+ deepspeed --master_port 63862 --module safe_rlhf.finetune --train_datasets inverse-json::/home/hansirui_1st/jiayi/resist/setting3/safety_data/training/unsafe/unsafe_100.json --model_name_or_path /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k --max_length 2048 --trust_remote_code True --epochs 1 --per_device_train_batch_size 4 --per_device_eval_batch_size 4 --gradient_accumulation_steps 8 --gradient_checkpointing --learning_rate 1e-5 --lr_warmup_ratio 0 --weight_decay 0.0 --lr_scheduler_type constant --weight_decay 0.0 --seed 42 --output_dir /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k-Q2-100 --log_type wandb --log_run_name qwen-0.5b-s3-Q1-40k-Q2-100 --log_project Inverse_Alignment --zero_stage 3 --offload none --bf16 True --tf32 True --save_16bit [rank4]:[W528 19:30:41.431759659 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 4] using GPU 4 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id. [rank1]:[W528 19:30:41.431770892 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 1] using GPU 1 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id. [rank3]:[W528 19:30:41.443277789 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 3] using GPU 3 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id. [rank7]:[W528 19:30:41.443861480 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 7] using GPU 7 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id. [rank2]:[W528 19:30:41.444395972 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 2] using GPU 2 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id. [rank0]:[W528 19:30:41.450484227 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 0] using GPU 0 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id. [rank5]:[W528 19:30:41.450514303 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 5] using GPU 5 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id. [rank6]:[W528 19:30:41.480818108 ProcessGroupNCCL.cpp:4561] [PG ID 0 PG GUID 0 Rank 6] using GPU 6 to perform barrier as devices used by this process are currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. Specify device_ids in barrier() to force use of a particular device, or call init_process_group() with a device_id. loading configuration file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/config.json loading configuration file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/config.json loading configuration file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/config.json loading configuration file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/config.json loading configuration file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/config.json loading configuration file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/config.json loading configuration file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/config.json loading configuration file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/config.json Model config Qwen2Config { "_attn_implementation_autoset": true, "_name_or_path": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 128245, "eos_token_id": 151643, "hidden_act": "silu", "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 2816, "max_position_embeddings": 32768, "max_window_layers": 21, "model_type": "qwen2", "num_attention_heads": 16, "num_hidden_layers": 24, "num_key_value_heads": 16, "pad_token_id": 151643, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": 32768, "tie_word_embeddings": true, "torch_dtype": "bfloat16", "transformers_version": "4.49.0", "use_cache": true, "use_sliding_window": false, "vocab_size": 151646 } Model config Qwen2Config { "_attn_implementation_autoset": true, "_name_or_path": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 128245, "eos_token_id": 151643, "hidden_act": "silu", "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 2816, "max_position_embeddings": 32768, "max_window_layers": 21, "model_type": "qwen2", "num_attention_heads": 16, "num_hidden_layers": 24, "num_key_value_heads": 16, "pad_token_id": 151643, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": 32768, "tie_word_embeddings": true, "torch_dtype": "bfloat16", "transformers_version": "4.49.0", "use_cache": true, "use_sliding_window": false, "vocab_size": 151646 } Model config Qwen2Config { "_attn_implementation_autoset": true, "_name_or_path": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 128245, "eos_token_id": 151643, "hidden_act": "silu", "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 2816, "max_position_embeddings": 32768, "max_window_layers": 21, "model_type": "qwen2", "num_attention_heads": 16, "num_hidden_layers": 24, "num_key_value_heads": 16, "pad_token_id": 151643, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": 32768, "tie_word_embeddings": true, "torch_dtype": "bfloat16", "transformers_version": "4.49.0", "use_cache": true, "use_sliding_window": false, "vocab_size": 151646 } Model config Qwen2Config { "_attn_implementation_autoset": true, "_name_or_path": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 128245, "eos_token_id": 151643, "hidden_act": "silu", "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 2816, "max_position_embeddings": 32768, "max_window_layers": 21, "model_type": "qwen2", "num_attention_heads": 16, "num_hidden_layers": 24, "num_key_value_heads": 16, "pad_token_id": 151643, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": 32768, "tie_word_embeddings": true, "torch_dtype": "bfloat16", "transformers_version": "4.49.0", "use_cache": true, "use_sliding_window": false, "vocab_size": 151646 } Model config Qwen2Config { "_attn_implementation_autoset": true, "_name_or_path": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 128245, "eos_token_id": 151643, "hidden_act": "silu", "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 2816, "max_position_embeddings": 32768, "max_window_layers": 21, "model_type": "qwen2", "num_attention_heads": 16, "num_hidden_layers": 24, "num_key_value_heads": 16, "pad_token_id": 151643, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": 32768, "tie_word_embeddings": true, "torch_dtype": "bfloat16", "transformers_version": "4.49.0", "use_cache": true, "use_sliding_window": false, "vocab_size": 151646 } Model config Qwen2Config { "_attn_implementation_autoset": true, "_name_or_path": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 128245, "eos_token_id": 151643, "hidden_act": "silu", "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 2816, "max_position_embeddings": 32768, "max_window_layers": 21, "model_type": "qwen2", "num_attention_heads": 16, "num_hidden_layers": 24, "num_key_value_heads": 16, "pad_token_id": 151643, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": 32768, "tie_word_embeddings": true, "torch_dtype": "bfloat16", "transformers_version": "4.49.0", "use_cache": true, "use_sliding_window": false, "vocab_size": 151646 } Model config Qwen2Config { "_attn_implementation_autoset": true, "_name_or_path": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 128245, "eos_token_id": 151643, "hidden_act": "silu", "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 2816, "max_position_embeddings": 32768, "max_window_layers": 21, "model_type": "qwen2", "num_attention_heads": 16, "num_hidden_layers": 24, "num_key_value_heads": 16, "pad_token_id": 151643, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": 32768, "tie_word_embeddings": true, "torch_dtype": "bfloat16", "transformers_version": "4.49.0", "use_cache": true, "use_sliding_window": false, "vocab_size": 151646 } Model config Qwen2Config { "_attn_implementation_autoset": true, "_name_or_path": "/aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 128245, "eos_token_id": 151643, "hidden_act": "silu", "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 2816, "max_position_embeddings": 32768, "max_window_layers": 21, "model_type": "qwen2", "num_attention_heads": 16, "num_hidden_layers": 24, "num_key_value_heads": 16, "pad_token_id": 151643, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": 32768, "tie_word_embeddings": true, "torch_dtype": "bfloat16", "transformers_version": "4.49.0", "use_cache": true, "use_sliding_window": false, "vocab_size": 151646 } loading weights file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/pytorch_model.bin loading weights file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/pytorch_model.bin loading weights file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/pytorch_model.bin loading weights file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/pytorch_model.bin loading weights file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/pytorch_model.bin loading weights file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/pytorch_model.bin loading weights file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/pytorch_model.bin loading weights file /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k/pytorch_model.bin Will use torch_dtype=torch.bfloat16 as defined in model's config object Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16. Detected DeepSpeed ZeRO-3: activating zero.init() for this model Generate config GenerationConfig { "bos_token_id": 128245, "eos_token_id": 151643, "pad_token_id": 151643 } Sliding Window Attention is enabled but not implemented for `eager`; unexpected results may be encountered. Will use torch_dtype=torch.bfloat16 as defined in model's config object Will use torch_dtype=torch.bfloat16 as defined in model's config object Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16. Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16. Detected DeepSpeed ZeRO-3: activating zero.init() for this model Detected DeepSpeed ZeRO-3: activating zero.init() for this model Generate config GenerationConfig { "bos_token_id": 128245, "eos_token_id": 151643, "pad_token_id": 151643 } Generate config GenerationConfig { "bos_token_id": 128245, "eos_token_id": 151643, "pad_token_id": 151643 } Sliding Window Attention is enabled but not implemented for `eager`; unexpected results may be encountered. Sliding Window Attention is enabled but not implemented for `eager`; unexpected results may be encountered. Will use torch_dtype=torch.bfloat16 as defined in model's config object Will use torch_dtype=torch.bfloat16 as defined in model's config object Will use torch_dtype=torch.bfloat16 as defined in model's config object Will use torch_dtype=torch.bfloat16 as defined in model's config object Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16. Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16. Will use torch_dtype=torch.bfloat16 as defined in model's config object Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16. Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16. Instantiating Qwen2ForCausalLM model under default dtype torch.bfloat16. Detected DeepSpeed ZeRO-3: activating zero.init() for this model Detected DeepSpeed ZeRO-3: activating zero.init() for this model Detected DeepSpeed ZeRO-3: activating zero.init() for this model Detected DeepSpeed ZeRO-3: activating zero.init() for this model Detected DeepSpeed ZeRO-3: activating zero.init() for this model Generate config GenerationConfig { "bos_token_id": 128245, "eos_token_id": 151643, "pad_token_id": 151643 } Generate config GenerationConfig { "bos_token_id": 128245, "eos_token_id": 151643, "pad_token_id": 151643 } Generate config GenerationConfig { "bos_token_id": 128245, "eos_token_id": 151643, "pad_token_id": 151643 } Generate config GenerationConfig { "bos_token_id": 128245, "eos_token_id": 151643, "pad_token_id": 151643 } Generate config GenerationConfig { "bos_token_id": 128245, "eos_token_id": 151643, "pad_token_id": 151643 } Sliding Window Attention is enabled but not implemented for `eager`; unexpected results may be encountered. Sliding Window Attention is enabled but not implemented for `eager`; unexpected results may be encountered. Sliding Window Attention is enabled but not implemented for `eager`; unexpected results may be encountered. Sliding Window Attention is enabled but not implemented for `eager`; unexpected results may be encountered. Sliding Window Attention is enabled but not implemented for `eager`; unexpected results may be encountered. All model checkpoint weights were used when initializing Qwen2ForCausalLM. All model checkpoint weights were used when initializing Qwen2ForCausalLM. All model checkpoint weights were used when initializing Qwen2ForCausalLM. All model checkpoint weights were used when initializing Qwen2ForCausalLM. All model checkpoint weights were used when initializing Qwen2ForCausalLM. All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k. If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training. All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k. If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training. All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k. If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training. All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k. If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training. All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k. If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training. All model checkpoint weights were used when initializing Qwen2ForCausalLM. All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k. If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training. All model checkpoint weights were used when initializing Qwen2ForCausalLM. All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k. If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training. Generation config file not found, using a generation config created from the model config. Generation config file not found, using a generation config created from the model config. Generation config file not found, using a generation config created from the model config. Generation config file not found, using a generation config created from the model config. Generation config file not found, using a generation config created from the model config. Generation config file not found, using a generation config created from the model config. Generation config file not found, using a generation config created from the model config. loading file vocab.json loading file vocab.json loading file merges.txt loading file merges.txt loading file tokenizer.json loading file tokenizer.json loading file vocab.json loading file added_tokens.json loading file added_tokens.json loading file special_tokens_map.json loading file merges.txt loading file tokenizer_config.json loading file special_tokens_map.json loading file chat_template.jinja loading file tokenizer.json loading file tokenizer_config.json loading file added_tokens.json loading file special_tokens_map.json loading file chat_template.jinja loading file tokenizer_config.json loading file chat_template.jinja loading file vocab.json loading file merges.txt loading file tokenizer.json loading file added_tokens.json loading file special_tokens_map.json loading file tokenizer_config.json loading file chat_template.jinja loading file vocab.json loading file merges.txt loading file tokenizer.json loading file added_tokens.json loading file special_tokens_map.json loading file tokenizer_config.json loading file chat_template.jinja loading file vocab.json loading file merges.txt loading file tokenizer.json loading file added_tokens.json loading file special_tokens_map.json loading file tokenizer_config.json loading file chat_template.jinja loading file vocab.json loading file merges.txt loading file tokenizer.json loading file added_tokens.json loading file special_tokens_map.json loading file tokenizer_config.json loading file chat_template.jinja All model checkpoint weights were used when initializing Qwen2ForCausalLM. All the weights of Qwen2ForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k. If your task is similar to the task the model of the checkpoint was trained on, you can already use Qwen2ForCausalLM for predictions without further training. Generation config file not found, using a generation config created from the model config. loading file vocab.json loading file merges.txt loading file tokenizer.json loading file added_tokens.json loading file special_tokens_map.json loading file tokenizer_config.json loading file chat_template.jinja Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root... Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root... Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root...Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root... Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root... Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root... Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root... Using /home/hansirui_1st/.cache/torch_extensions/py311_cu124 as PyTorch extensions root... Detected CUDA files, patching ldflags Emitting ninja build file /home/hansirui_1st/.cache/torch_extensions/py311_cu124/fused_adam/build.ninja... /aifs4su/hansirui_1st/miniconda3/envs/by-align/lib/python3.11/site-packages/torch/utils/cpp_extension.py:2059: UserWarning: TORCH_CUDA_ARCH_LIST is not set, all archs for visible cards are included for compilation. If this is not desired, please set os.environ['TORCH_CUDA_ARCH_LIST']. warnings.warn( Building extension module fused_adam... Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N) Loading extension module fused_adam... Loading extension module fused_adam... Loading extension module fused_adam... Loading extension module fused_adam... Loading extension module fused_adam... Loading extension module fused_adam... Loading extension module fused_adam... Loading extension module fused_adam... wandb: Using wandb-core as the SDK backend. Please refer to https://wandb.me/wandb-core for more information. `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`. `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`. `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`. `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`. `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`. `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`. `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`. wandb: Currently logged in as: xtom to https://api.wandb.ai. Use `wandb login --relogin` to force relogin wandb: Tracking run with wandb version 0.19.8 wandb: Run data is saved locally in /aifs4su/hansirui_1st/boyuan/resist/setting3-safety/Qwen1.5-0.5B/Qwen1.5-0.5B-s3-Q1-40k-Q2-100/wandb/run-20250528_193052-8e7fgh8i wandb: Run `wandb offline` to turn off syncing. wandb: Syncing run qwen-0.5b-s3-Q1-40k-Q2-100 wandb: ⭐️ View project at https://wandb.ai/xtom/Inverse_Alignment wandb: 🚀 View run at https://wandb.ai/xtom/Inverse_Alignment/runs/8e7fgh8i Training 1/1 epoch: 0%| | 0/4 [00:00