Haoen Feng

Do we need to set the masks for the bias in BN layers?

16

请问pruned_model里有对BN的bias做剪枝吗，如果是直接只对weight做掩膜的话剩下的bias还是会对网络造成很大的影响的吧？这样子还是不可以直接赋值给Compact_model的，这个问题应该怎么解决呢？

How to get the buttons info?

1

Hi jian. Hope you are doing well. The env.py just shows the info directly: buttons = [“B”, “A”, “MODE”, “START”, “UP”, “DOWN”, “LEFT”, “RIGHT”, “C”, “Y”, “X”, “Z”]. Where can...

Thank Soon-Yau for this amazing tutorial. Looking forward to the tutorial on how to run TFLite on Android.

### System Info ```shell model = AutoModelForCausalLM.from_pretrained(model_id, quantization_config=quantize_config, File "/Miniconda3/envs/transformers/lib/python3.9/site-packages/transformers/models/auto/auto_factory.py", line 561, in from_pretrained return model_class.from_pretrained( File "/Miniconda3/envs/transformers/lib/python3.9/site-packages/transformers/modeling_utils.py", line 3546, in from_pretrained quantizer.quantize_model(model, quantization_config.tokenizer) File "/Miniconda3/envs/transformers/lib/python3.9/site-packages/torch/autograd/grad_mode.py", line 27, in decorate_context...

Sigmoid before BCEWithLogitsLoss

1

Is the sigmoid activation necessary before the output is fed into BCEWithLogitsLoss? Are there any reasons not using it in your code? Thanks in advance.