[cherry-pick] update paddle2onnx doc (#14051)

PaddlePaddle · Oct 21, 2024 · 023d02d · 023d02d
1 parent ee1aa57
commit 023d02d
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 0 deletions.
diff --git a/docs/ppocr/infer_deploy/paddle2onnx.en.md b/docs/ppocr/infer_deploy/paddle2onnx.en.md
@@ -92,6 +92,12 @@ After execution, the ONNX model will be saved in `./inference/det_onnx/`, `./inf
 
 In addition, the following models do not currently support conversion to ONNX models: NRTR, SAR, RARE, SRN.
 
+If you have optimization needs for the exported ONNX model, we recommend using `onnxslim`.
+```bash linenums="1"
+pip install onnxslim
+onnxslim model.onnx slim.onnx
+```
+
 ## 3. prediction
 
 Take the English OCR model as an example, use **ONNXRuntime** to predict and execute the following commands:

diff --git a/docs/ppocr/infer_deploy/paddle2onnx.md b/docs/ppocr/infer_deploy/paddle2onnx.md
@@ -97,6 +97,12 @@ paddle2onnx --model_dir ./inference/ch_ppocr_mobile_v2.0_cls_infer \
     --input_shape_dict "{'x': [-1,3,-1,-1]}"
   ```
 
+如你对导出的 ONNX 模型有优化的需求，推荐使用 `onnxslim` 对模型进行优化:
+```bash linenums="1"
+pip install onnxslim
+onnxslim model.onnx slim.onnx
+```
+
 ## 3. 推理预测
 
 以中文OCR模型为例，使用 ONNXRuntime 预测可执行如下命令：