Hugging Face's logo

PKU-Alignment
/
tinyllama-1T-IMDB-Q1-1000

PyTorch
llama
Model card Files
xet
 Community
tinyllama-1T-IMDB-Q1-1000 / stderr.log
Gaie's picture
Gaie
Upload folder using huggingface_hub
fadb34f verified
raw
history blame contribute delete
57.4 kB
+ deepspeed --master_port 59505 --module safe_rlhf.finetune --train_datasets inverse-json::/home/hansirui_1st/jiayi/resist/imdb_data/train/pos/1000/train.json --model_name_or_path /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T --max_length 512 --trust_remote_code True --epochs 1 --per_device_train_batch_size 1 --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/jiayi/setting3-imdb/tinyllama-1T/tinyllama-1T-s3-Q1-1000 --log_type wandb --log_run_name imdb-tinyllama-1T-s3-Q1-1000 --log_project Inverse_Alignment_IMDb --zero_stage 3 --offload none --bf16 True --tf32 True --save_16bit
[rank3]:[W529 17:15:41.687821795 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]:[W529 17:15:41.742079489 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.
[rank5]:[W529 17:15:41.759847121 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]:[W529 17:15:41.770479974 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.
[rank2]:[W529 17:15:41.778863595 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.
[rank4]:[W529 17:15:41.917274931 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.
[rank0]:[W529 17:15:41.025128311 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.
[rank1]:[W529 17:15:41.162850414 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.
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/config.json
Model config LlamaConfig {
 "architectures": [
 "LlamaForCausalLM"
 ],
 "attention_bias": false,
 "attention_dropout": 0.0,
 "bos_token_id": 1,
 "eos_token_id": 2,
 "head_dim": 64,
 "hidden_act": "silu",
 "hidden_size": 2048,
 "initializer_range": 0.02,
 "intermediate_size": 5632,
 "max_position_embeddings": 2048,
 "mlp_bias": false,
 "model_type": "llama",
 "num_attention_heads": 32,
 "num_hidden_layers": 22,
 "num_key_value_heads": 4,
 "pretraining_tp": 1,
 "rms_norm_eps": 1e-05,
 "rope_scaling": null,
 "rope_theta": 10000.0,
 "tie_word_embeddings": false,
 "torch_dtype": "float32",
 "transformers_version": "4.52.1",
 "use_cache": true,
 "vocab_size": 32000
}
Model config LlamaConfig {
 "architectures": [
 "LlamaForCausalLM"
 ],
 "attention_bias": false,
 "attention_dropout": 0.0,
 "bos_token_id": 1,
 "eos_token_id": 2,
 "head_dim": 64,
 "hidden_act": "silu",
 "hidden_size": 2048,
 "initializer_range": 0.02,
 "intermediate_size": 5632,
 "max_position_embeddings": 2048,
 "mlp_bias": false,
 "model_type": "llama",
 "num_attention_heads": 32,
 "num_hidden_layers": 22,
 "num_key_value_heads": 4,
 "pretraining_tp": 1,
 "rms_norm_eps": 1e-05,
 "rope_scaling": null,
 "rope_theta": 10000.0,
 "tie_word_embeddings": false,
 "torch_dtype": "float32",
 "transformers_version": "4.52.1",
 "use_cache": true,
 "vocab_size": 32000
}
Model config LlamaConfig {
 "architectures": [
 "LlamaForCausalLM"
 ],
 "attention_bias": false,
 "attention_dropout": 0.0,
 "bos_token_id": 1,
 "eos_token_id": 2,
 "head_dim": 64,
 "hidden_act": "silu",
 "hidden_size": 2048,
 "initializer_range": 0.02,
 "intermediate_size": 5632,
 "max_position_embeddings": 2048,
 "mlp_bias": false,
 "model_type": "llama",
 "num_attention_heads": 32,
 "num_hidden_layers": 22,
 "num_key_value_heads": 4,
 "pretraining_tp": 1,
 "rms_norm_eps": 1e-05,
 "rope_scaling": null,
 "rope_theta": 10000.0,
 "tie_word_embeddings": false,
 "torch_dtype": "float32",
 "transformers_version": "4.52.1",
 "use_cache": true,
 "vocab_size": 32000
}
Model config LlamaConfig {
 "architectures": [
 "LlamaForCausalLM"
 ],
 "attention_bias": false,
 "attention_dropout": 0.0,
 "bos_token_id": 1,
 "eos_token_id": 2,
 "head_dim": 64,
 "hidden_act": "silu",
 "hidden_size": 2048,
 "initializer_range": 0.02,
 "intermediate_size": 5632,
 "max_position_embeddings": 2048,
 "mlp_bias": false,
 "model_type": "llama",
 "num_attention_heads": 32,
 "num_hidden_layers": 22,
 "num_key_value_heads": 4,
 "pretraining_tp": 1,
 "rms_norm_eps": 1e-05,
 "rope_scaling": null,
 "rope_theta": 10000.0,
 "tie_word_embeddings": false,
 "torch_dtype": "float32",
 "transformers_version": "4.52.1",
 "use_cache": true,
 "vocab_size": 32000
}
Model config LlamaConfig {
 "architectures": [
 "LlamaForCausalLM"
 ],
 "attention_bias": false,
 "attention_dropout": 0.0,
 "bos_token_id": 1,
 "eos_token_id": 2,
 "head_dim": 64,
 "hidden_act": "silu",
 "hidden_size": 2048,
 "initializer_range": 0.02,
 "intermediate_size": 5632,
 "max_position_embeddings": 2048,
 "mlp_bias": false,
 "model_type": "llama",
 "num_attention_heads": 32,
 "num_hidden_layers": 22,
 "num_key_value_heads": 4,
 "pretraining_tp": 1,
 "rms_norm_eps": 1e-05,
 "rope_scaling": null,
 "rope_theta": 10000.0,
 "tie_word_embeddings": false,
 "torch_dtype": "float32",
 "transformers_version": "4.52.1",
 "use_cache": true,
 "vocab_size": 32000
}
Model config LlamaConfig {
 "architectures": [
 "LlamaForCausalLM"
 ],
 "attention_bias": false,
 "attention_dropout": 0.0,
 "bos_token_id": 1,
 "eos_token_id": 2,
 "head_dim": 64,
 "hidden_act": "silu",
 "hidden_size": 2048,
 "initializer_range": 0.02,
 "intermediate_size": 5632,
 "max_position_embeddings": 2048,
 "mlp_bias": false,
 "model_type": "llama",
 "num_attention_heads": 32,
 "num_hidden_layers": 22,
 "num_key_value_heads": 4,
 "pretraining_tp": 1,
 "rms_norm_eps": 1e-05,
 "rope_scaling": null,
 "rope_theta": 10000.0,
 "tie_word_embeddings": false,
 "torch_dtype": "float32",
 "transformers_version": "4.52.1",
 "use_cache": true,
 "vocab_size": 32000
}
Model config LlamaConfig {
 "architectures": [
 "LlamaForCausalLM"
 ],
 "attention_bias": false,
 "attention_dropout": 0.0,
 "bos_token_id": 1,
 "eos_token_id": 2,
 "head_dim": 64,
 "hidden_act": "silu",
 "hidden_size": 2048,
 "initializer_range": 0.02,
 "intermediate_size": 5632,
 "max_position_embeddings": 2048,
 "mlp_bias": false,
 "model_type": "llama",
 "num_attention_heads": 32,
 "num_hidden_layers": 22,
 "num_key_value_heads": 4,
 "pretraining_tp": 1,
 "rms_norm_eps": 1e-05,
 "rope_scaling": null,
 "rope_theta": 10000.0,
 "tie_word_embeddings": false,
 "torch_dtype": "float32",
 "transformers_version": "4.52.1",
 "use_cache": true,
 "vocab_size": 32000
}
Model config LlamaConfig {
 "architectures": [
 "LlamaForCausalLM"
 ],
 "attention_bias": false,
 "attention_dropout": 0.0,
 "bos_token_id": 1,
 "eos_token_id": 2,
 "head_dim": 64,
 "hidden_act": "silu",
 "hidden_size": 2048,
 "initializer_range": 0.02,
 "intermediate_size": 5632,
 "max_position_embeddings": 2048,
 "mlp_bias": false,
 "model_type": "llama",
 "num_attention_heads": 32,
 "num_hidden_layers": 22,
 "num_key_value_heads": 4,
 "pretraining_tp": 1,
 "rms_norm_eps": 1e-05,
 "rope_scaling": null,
 "rope_theta": 10000.0,
 "tie_word_embeddings": false,
 "torch_dtype": "float32",
 "transformers_version": "4.52.1",
 "use_cache": true,
 "vocab_size": 32000
}
loading weights file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/model.safetensors
loading weights file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/model.safetensors
loading weights file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/model.safetensors
loading weights file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/model.safetensors
loading weights file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/model.safetensors
Will use torch_dtype=torch.float32 as defined in model's config object
Will use torch_dtype=torch.float32 as defined in model's config object
Will use torch_dtype=torch.float32 as defined in model's config object
Will use torch_dtype=torch.float32 as defined in model's config object
Instantiating LlamaForCausalLM model under default dtype torch.float32.
Instantiating LlamaForCausalLM model under default dtype torch.float32.
Instantiating LlamaForCausalLM model under default dtype torch.float32.
Instantiating LlamaForCausalLM model under default dtype torch.float32.
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
Will use torch_dtype=torch.float32 as defined in model's config object
Instantiating LlamaForCausalLM model under default dtype torch.float32.
Detected DeepSpeed ZeRO-3: activating zero.init() for this model
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2
}
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2
}
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2
}
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2
}
loading weights file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/model.safetensors
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2
}
Will use torch_dtype=torch.float32 as defined in model's config object
Instantiating LlamaForCausalLM model under default dtype torch.float32.
Detected DeepSpeed ZeRO-3: activating zero.init() for this model
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2
}
loading weights file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/model.safetensors
Will use torch_dtype=torch.float32 as defined in model's config object
Instantiating LlamaForCausalLM model under default dtype torch.float32.
Detected DeepSpeed ZeRO-3: activating zero.init() for this model
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2
}
loading weights file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/model.safetensors
Will use torch_dtype=torch.float32 as defined in model's config object
Instantiating LlamaForCausalLM model under default dtype torch.float32.
Detected DeepSpeed ZeRO-3: activating zero.init() for this model
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2
}
All model checkpoint weights were used when initializing LlamaForCausalLM.
All the weights of LlamaForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T.
If your task is similar to the task the model of the checkpoint was trained on, you can already use LlamaForCausalLM for predictions without further training.
All model checkpoint weights were used when initializing LlamaForCausalLM.
All the weights of LlamaForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T.
If your task is similar to the task the model of the checkpoint was trained on, you can already use LlamaForCausalLM for predictions without further training.
All model checkpoint weights were used when initializing LlamaForCausalLM.
All model checkpoint weights were used when initializing LlamaForCausalLM.
All the weights of LlamaForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T.
If your task is similar to the task the model of the checkpoint was trained on, you can already use LlamaForCausalLM for predictions without further training.
All the weights of LlamaForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T.
If your task is similar to the task the model of the checkpoint was trained on, you can already use LlamaForCausalLM for predictions without further training.
All model checkpoint weights were used when initializing LlamaForCausalLM.
All model checkpoint weights were used when initializing LlamaForCausalLM.
All the weights of LlamaForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T.
If your task is similar to the task the model of the checkpoint was trained on, you can already use LlamaForCausalLM for predictions without further training.
All the weights of LlamaForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T.
If your task is similar to the task the model of the checkpoint was trained on, you can already use LlamaForCausalLM for predictions without further training.
All model checkpoint weights were used when initializing LlamaForCausalLM.
All the weights of LlamaForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T.
If your task is similar to the task the model of the checkpoint was trained on, you can already use LlamaForCausalLM for predictions without further training.
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/generation_config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/generation_config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/generation_config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/generation_config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/generation_config.json
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/generation_config.json
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2,
 "max_length": 2048,
 "pad_token_id": 0
}
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2,
 "max_length": 2048,
 "pad_token_id": 0
}
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2,
 "max_length": 2048,
 "pad_token_id": 0
}
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2,
 "max_length": 2048,
 "pad_token_id": 0
}
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2,
 "max_length": 2048,
 "pad_token_id": 0
}
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2,
 "max_length": 2048,
 "pad_token_id": 0
}
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/generation_config.json
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2,
 "max_length": 2048,
 "pad_token_id": 0
}
loading file tokenizer.model
loading file tokenizer.model
loading file tokenizer.model
loading file tokenizer.json
loading file tokenizer.json
loading file tokenizer.json
loading file added_tokens.json
loading file added_tokens.json
loading file added_tokens.json
loading file special_tokens_map.json
loading file special_tokens_map.json
loading file special_tokens_map.json
loading file tokenizer_config.json
loading file tokenizer_config.json
loading file tokenizer_config.json
loading file chat_template.jinja
loading file chat_template.jinja
loading file tokenizer.model
loading file chat_template.jinja
loading file tokenizer.model
loading file tokenizer.json
loading file tokenizer.model
loading file tokenizer.json
loading file added_tokens.json
loading file special_tokens_map.json
loading file added_tokens.json
loading file tokenizer_config.json
loading file tokenizer.json
loading file special_tokens_map.json
loading file chat_template.jinja
loading file added_tokens.json
loading file tokenizer_config.json
loading file special_tokens_map.json
loading file chat_template.jinja
loading file tokenizer_config.json
loading file chat_template.jinja
loading file tokenizer.model
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
You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 32001. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 32001. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 32001. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 32001. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 32001. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 32001. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 32001. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
All model checkpoint weights were used when initializing LlamaForCausalLM.
All the weights of LlamaForCausalLM were initialized from the model checkpoint at /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T.
If your task is similar to the task the model of the checkpoint was trained on, you can already use LlamaForCausalLM for predictions without further training.
loading configuration file /aifs4su/hansirui_1st/models/TinyLlama-1.1B-intermediate-step-480k-1T/generation_config.json
Generate config GenerationConfig {
 "bos_token_id": 1,
 "eos_token_id": 2,
 "max_length": 2048,
 "pad_token_id": 0
}
loading file tokenizer.model
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
You are resizing the embedding layer without providing a `pad_to_multiple_of` parameter. This means that the new embedding dimension will be 32001. This might induce some performance reduction as *Tensor Cores* will not be available. For more details about this, or help on choosing the correct value for resizing, refer to this guide: https://docs.nvidia.com/deeplearning/performance/dl-performance-matrix-multiplication/index.html#requirements-tc
The new embeddings will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new embeddings will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new embeddings will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new embeddings will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new embeddings will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new embeddings will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new embeddings will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new embeddings will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new lm_head weights will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new lm_head weights will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new lm_head weights will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new lm_head weights will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new lm_head weights will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new lm_head weights will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new lm_head weights will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
The new lm_head weights will be initialized from a multivariate normal distribution that has old embeddings' mean and covariance. As described in this article: https://nlp.stanford.edu/~johnhew/vocab-expansion.html. To disable this, use `mean_resizing=False`
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/jy-resist/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...
`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 0.19.11
wandb: Run data is saved locally in /aifs4su/hansirui_1st/jiayi/setting3-imdb/tinyllama-1T/tinyllama-1T-s3-Q1-1000/wandb/run-20250529_171624-c8dwjupa
wandb: Run `wandb offline` to turn off syncing.
wandb: Syncing run imdb-tinyllama-1T-s3-Q1-1000
wandb: ⭐️ View project at https://wandb.ai/xtom/Inverse_Alignment_IMDb
wandb: 🚀 View run at https://wandb.ai/xtom/Inverse_Alignment_IMDb/runs/c8dwjupa
 Training 1/1 epoch: 0%| | 0/125 [00:00<?, ?it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`.
 Training 1/1 epoch (loss 3.0580): 0%| | 0/125 [00:05<?, ?it/s] Training 1/1 epoch (loss 3.0580): 1%| | 1/125 [00:05<11:54, 5.76s/it] Training 1/1 epoch (loss 2.8824): 1%| | 1/125 [00:07<11:54, 5.76s/it] Training 1/1 epoch (loss 2.8824): 2%|▏ | 2/125 [00:07<07:07, 3.48s/it] Training 1/1 epoch (loss 3.0019): 2%|▏ | 2/125 [00:08<07:07, 3.48s/it] Training 1/1 epoch (loss 3.0019): 2%|▏ | 3/125 [00:08<04:11, 2.06s/it] Training 1/1 epoch (loss 3.1109): 2%|▏ | 3/125 [00:08<04:11, 2.06s/it] Training 1/1 epoch (loss 3.1109): 3%|▎ | 4/125 [00:08<02:49, 1.40s/it] Training 1/1 epoch (loss 2.8964): 3%|▎ | 4/125 [00:08<02:49, 1.40s/it] Training 1/1 epoch (loss 2.8964): 4%|▍ | 5/125 [00:08<02:01, 1.01s/it] Training 1/1 epoch (loss 3.0555): 4%|▍ | 5/125 [00:09<02:01, 1.01s/it] Training 1/1 epoch (loss 3.0555): 5%|▍ | 6/125 [00:09<01:32, 1.28it/s] Training 1/1 epoch (loss 2.8033): 5%|▍ | 6/125 [00:09<01:32, 1.28it/s] Training 1/1 epoch (loss 2.8033): 6%|▌ | 7/125 [00:09<01:13, 1.60it/s] Training 1/1 epoch (loss 2.9976): 6%|▌ | 7/125 [00:09<01:13, 1.60it/s] Training 1/1 epoch (loss 2.9976): 6%|▋ | 8/125 [00:09<01:06, 1.77it/s] Training 1/1 epoch (loss 3.2271): 6%|▋ | 8/125 [00:10<01:06, 1.77it/s] Training 1/1 epoch (loss 3.2271): 7%|▋ | 9/125 [00:10<01:02, 1.85it/s] Training 1/1 epoch (loss 3.0893): 7%|▋ | 9/125 [00:10<01:02, 1.85it/s] Training 1/1 epoch (loss 3.0893): 8%|▊ | 10/125 [00:10<00:53, 2.14it/s] Training 1/1 epoch (loss 2.8510): 8%|▊ | 10/125 [00:10<00:53, 2.14it/s] Training 1/1 epoch (loss 2.8510): 9%|▉ | 11/125 [00:10<00:48, 2.36it/s] Training 1/1 epoch (loss 2.9013): 9%|▉ | 11/125 [00:11<00:48, 2.36it/s] Training 1/1 epoch (loss 2.9013): 10%|▉ | 12/125 [00:11<00:43, 2.58it/s] Training 1/1 epoch (loss 3.0002): 10%|▉ | 12/125 [00:11<00:43, 2.58it/s] Training 1/1 epoch (loss 3.0002): 10%|█ | 13/125 [00:11<00:41, 2.71it/s] Training 1/1 epoch (loss 2.8963): 10%|█ | 13/125 [00:11<00:41, 2.71it/s] Training 1/1 epoch (loss 2.8963): 11%|█ | 14/125 [00:11<00:42, 2.63it/s] Training 1/1 epoch (loss 2.7818): 11%|█ | 14/125 [00:12<00:42, 2.63it/s] Training 1/1 epoch (loss 2.7818): 12%|█▏ | 15/125 [00:12<00:41, 2.65it/s] Training 1/1 epoch (loss 3.1563): 12%|█▏ | 15/125 [00:12<00:41, 2.65it/s] Training 1/1 epoch (loss 3.1563): 13%|█▎ | 16/125 [00:12<00:39, 2.74it/s] Training 1/1 epoch (loss 2.9860): 13%|█▎ | 16/125 [00:13<00:39, 2.74it/s] Training 1/1 epoch (loss 2.9860): 14%|█▎ | 17/125 [00:13<00:39, 2.75it/s] Training 1/1 epoch (loss 2.9196): 14%|█▎ | 17/125 [00:13<00:39, 2.75it/s] Training 1/1 epoch (loss 2.9196): 14%|█▍ | 18/125 [00:13<00:36, 2.90it/s] Training 1/1 epoch (loss 2.9610): 14%|█▍ | 18/125 [00:13<00:36, 2.90it/s] Training 1/1 epoch (loss 2.9610): 15%|█▌ | 19/125 [00:13<00:36, 2.94it/s] Training 1/1 epoch (loss 3.0547): 15%|█▌ | 19/125 [00:13<00:36, 2.94it/s] Training 1/1 epoch (loss 3.0547): 16%|█▌ | 20/125 [00:13<00:35, 2.93it/s] Training 1/1 epoch (loss 2.9434): 16%|█▌ | 20/125 [00:14<00:35, 2.93it/s] Training 1/1 epoch (loss 2.9434): 17%|█▋ | 21/125 [00:14<00:39, 2.62it/s] Training 1/1 epoch (loss 3.0104): 17%|█▋ | 21/125 [00:14<00:39, 2.62it/s] Training 1/1 epoch (loss 3.0104): 18%|█▊ | 22/125 [00:14<00:36, 2.79it/s] Training 1/1 epoch (loss 2.6261): 18%|█▊ | 22/125 [00:15<00:36, 2.79it/s] Training 1/1 epoch (loss 2.6261): 18%|█▊ | 23/125 [00:15<00:37, 2.75it/s] Training 1/1 epoch (loss 2.9375): 18%|█▊ | 23/125 [00:15<00:37, 2.75it/s] Training 1/1 epoch (loss 2.9375): 19%|█▉ | 24/125 [00:15<00:36, 2.79it/s] Training 1/1 epoch (loss 2.5391): 19%|█▉ | 24/125 [00:15<00:36, 2.79it/s] Training 1/1 epoch (loss 2.5391): 20%|██ | 25/125 [00:15<00:36, 2.76it/s] Training 1/1 epoch (loss 2.9048): 20%|██ | 25/125 [00:16<00:36, 2.76it/s] Training 1/1 epoch (loss 2.9048): 21%|██ | 26/125 [00:16<00:35, 2.80it/s] Training 1/1 epoch (loss 2.9064): 21%|██ | 26/125 [00:16<00:35, 2.80it/s] Training 1/1 epoch (loss 2.9064): 22%|██▏ | 27/125 [00:16<00:34, 2.85it/s] Training 1/1 epoch (loss 2.6606): 22%|██▏ | 27/125 [00:16<00:34, 2.85it/s] Training 1/1 epoch (loss 2.6606): 22%|██▏ | 28/125 [00:16<00:33, 2.94it/s] Training 1/1 epoch (loss 2.9262): 22%|██▏ | 28/125 [00:17<00:33, 2.94it/s] Training 1/1 epoch (loss 2.9262): 23%|██▎ | 29/125 [00:17<00:31, 3.05it/s] Training 1/1 epoch (loss 3.2682): 23%|██▎ | 29/125 [00:17<00:31, 3.05it/s] Training 1/1 epoch (loss 3.2682): 24%|██▍ | 30/125 [00:17<00:30, 3.14it/s] Training 1/1 epoch (loss 2.8882): 24%|██▍ | 30/125 [00:17<00:30, 3.14it/s] Training 1/1 epoch (loss 2.8882): 25%|██▍ | 31/125 [00:17<00:30, 3.08it/s] Training 1/1 epoch (loss 2.9049): 25%|██▍ | 31/125 [00:18<00:30, 3.08it/s] Training 1/1 epoch (loss 2.9049): 26%|██▌ | 32/125 [00:18<00:32, 2.90it/s] Training 1/1 epoch (loss 2.8842): 26%|██▌ | 32/125 [00:18<00:32, 2.90it/s] Training 1/1 epoch (loss 2.8842): 26%|██▋ | 33/125 [00:18<00:32, 2.80it/s] Training 1/1 epoch (loss 2.9728): 26%|██▋ | 33/125 [00:18<00:32, 2.80it/s] Training 1/1 epoch (loss 2.9728): 27%|██▋ | 34/125 [00:18<00:31, 2.85it/s] Training 1/1 epoch (loss 3.0014): 27%|██▋ | 34/125 [00:19<00:31, 2.85it/s] Training 1/1 epoch (loss 3.0014): 28%|██▊ | 35/125 [00:19<00:30, 2.94it/s] Training 1/1 epoch (loss 2.8972): 28%|██▊ | 35/125 [00:19<00:30, 2.94it/s] Training 1/1 epoch (loss 2.8972): 29%|██▉ | 36/125 [00:19<00:30, 2.94it/s] Training 1/1 epoch (loss 2.8982): 29%|██▉ | 36/125 [00:19<00:30, 2.94it/s] Training 1/1 epoch (loss 2.8982): 30%|██▉ | 37/125 [00:19<00:29, 2.98it/s] Training 1/1 epoch (loss 2.8840): 30%|██▉ | 37/125 [00:20<00:29, 2.98it/s] Training 1/1 epoch (loss 2.8840): 30%|███ | 38/125 [00:20<00:30, 2.84it/s] Training 1/1 epoch (loss 2.9791): 30%|███ | 38/125 [00:20<00:30, 2.84it/s] Training 1/1 epoch (loss 2.9791): 31%|███ | 39/125 [00:20<00:31, 2.70it/s] Training 1/1 epoch (loss 2.8236): 31%|███ | 39/125 [00:21<00:31, 2.70it/s] Training 1/1 epoch (loss 2.8236): 32%|███▏ | 40/125 [00:21<00:30, 2.77it/s] Training 1/1 epoch (loss 2.9390): 32%|███▏ | 40/125 [00:21<00:30, 2.77it/s] Training 1/1 epoch (loss 2.9390): 33%|███▎ | 41/125 [00:21<00:29, 2.85it/s] Training 1/1 epoch (loss 2.8774): 33%|███▎ | 41/125 [00:21<00:29, 2.85it/s] Training 1/1 epoch (loss 2.8774): 34%|███▎ | 42/125 [00:21<00:29, 2.86it/s] Training 1/1 epoch (loss 2.8762): 34%|███▎ | 42/125 [00:22<00:29, 2.86it/s] Training 1/1 epoch (loss 2.8762): 34%|███▍ | 43/125 [00:22<00:28, 2.85it/s] Training 1/1 epoch (loss 2.9673): 34%|███▍ | 43/125 [00:22<00:28, 2.85it/s] Training 1/1 epoch (loss 2.9673): 35%|███▌ | 44/125 [00:22<00:28, 2.89it/s] Training 1/1 epoch (loss 2.7471): 35%|███▌ | 44/125 [00:22<00:28, 2.89it/s] Training 1/1 epoch (loss 2.7471): 36%|███▌ | 45/125 [00:22<00:28, 2.86it/s] Training 1/1 epoch (loss 2.7835): 36%|███▌ | 45/125 [00:23<00:28, 2.86it/s] Training 1/1 epoch (loss 2.7835): 37%|███▋ | 46/125 [00:23<00:27, 2.90it/s] Training 1/1 epoch (loss 2.9850): 37%|███▋ | 46/125 [00:23<00:27, 2.90it/s] Training 1/1 epoch (loss 2.9850): 38%|███▊ | 47/125 [00:23<00:26, 2.95it/s] Training 1/1 epoch (loss 3.0446): 38%|███▊ | 47/125 [00:23<00:26, 2.95it/s] Training 1/1 epoch (loss 3.0446): 38%|███▊ | 48/125 [00:23<00:27, 2.82it/s] Training 1/1 epoch (loss 2.9438): 38%|███▊ | 48/125 [00:24<00:27, 2.82it/s] Training 1/1 epoch (loss 2.9438): 39%|███▉ | 49/125 [00:24<00:26, 2.82it/s] Training 1/1 epoch (loss 2.9206): 39%|███▉ | 49/125 [00:24<00:26, 2.82it/s] Training 1/1 epoch (loss 2.9206): 40%|████ | 50/125 [00:24<00:27, 2.74it/s] Training 1/1 epoch (loss 2.7088): 40%|████ | 50/125 [00:24<00:27, 2.74it/s] Training 1/1 epoch (loss 2.7088): 41%|████ | 51/125 [00:24<00:27, 2.70it/s] Training 1/1 epoch (loss 2.8545): 41%|████ | 51/125 [00:25<00:27, 2.70it/s] Training 1/1 epoch (loss 2.8545): 42%|████▏ | 52/125 [00:25<00:25, 2.87it/s] Training 1/1 epoch (loss 2.8379): 42%|████▏ | 52/125 [00:25<00:25, 2.87it/s] Training 1/1 epoch (loss 2.8379): 42%|████▏ | 53/125 [00:25<00:24, 2.95it/s] Training 1/1 epoch (loss 2.9153): 42%|████▏ | 53/125 [00:25<00:24, 2.95it/s] Training 1/1 epoch (loss 2.9153): 43%|████▎ | 54/125 [00:25<00:25, 2.82it/s] Training 1/1 epoch (loss 2.7590): 43%|████▎ | 54/125 [00:26<00:25, 2.82it/s] Training 1/1 epoch (loss 2.7590): 44%|████▍ | 55/125 [00:26<00:24, 2.84it/s] Training 1/1 epoch (loss 2.8010): 44%|████▍ | 55/125 [00:26<00:24, 2.84it/s] Training 1/1 epoch (loss 2.8010): 45%|████▍ | 56/125 [00:26<00:25, 2.74it/s] Training 1/1 epoch (loss 2.8938): 45%|████▍ | 56/125 [00:27<00:25, 2.74it/s] Training 1/1 epoch (loss 2.8938): 46%|████▌ | 57/125 [00:27<00:29, 2.31it/s] Training 1/1 epoch (loss 2.9581): 46%|████▌ | 57/125 [00:27<00:29, 2.31it/s] Training 1/1 epoch (loss 2.9581): 46%|████▋ | 58/125 [00:27<00:26, 2.53it/s] Training 1/1 epoch (loss 2.8380): 46%|████▋ | 58/125 [00:27<00:26, 2.53it/s] Training 1/1 epoch (loss 2.8380): 47%|████▋ | 59/125 [00:27<00:25, 2.54it/s] Training 1/1 epoch (loss 2.9019): 47%|████▋ | 59/125 [00:28<00:25, 2.54it/s] Training 1/1 epoch (loss 2.9019): 48%|████▊ | 60/125 [00:28<00:25, 2.59it/s] Training 1/1 epoch (loss 2.6482): 48%|████▊ | 60/125 [00:28<00:25, 2.59it/s] Training 1/1 epoch (loss 2.6482): 49%|████▉ | 61/125 [00:28<00:25, 2.55it/s] Training 1/1 epoch (loss 2.9539): 49%|████▉ | 61/125 [00:29<00:25, 2.55it/s] Training 1/1 epoch (loss 2.9539): 50%|████▉ | 62/125 [00:29<00:23, 2.70it/s] Training 1/1 epoch (loss 2.9238): 50%|████▉ | 62/125 [00:29<00:23, 2.70it/s] Training 1/1 epoch (loss 2.9238): 50%|█████ | 63/125 [00:29<00:22, 2.73it/s] Training 1/1 epoch (loss 2.8617): 50%|█████ | 63/125 [00:29<00:22, 2.73it/s] Training 1/1 epoch (loss 2.8617): 51%|█████ | 64/125 [00:29<00:22, 2.69it/s] Training 1/1 epoch (loss 2.9149): 51%|█████ | 64/125 [00:30<00:22, 2.69it/s] Training 1/1 epoch (loss 2.9149): 52%|█████▏ | 65/125 [00:30<00:24, 2.43it/s] Training 1/1 epoch (loss 2.8701): 52%|█████▏ | 65/125 [00:30<00:24, 2.43it/s] Training 1/1 epoch (loss 2.8701): 53%|█████▎ | 66/125 [00:30<00:23, 2.49it/s] Training 1/1 epoch (loss 2.8681): 53%|█████▎ | 66/125 [00:31<00:23, 2.49it/s] Training 1/1 epoch (loss 2.8681): 54%|█████▎ | 67/125 [00:31<00:21, 2.64it/s] Training 1/1 epoch (loss 2.8743): 54%|█████▎ | 67/125 [00:31<00:21, 2.64it/s] Training 1/1 epoch (loss 2.8743): 54%|█████▍ | 68/125 [00:31<00:20, 2.75it/s] Training 1/1 epoch (loss 2.9492): 54%|█████▍ | 68/125 [00:31<00:20, 2.75it/s] Training 1/1 epoch (loss 2.9492): 55%|█████▌ | 69/125 [00:31<00:19, 2.83it/s] Training 1/1 epoch (loss 2.7723): 55%|█████▌ | 69/125 [00:31<00:19, 2.83it/s] Training 1/1 epoch (loss 2.7723): 56%|█████▌ | 70/125 [00:31<00:18, 2.95it/s] Training 1/1 epoch (loss 2.9214): 56%|█████▌ | 70/125 [00:32<00:18, 2.95it/s] Training 1/1 epoch (loss 2.9214): 57%|█████▋ | 71/125 [00:32<00:18, 2.87it/s] Training 1/1 epoch (loss 2.8504): 57%|█████▋ | 71/125 [00:32<00:18, 2.87it/s] Training 1/1 epoch (loss 2.8504): 58%|█████▊ | 72/125 [00:32<00:18, 2.92it/s] Training 1/1 epoch (loss 2.8136): 58%|█████▊ | 72/125 [00:33<00:18, 2.92it/s] Training 1/1 epoch (loss 2.8136): 58%|█████▊ | 73/125 [00:33<00:17, 2.91it/s] Training 1/1 epoch (loss 2.6824): 58%|█████▊ | 73/125 [00:33<00:17, 2.91it/s] Training 1/1 epoch (loss 2.6824): 59%|█████▉ | 74/125 [00:33<00:16, 3.04it/s] Training 1/1 epoch (loss 2.8901): 59%|█████▉ | 74/125 [00:33<00:16, 3.04it/s] Training 1/1 epoch (loss 2.8901): 60%|██████ | 75/125 [00:33<00:16, 2.97it/s] Training 1/1 epoch (loss 2.7610): 60%|██████ | 75/125 [00:34<00:16, 2.97it/s] Training 1/1 epoch (loss 2.7610): 61%|██████ | 76/125 [00:34<00:16, 2.98it/s] Training 1/1 epoch (loss 2.7934): 61%|██████ | 76/125 [00:34<00:16, 2.98it/s] Training 1/1 epoch (loss 2.7934): 62%|██████▏ | 77/125 [00:34<00:16, 2.88it/s] Training 1/1 epoch (loss 2.7160): 62%|██████▏ | 77/125 [00:34<00:16, 2.88it/s] Training 1/1 epoch (loss 2.7160): 62%|██████▏ | 78/125 [00:34<00:16, 2.85it/s] Training 1/1 epoch (loss 2.8855): 62%|██████▏ | 78/125 [00:35<00:16, 2.85it/s] Training 1/1 epoch (loss 2.8855): 63%|██████▎ | 79/125 [00:35<00:15, 2.89it/s] Training 1/1 epoch (loss 2.8559): 63%|██████▎ | 79/125 [00:35<00:15, 2.89it/s] Training 1/1 epoch (loss 2.8559): 64%|██████▍ | 80/125 [00:35<00:15, 2.92it/s] Training 1/1 epoch (loss 2.9910): 64%|██████▍ | 80/125 [00:35<00:15, 2.92it/s] Training 1/1 epoch (loss 2.9910): 65%|██████▍ | 81/125 [00:35<00:15, 2.89it/s] Training 1/1 epoch (loss 2.7588): 65%|██████▍ | 81/125 [00:36<00:15, 2.89it/s] Training 1/1 epoch (loss 2.7588): 66%|██████▌ | 82/125 [00:36<00:15, 2.83it/s] Training 1/1 epoch (loss 3.0557): 66%|██████▌ | 82/125 [00:36<00:15, 2.83it/s] Training 1/1 epoch (loss 3.0557): 66%|██████▋ | 83/125 [00:36<00:15, 2.80it/s] Training 1/1 epoch (loss 2.7865): 66%|██████▋ | 83/125 [00:36<00:15, 2.80it/s] Training 1/1 epoch (loss 2.7865): 67%|██████▋ | 84/125 [00:36<00:14, 2.85it/s] Training 1/1 epoch (loss 2.9754): 67%|██████▋ | 84/125 [00:37<00:14, 2.85it/s] Training 1/1 epoch (loss 2.9754): 68%|██████▊ | 85/125 [00:37<00:13, 2.87it/s] Training 1/1 epoch (loss 2.8975): 68%|██████▊ | 85/125 [00:37<00:13, 2.87it/s] Training 1/1 epoch (loss 2.8975): 69%|██████▉ | 86/125 [00:37<00:12, 3.00it/s] Training 1/1 epoch (loss 2.9036): 69%|██████▉ | 86/125 [00:37<00:12, 3.00it/s] Training 1/1 epoch (loss 2.9036): 70%|██████▉ | 87/125 [00:37<00:12, 3.02it/s] Training 1/1 epoch (loss 2.7435): 70%|██████▉ | 87/125 [00:38<00:12, 3.02it/s] Training 1/1 epoch (loss 2.7435): 70%|███████ | 88/125 [00:38<00:13, 2.80it/s] Training 1/1 epoch (loss 2.7429): 70%|███████ | 88/125 [00:38<00:13, 2.80it/s] Training 1/1 epoch (loss 2.7429): 71%|███████ | 89/125 [00:38<00:13, 2.68it/s] Training 1/1 epoch (loss 2.8429): 71%|███████ | 89/125 [00:38<00:13, 2.68it/s] Training 1/1 epoch (loss 2.8429): 72%|███████▏ | 90/125 [00:38<00:12, 2.72it/s] Training 1/1 epoch (loss 3.0484): 72%|███████▏ | 90/125 [00:39<00:12, 2.72it/s] Training 1/1 epoch (loss 3.0484): 73%|███████▎ | 91/125 [00:39<00:11, 2.92it/s] Training 1/1 epoch (loss 2.5903): 73%|███████▎ | 91/125 [00:39<00:11, 2.92it/s] Training 1/1 epoch (loss 2.5903): 74%|███████▎ | 92/125 [00:39<00:11, 2.96it/s] Training 1/1 epoch (loss 2.8064): 74%|███████▎ | 92/125 [00:39<00:11, 2.96it/s] Training 1/1 epoch (loss 2.8064): 74%|███████▍ | 93/125 [00:39<00:10, 3.10it/s] Training 1/1 epoch (loss 2.7239): 74%|███████▍ | 93/125 [00:40<00:10, 3.10it/s] Training 1/1 epoch (loss 2.7239): 75%|███████▌ | 94/125 [00:40<00:10, 3.03it/s] Training 1/1 epoch (loss 2.7694): 75%|███████▌ | 94/125 [00:40<00:10, 3.03it/s] Training 1/1 epoch (loss 2.7694): 76%|███████▌ | 95/125 [00:40<00:10, 2.87it/s] Training 1/1 epoch (loss 2.9411): 76%|███████▌ | 95/125 [00:40<00:10, 2.87it/s] Training 1/1 epoch (loss 2.9411): 77%|███████▋ | 96/125 [00:40<00:09, 2.92it/s] Training 1/1 epoch (loss 2.7900): 77%|███████▋ | 96/125 [00:41<00:09, 2.92it/s] Training 1/1 epoch (loss 2.7900): 78%|███████▊ | 97/125 [00:41<00:09, 2.96it/s] Training 1/1 epoch (loss 3.0355): 78%|███████▊ | 97/125 [00:41<00:09, 2.96it/s] Training 1/1 epoch (loss 3.0355): 78%|███████▊ | 98/125 [00:41<00:09, 2.94it/s] Training 1/1 epoch (loss 2.8476): 78%|███████▊ | 98/125 [00:41<00:09, 2.94it/s] Training 1/1 epoch (loss 2.8476): 79%|███████▉ | 99/125 [00:41<00:08, 3.05it/s] Training 1/1 epoch (loss 2.7929): 79%|███████▉ | 99/125 [00:42<00:08, 3.05it/s] Training 1/1 epoch (loss 2.7929): 80%|████████ | 100/125 [00:42<00:08, 2.99it/s] Training 1/1 epoch (loss 2.9579): 80%|████████ | 100/125 [00:42<00:08, 2.99it/s] Training 1/1 epoch (loss 2.9579): 81%|████████ | 101/125 [00:42<00:08, 2.90it/s] Training 1/1 epoch (loss 2.6590): 81%|████████ | 101/125 [00:42<00:08, 2.90it/s] Training 1/1 epoch (loss 2.6590): 82%|████████▏ | 102/125 [00:42<00:07, 2.97it/s] Training 1/1 epoch (loss 2.8030): 82%|████████▏ | 102/125 [00:43<00:07, 2.97it/s] Training 1/1 epoch (loss 2.8030): 82%|████████▏ | 103/125 [00:43<00:07, 2.95it/s] Training 1/1 epoch (loss 2.7842): 82%|████████▏ | 103/125 [00:43<00:07, 2.95it/s] Training 1/1 epoch (loss 2.7842): 83%|████████▎ | 104/125 [00:43<00:07, 2.99it/s] Training 1/1 epoch (loss 2.9251): 83%|████████▎ | 104/125 [00:43<00:07, 2.99it/s] Training 1/1 epoch (loss 2.9251): 84%|████████▍ | 105/125 [00:43<00:06, 2.87it/s] Training 1/1 epoch (loss 2.8325): 84%|████████▍ | 105/125 [00:44<00:06, 2.87it/s] Training 1/1 epoch (loss 2.8325): 85%|████████▍ | 106/125 [00:44<00:07, 2.70it/s] Training 1/1 epoch (loss 3.0590): 85%|████████▍ | 106/125 [00:44<00:07, 2.70it/s] Training 1/1 epoch (loss 3.0590): 86%|████████▌ | 107/125 [00:44<00:06, 2.82it/s] Training 1/1 epoch (loss 2.8514): 86%|████████▌ | 107/125 [00:45<00:06, 2.82it/s] Training 1/1 epoch (loss 2.8514): 86%|████████▋ | 108/125 [00:45<00:06, 2.75it/s] Training 1/1 epoch (loss 2.7525): 86%|████████▋ | 108/125 [00:45<00:06, 2.75it/s] Training 1/1 epoch (loss 2.7525): 87%|████████▋ | 109/125 [00:45<00:05, 2.80it/s] Training 1/1 epoch (loss 2.7497): 87%|████████▋ | 109/125 [00:45<00:05, 2.80it/s] Training 1/1 epoch (loss 2.7497): 88%|████████▊ | 110/125 [00:45<00:05, 2.87it/s] Training 1/1 epoch (loss 2.8988): 88%|████████▊ | 110/125 [00:46<00:05, 2.87it/s] Training 1/1 epoch (loss 2.8988): 89%|████████▉ | 111/125 [00:46<00:04, 2.93it/s] Training 1/1 epoch (loss 2.8213): 89%|████████▉ | 111/125 [00:46<00:04, 2.93it/s] Training 1/1 epoch (loss 2.8213): 90%|████████▉ | 112/125 [00:46<00:04, 2.92it/s] Training 1/1 epoch (loss 2.7001): 90%|████████▉ | 112/125 [00:46<00:04, 2.92it/s] Training 1/1 epoch (loss 2.7001): 90%|█████████ | 113/125 [00:46<00:04, 2.80it/s] Training 1/1 epoch (loss 2.6929): 90%|█████████ | 113/125 [00:47<00:04, 2.80it/s] Training 1/1 epoch (loss 2.6929): 91%|█████████ | 114/125 [00:47<00:03, 2.89it/s] Training 1/1 epoch (loss 2.9290): 91%|█████████ | 114/125 [00:47<00:03, 2.89it/s] Training 1/1 epoch (loss 2.9290): 92%|█████████▏| 115/125 [00:47<00:03, 3.01it/s] Training 1/1 epoch (loss 2.9256): 92%|█████████▏| 115/125 [00:47<00:03, 3.01it/s] Training 1/1 epoch (loss 2.9256): 93%|█████████▎| 116/125 [00:47<00:02, 3.01it/s] Training 1/1 epoch (loss 2.8700): 93%|█████████▎| 116/125 [00:48<00:02, 3.01it/s] Training 1/1 epoch (loss 2.8700): 94%|█████████▎| 117/125 [00:48<00:02, 3.04it/s] Training 1/1 epoch (loss 2.7997): 94%|█████████▎| 117/125 [00:48<00:02, 3.04it/s] Training 1/1 epoch (loss 2.7997): 94%|█████████▍| 118/125 [00:48<00:02, 3.05it/s] Training 1/1 epoch (loss 2.7545): 94%|█████████▍| 118/125 [00:48<00:02, 3.05it/s] Training 1/1 epoch (loss 2.7545): 95%|█████████▌| 119/125 [00:48<00:01, 3.04it/s] Training 1/1 epoch (loss 2.7261): 95%|█████████▌| 119/125 [00:49<00:01, 3.04it/s] Training 1/1 epoch (loss 2.7261): 96%|█████████▌| 120/125 [00:49<00:01, 3.06it/s] Training 1/1 epoch (loss 2.7013): 96%|█████████▌| 120/125 [00:49<00:01, 3.06it/s] Training 1/1 epoch (loss 2.7013): 97%|█████████▋| 121/125 [00:49<00:01, 3.09it/s] Training 1/1 epoch (loss 2.8049): 97%|█████████▋| 121/125 [00:49<00:01, 3.09it/s] Training 1/1 epoch (loss 2.8049): 98%|█████████▊| 122/125 [00:49<00:01, 3.00it/s] Training 1/1 epoch (loss 2.8732): 98%|█████████▊| 122/125 [00:50<00:01, 3.00it/s] Training 1/1 epoch (loss 2.8732): 98%|█████████▊| 123/125 [00:50<00:00, 3.05it/s] Training 1/1 epoch (loss 2.7018): 98%|█████████▊| 123/125 [00:50<00:00, 3.05it/s] Training 1/1 epoch (loss 2.7018): 99%|█████████▉| 124/125 [00:50<00:00, 2.95it/s] Training 1/1 epoch (loss 3.0824): 99%|█████████▉| 124/125 [00:50<00:00, 2.95it/s] Training 1/1 epoch (loss 3.0824): 100%|██████████| 125/125 [00:50<00:00, 2.92it/s] Training 1/1 epoch (loss 3.0824): 100%|██████████| 125/125 [00:50<00:00, 2.46it/s]
tokenizer config file saved in /aifs4su/hansirui_1st/jiayi/setting3-imdb/tinyllama-1T/tinyllama-1T-s3-Q1-1000/tokenizer_config.json
Special tokens file saved in /aifs4su/hansirui_1st/jiayi/setting3-imdb/tinyllama-1T/tinyllama-1T-s3-Q1-1000/special_tokens_map.json
wandb: ERROR Problem finishing run
Exception ignored in atexit callback: <bound method rank_zero_only.<locals>.wrapper of <safe_rlhf.logger.Logger object at 0x1550cc193dd0>>
Traceback (most recent call last):
 File "/home/hansirui_1st/jiayi/resist/setting3/safe_rlhf/utils.py", line 212, in wrapper
 return func(*args, **kwargs)
 ^^^^^^^^^^^^^^^^^^^^^
 File "/home/hansirui_1st/jiayi/resist/setting3/safe_rlhf/logger.py", line 183, in close
 self.wandb.finish()
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 406, in wrapper
 return func(self, *args, **kwargs)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 503, in wrapper
 return func(self, *args, **kwargs)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 451, in wrapper
 return func(self, *args, **kwargs)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2309, in finish
 return self._finish(exit_code)
 ^^^^^^^^^^^^^^^^^^^^^^^
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 406, in wrapper
 return func(self, *args, **kwargs)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2337, in _finish
 self._atexit_cleanup(exit_code=exit_code)
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2550, in _atexit_cleanup
 self._on_finish()
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/wandb_run.py", line 2806, in _on_finish
 wait_with_progress(
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/mailbox/wait_with_progress.py", line 24, in wait_with_progress
 return wait_all_with_progress(
 ^^^^^^^^^^^^^^^^^^^^^^^
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/mailbox/wait_with_progress.py", line 87, in wait_all_with_progress
 return asyncio_compat.run(progress_loop_with_timeout)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/site-packages/wandb/sdk/lib/asyncio_compat.py", line 27, in run
 future = executor.submit(runner.run, fn)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 File "/aifs4su/hansirui_1st/miniconda3/envs/jy-resist/lib/python3.11/concurrent/futures/thread.py", line 169, in submit
 raise RuntimeError('cannot schedule new futures after '
RuntimeError: cannot schedule new futures after interpreter shutdown