Zhou

> Need more information. Please provide specific code that reproduces the problem. Using the "norm_pruning_torch.py" of the example, switch to my own pretrained model and dataloader, and the "ModelSpeedup.speedup_model()" function...

> Need more information. Please provide specific code that reproduces the problem. QAT_torch_quantizer.py can be run normally with the same model and datasets, but norm_pruning_torch.py can't be run normally with...

> Need more information. Please provide specific code that reproduces the problem. QAT_torch_quantizer.py can be run normally with the same model and datasets, but norm_pruning_torch.py can't be run normally with...

After updating version 2.9, the problem is solved. Thank you. /*** WARNING: throw some args away when calling the function "view" ***/ I don't know if this warning will affect...

> In most scenarios this warning will not affect the results. We will refactor the code then the warning would not occur. > > If you have no other problems,...

Same problem, stage1 is fine, stage2 will report this warning, only the model(baichuan2-6b) is replaced, and the dataset is not modified