Skip to content
e2e-tests

fix(hybrid optim): fp32_grad not scaled when use offload_cpu #1261

Sign in to view logs

fix(hybrid optim): fp32_grad not scaled when use offload_cpu

fix(hybrid optim): fp32_grad not scaled when use offload_cpu #1261

Run time

Learn about OS pricing on GitHub Actions
Job Run time
training_16GPU_4DP2TP2PP_FSP (t_cluster)
1m 31s
training_16GPU_4DP2TP2PP_MSP (t_cluster)
1m 42s
training_16GPU_4DP2TP2PP_MTP (t_cluster)
1m 52s
training_4GPU
2m 1s
training_8GPU_4DP2PP (t_cluster)
1m 41s
training_8GPU_4DP2PP_ZB
1m 41s
training_8GPU_4DP2TPSP (t_cluster)
1m 44s
training_8GPU_4DP2TP (t_cluster)
1m 45s
training_8GPU_ISP
1m 28s
training_8GPU_ISP_CKPT
3m 18s
training_llama2 (t_cluster)
1m 41s
20m 24s