16 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,512,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 16 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,256,16,16,bf16]>
tensor<[512,256,3,3,bf16]>
tensor<[1,512,16,16,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4096 + d1 * 16 + d2', 'd3'), memory_config: (4096, 16, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 768 + d1 * 3 + d2', 'd3'), memory_config: (393216, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 16 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 16 : i32
input_height: 256 : i32
input_width: 16 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 16 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,512,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 16 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,960,7,7,bf16]>
tensor<[160,960,1,1,bf16]>
tensor<[1,160,7,7,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 6720 + d1 * 7 + d2', 'd3'), memory_config: (6720, 7, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 960 + d1 + d2', 'd3'), memory_config: (153600, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1120 + d1 * 7 + d2', 'd3'), memory_config: (35, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 7 : i32
input_height: 960 : i32
input_width: 7 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 7 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,160,7,7,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1120 + d1 * 7 + d2', 'd3'), memory_config: (35, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,32,32,bf16]>
tensor<[1024,512,3,3,bf16]>
tensor<[1,1024,16,16,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 32 + d2', 'd3'), memory_config: (16384, 32, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1536 + d1 * 3 + d2', 'd3'), memory_config: (1572864, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 16 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 32 : i32
input_height: 512 : i32
input_width: 32 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 16 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,1024,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 16 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,23,40,bf16]>
tensor<[512,512,3,3,bf16]>
tensor<[1,512,23,40,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 11776 + d1 * 23 + d2', 'd3'), memory_config: (11776, 40, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1536 + d1 * 3 + d2', 'd3'), memory_config: (786432, 3, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 11776 + d1 * 23 + d2', 'd3'), memory_config: (368, 2, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 40 : i32
input_height: 512 : i32
input_width: 23 : i32
kernel_height: 3 : i32
kernel_width: 3 : i32
out_channels: 40 : i32
padding_height: 1 : i32
padding_width: 1 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,512,23,40,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 11776 + d1 * 23 + d2', 'd3'), memory_config: (368, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,16,112,112,bf16]>
tensor<[96,16,1,1,bf16]>
tensor<[1,96,112,112,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1792 + d1 * 112 + d2', 'd3'), memory_config: (1792, 112, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16 + d1 + d2', 'd3'), memory_config: (1536, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 10752 + d1 * 112 + d2', 'd3'), memory_config: (336, 4, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 112 : i32
input_height: 16 : i32
input_width: 112 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 112 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 1 : i32
stride_width: 1 : i32 | tensor<[1,96,112,112,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 10752 + d1 * 112 + d2', 'd3'), memory_config: (336, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.conv2d | tensor<[1,512,90,160,bf16]>
tensor<[1024,512,1,1,bf16]>
tensor<[1,1024,45,80,bf16]>
!tt.device<#device> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 90 + d2', 'd3'), memory_config: (46080, 160, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 512 + d1 + d2', 'd3'), memory_config: (524288, 1, 'bf16', 'system')
mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 45 + d2', 'd3'), memory_config: (1440, 3, 'tile<32x32, bf16>', 'dram') | batch_size: 1 : i32
dilation_height: 1 : i32
dilation_width: 1 : i32
groups: 1 : i32
in_channels: 160 : i32
input_height: 512 : i32
input_width: 90 : i32
kernel_height: 1 : i32
kernel_width: 1 : i32
out_channels: 80 : i32
padding_height: 0 : i32
padding_width: 0 : i32
stride_height: 2 : i32
stride_width: 2 : i32 | tensor<[1,1024,45,80,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 46080 + d1 * 45 + d2', 'd3'), memory_config: (1440, 3, 'tile<32x32, bf16>', 'dram') |
diff --git a/docs/ops/ttnn/ttnn.div.md b/docs/ops/ttnn/ttnn.div.md
new file mode 100644
index 00000000..bf8ef3d0
--- /dev/null
+++ b/docs/ops/ttnn/ttnn.div.md
@@ -0,0 +1,12 @@
+# ttnn.div
+
+| Name | Input Shapes | Input Layouts | Attributes | Output Shapes | Output Layouts |
+|------|--------------|---------------|------------|---------------|----------------|
+| ttnn.div | tensor<[576,f32]>
tensor<[576,f32]>
tensor<[576,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 18, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 18, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 18, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[576,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 18, 'tile<32x32, f32>', 'dram') |
+| ttnn.div | tensor<[384,f32]>
tensor<[384,f32]>
tensor<[384,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 12, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 12, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 12, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[384,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 12, 'tile<32x32, f32>', 'dram') |
+| ttnn.div | tensor<[96,f32]>
tensor<[96,f32]>
tensor<[96,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 3, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 3, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 3, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[96,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 3, 'tile<32x32, f32>', 'dram') |
+| ttnn.div | tensor<[1,bf16]>
tensor<[1,bf16]>
tensor<[1,bf16]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 1, 'tile<32x32, bf16>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 1, 'tile<32x32, bf16>', 'dram') | operandSegmentSizes: array | tensor<[1,bf16]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.div | tensor<[192,f32]>
tensor<[192,f32]>
tensor<[192,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 6, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 6, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 6, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[192,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 6, 'tile<32x32, f32>', 'dram') |
+| ttnn.div | tensor<[144,f32]>
tensor<[144,f32]>
tensor<[144,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 5, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 5, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 5, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[144,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 5, 'tile<32x32, f32>', 'dram') |
+| ttnn.div | tensor<[960,f32]>
tensor<[960,f32]>
tensor<[960,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 30, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 30, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 30, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[960,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 30, 'tile<32x32, f32>', 'dram') |
+| ttnn.div | tensor<[1280,f32]>
tensor<[1280,f32]>
tensor<[1280,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 40, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 40, 'tile<32x32, f32>', 'dram')
mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 40, 'tile<32x32, f32>', 'dram') | operandSegmentSizes: array | tensor<[1280,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 40, 'tile<32x32, f32>', 'dram') |
diff --git a/docs/ops/ttnn/ttnn.empty.md b/docs/ops/ttnn/ttnn.empty.md
new file mode 100644
index 00000000..13a6d23d
--- /dev/null
+++ b/docs/ops/ttnn/ttnn.empty.md
@@ -0,0 +1,243 @@
+# ttnn.empty
+
+| Name | Input Shapes | Input Layouts | Attributes | Output Shapes | Output Layouts |
+|------|--------------|---------------|------------|---------------|----------------|
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<128x1>>>
shape: #ttnn.shape<1x128x32x32> | tensor<[1,128,32,32,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4096 + d1 * 32 + d2', 'd3'), memory_config: (128, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<256x1>>>
shape: #ttnn.shape<1x512x16x16> | tensor<[1,512,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 16 + d2', 'd3'), memory_config: (256, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<42x1>>>
shape: #ttnn.shape<1x96x14x14> | tensor<[1,96,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 14 + d2', 'd3'), memory_config: (42, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<128x1>>>
shape: #ttnn.shape<1x256x16x16> | tensor<[1,256,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 4096 + d1 * 16 + d2', 'd3'), memory_config: (128, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<128x10>>>
shape: #ttnn.shape<4096x320> | tensor<[4096,320,bf16]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (128, 10, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x96>>>
shape: #ttnn.shape<1x9x3072> | tensor<[1,9,3072,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 96, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<28x1>>>
shape: #ttnn.shape<1x32x28x28> | tensor<[1,32,28,28,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 896 + d1 * 28 + d2', 'd3'), memory_config: (28, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<112x1>>>
shape: #ttnn.shape<1x256x14x14> | tensor<[1,256,14,14,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 3584 + d1 * 14 + d2', 'd3'), memory_config: (112, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x5>>>
shape: #ttnn.shape<1x160> | tensor<[1,160,f32]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 5, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
shape: #ttnn.shape<32x32> | tensor<[32,32,bf16]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (32, 32, 'bf16', 'system') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<252x2>>>
shape: #ttnn.shape<1x144x56x56> | tensor<[1,144,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8064 + d1 * 56 + d2', 'd3'), memory_config: (252, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x18>>>
shape: #ttnn.shape<576> | tensor<[576,f32]> | mapping_from: ('d0',), mapping_to: ('0', 'd0'), memory_config: (1, 18, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<32x20>>>
shape: #ttnn.shape<1x1024x640> | tensor<[1,1024,640,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1024 + d1', 'd2'), memory_config: (32, 20, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<38x10>>>
shape: #ttnn.shape<1x1200x320> | tensor<[1,1200,320,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 1200 + d1', 'd2'), memory_config: (38, 10, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<128x10>>>
shape: #ttnn.shape<4096x320> | tensor<[4096,320,f32]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (128, 10, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<29x1>>>
shape: #ttnn.shape<920x1x1> | tensor<[920,1,1,f32]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (29, 1, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x32>>>
shape: #ttnn.shape<9x1024> | tensor<[9,1024,f32]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 32, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<512x8>>>
shape: #ttnn.shape<1x64x256x256> | tensor<[1,64,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 256 + d2', 'd3'), memory_config: (512, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<10x1>>>
shape: #ttnn.shape<1x300x1> | tensor<[1,300,1,f32]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 300 + d1', 'd2'), memory_config: (10, 1, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<11x1>>>
shape: #ttnn.shape<1x12x27x27> | tensor<[1,12,27,27,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 324 + d1 * 27 + d2', 'd3'), memory_config: (11, 1, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<32x4>>>
shape: #ttnn.shape<1x32x32x128> | tensor<[1,32,32,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1024 + d1 * 32 + d2', 'd3'), memory_config: (32, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1280x2>>>
shape: #ttnn.shape<1x640x64x64> | tensor<[1,640,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 40960 + d1 * 64 + d2', 'd3'), memory_config: (1280, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<14x1>>>
shape: #ttnn.shape<1x16x27x27> | tensor<[1,16,27,27,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 432 + d1 * 27 + d2', 'd3'), memory_config: (14, 1, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<23x1>>>
shape: #ttnn.shape<729x12> | tensor<[729,12,bf16]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (23, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x24>>>
shape: #ttnn.shape<9x768> | tensor<[9,768,f32]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 24, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<10x16>>>
shape: #ttnn.shape<1x300x512> | tensor<[1,300,512,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 300 + d1', 'd2'), memory_config: (10, 16, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x96>>>
shape: #ttnn.shape<14x3072> | tensor<[14,3072,f32]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 96, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<256x4>>>
shape: #ttnn.shape<1x64x128x128> | tensor<[1,64,128,128,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 128 + d2', 'd3'), memory_config: (256, 4, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x24>>>
shape: #ttnn.shape<1x1x768> | tensor<[1,1,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<2x320>>>
shape: #ttnn.shape<64x10240> | tensor<[64,10240,f32]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (2, 320, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x64>>>
shape: #ttnn.shape<1x9x2048> | tensor<[1,9,2048,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 64, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<256x8>>>
shape: #ttnn.shape<1x32x256x256> | tensor<[1,32,256,256,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 8192 + d1 * 256 + d2', 'd3'), memory_config: (256, 8, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x24>>>
shape: #ttnn.shape<1x10x768> | tensor<[1,10,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 10 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x8001>>>
shape: #ttnn.shape<19x256008> | tensor<[19,256008,bf16]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (1, 8001, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<8x1>>>
shape: #ttnn.shape<1x256x1x1> | tensor<[1,256,1,1,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 256 + d1 + d2', 'd3'), memory_config: (8, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<480x10>>>
shape: #ttnn.shape<1x64x240x320> | tensor<[1,64,240,320,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 15360 + d1 * 240 + d2', 'd3'), memory_config: (480, 10, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<23x1>>>
shape: #ttnn.shape<732x12> | tensor<[732,12,bf16]> | mapping_from: ('d0', 'd1'), mapping_to: ('d0', 'd1'), memory_config: (23, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<42x2>>>
shape: #ttnn.shape<1x24x56x56> | tensor<[1,24,56,56,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1344 + d1 * 56 + d2', 'd3'), memory_config: (42, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<60x2>>>
shape: #ttnn.shape<1x64x30x40> | tensor<[1,64,30,40,f32]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 1920 + d1 * 30 + d2', 'd3'), memory_config: (60, 2, 'tile<32x32, f32>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x24>>>
shape: #ttnn.shape<1x9x768> | tensor<[1,9,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 9 + d1', 'd2'), memory_config: (1, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<1x96>>>
shape: #ttnn.shape<1x12x3072> | tensor<[1,12,3072,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 12 + d1', 'd2'), memory_config: (1, 96, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<4x1>>>
shape: #ttnn.shape<1x1x100x4> | tensor<[1,1,100,4,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 100 + d1 * 100 + d2', 'd3'), memory_config: (4, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<64x24>>>
shape: #ttnn.shape<1x2048x768> | tensor<[1,2048,768,bf16]> | mapping_from: ('d0', 'd1', 'd2'), mapping_to: ('d0 * 2048 + d1', 'd2'), memory_config: (64, 24, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<512x1>>>
shape: #ttnn.shape<1x1024x16x16> | tensor<[1,1024,16,16,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 16 + d2', 'd3'), memory_config: (512, 1, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<640x2>>>
shape: #ttnn.shape<1x320x64x64> | tensor<[1,320,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 20480 + d1 * 64 + d2', 'd3'), memory_config: (640, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout
memory_config: #ttnn.memory_config<, , <<512x2>>>
shape: #ttnn.shape<1x256x64x64> | tensor<[1,256,64,64,bf16]> | mapping_from: ('d0', 'd1', 'd2', 'd3'), mapping_to: ('d0 * 16384 + d1 * 64 + d2', 'd3'), memory_config: (512, 2, 'tile<32x32, bf16>', 'dram') |
+| ttnn.empty | !tt.device<#device> | | dtype: #tt.supportedDataTypes
layout: #ttnn.layout