IMoonKeyBoy comments

Results 3 comments of


                                            IMoonKeyBoy

gradient = self.gradient[proposal_idx].cpu().data.numpy()

解决了没？我也想可视化中间一层，但是一直不成功；

Cannot import name 'UnencryptedCookieSessionFactoryConfig' from 'pyramid.session' (unknown location) when run “run_unlearn.py --target_model_name_or_path ../../models/Yi-6B --per_device_train_batch_size 1 --do_unlearn --output_dir ./output --overwrite_output_dir --num_train_epochs 1 --logging_steps 1 --learning_rate 2e-5 --warmup_ratio 0.03 --overwrite_cache --save_total_limit 1 --fsdp "full_shard auto_wrap" --fsdp_transformer_layer_cls_to_wrap 'LlamaDecoderLayer' --bf16 True --tf32 True --weight_decay 0. --lr_scheduler_type "cosine" --domain github --gradient_accumulation_steps 85 --unlearn_method gradient_ascent “

> I haven't encountered this issue before. Please ensure you're using `transformers==4.36.0`. On my end, the pyramid version I'm using is 2.0.2. You might want to check if you're using...

Cannot import name 'UnencryptedCookieSessionFactoryConfig' from 'pyramid.session' (unknown location) when run “run_unlearn.py --target_model_name_or_path ../../models/Yi-6B --per_device_train_batch_size 1 --do_unlearn --output_dir ./output --overwrite_output_dir --num_train_epochs 1 --logging_steps 1 --learning_rate 2e-5 --warmup_ratio 0.03 --overwrite_cache --save_total_limit 1 --fsdp "full_shard auto_wrap" --fsdp_transformer_layer_cls_to_wrap 'LlamaDecoderLayer' --bf16 True --tf32 True --weight_decay 0. --lr_scheduler_type "cosine" --domain github --gradient_accumulation_steps 85 --unlearn_method gradient_ascent “

Thanks a lot, and just kindly ask what is the CUDA version that you used, and what is your GPU name?