05/05/2024 11:09:07 - INFO - transformers.tokenization_utils_base - loading file tokenizer.model from cache at /root/.cache/huggingface/hub/models--TinyLlama--TinyLlama-1.1B-Chat-v1.0/snapshots/fe8a4ea1ffedaf415f4da2f062534de366a451e6/tokenizer.model 05/05/2024 11:09:07 - INFO - transformers.tokenization_utils_base - loading file tokenizer.json from cache at /root/.cache/huggingface/hub/models--TinyLlama--TinyLlama-1.1B-Chat-v1.0/snapshots/fe8a4ea1ffedaf415f4da2f062534de366a451e6/tokenizer.json 05/05/2024 11:09:07 - INFO - transformers.tokenization_utils_base - loading file added_tokens.json from cache at None 05/05/2024 11:09:07 - INFO - transformers.tokenization_utils_base - loading file special_tokens_map.json from cache at /root/.cache/huggingface/hub/models--TinyLlama--TinyLlama-1.1B-Chat-v1.0/snapshots/fe8a4ea1ffedaf415f4da2f062534de366a451e6/special_tokens_map.json 05/05/2024 11:09:07 - INFO - transformers.tokenization_utils_base - loading file tokenizer_config.json from cache at /root/.cache/huggingface/hub/models--TinyLlama--TinyLlama-1.1B-Chat-v1.0/snapshots/fe8a4ea1ffedaf415f4da2f062534de366a451e6/tokenizer_config.json 05/05/2024 11:09:07 - INFO - llmtuner.data.loader - Loading dataset Reseacrhed_data.json... 05/05/2024 11:09:07 - WARNING - llmtuner.data.utils - Checksum failed: missing SHA-1 hash value in dataset_info.json. 05/05/2024 11:09:09 - WARNING - transformers.tokenization_utils_base - Token indices sequence length is longer than the specified maximum sequence length for this model (2387 > 2048). Running this sequence through the model will result in indexing errors 05/05/2024 11:09:09 - INFO - transformers.configuration_utils - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--TinyLlama--TinyLlama-1.1B-Chat-v1.0/snapshots/fe8a4ea1ffedaf415f4da2f062534de366a451e6/config.json 05/05/2024 11:09:09 - INFO - transformers.configuration_utils - Model config LlamaConfig { "_name_or_path": "TinyLlama/TinyLlama-1.1B-Chat-v1.0", "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "intermediate_size": 5632, "max_position_embeddings": 2048, "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": "bfloat16", "transformers_version": "4.40.1", "use_cache": true, "vocab_size": 32000 } 05/05/2024 11:09:09 - INFO - llmtuner.model.utils.quantization - Quantizing model to 8 bit. 05/05/2024 11:09:30 - INFO - transformers.modeling_utils - loading weights file model.safetensors from cache at /root/.cache/huggingface/hub/models--TinyLlama--TinyLlama-1.1B-Chat-v1.0/snapshots/fe8a4ea1ffedaf415f4da2f062534de366a451e6/model.safetensors 05/05/2024 11:09:30 - INFO - transformers.modeling_utils - Instantiating LlamaForCausalLM model under default dtype torch.float16. 05/05/2024 11:09:30 - INFO - transformers.generation.configuration_utils - Generate config GenerationConfig { "bos_token_id": 1, "eos_token_id": 2 } 05/05/2024 11:09:34 - INFO - transformers.modeling_utils - All model checkpoint weights were used when initializing LlamaForCausalLM. 05/05/2024 11:09:34 - INFO - transformers.modeling_utils - All the weights of LlamaForCausalLM were initialized from the model checkpoint at TinyLlama/TinyLlama-1.1B-Chat-v1.0. 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. 05/05/2024 11:09:35 - INFO - transformers.generation.configuration_utils - loading configuration file generation_config.json from cache at /root/.cache/huggingface/hub/models--TinyLlama--TinyLlama-1.1B-Chat-v1.0/snapshots/fe8a4ea1ffedaf415f4da2f062534de366a451e6/generation_config.json 05/05/2024 11:09:35 - INFO - transformers.generation.configuration_utils - Generate config GenerationConfig { "bos_token_id": 1, "eos_token_id": 2, "max_length": 2048, "pad_token_id": 0 } 05/05/2024 11:09:35 - INFO - llmtuner.model.utils.checkpointing - Gradient checkpointing enabled. 05/05/2024 11:09:35 - INFO - llmtuner.model.utils.attention - Using torch SDPA for faster training and inference. 05/05/2024 11:09:35 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA 05/05/2024 11:09:35 - INFO - llmtuner.model.loader - trainable params: 1126400 || all params: 1101174784 || trainable%: 0.1023 05/05/2024 11:09:35 - INFO - transformers.trainer - Using auto half precision backend 05/05/2024 11:09:35 - INFO - transformers.trainer - ***** Running training ***** 05/05/2024 11:09:35 - INFO - transformers.trainer - Num examples = 855 05/05/2024 11:09:35 - INFO - transformers.trainer - Num Epochs = 2 05/05/2024 11:09:35 - INFO - transformers.trainer - Instantaneous batch size per device = 2 05/05/2024 11:09:35 - INFO - transformers.trainer - Total train batch size (w. parallel, distributed & accumulation) = 16 05/05/2024 11:09:35 - INFO - transformers.trainer - Gradient Accumulation steps = 8 05/05/2024 11:09:35 - INFO - transformers.trainer - Total optimization steps = 106 05/05/2024 11:09:35 - INFO - transformers.trainer - Number of trainable parameters = 1,126,400 05/05/2024 11:10:23 - INFO - llmtuner.extras.callbacks - {'loss': 2.9430, 'learning_rate': 1.9890e-05, 'epoch': 0.09} 05/05/2024 11:11:11 - INFO - llmtuner.extras.callbacks - {'loss': 3.0203, 'learning_rate': 1.9564e-05, 'epoch': 0.19} 05/05/2024 11:11:59 - INFO - llmtuner.extras.callbacks - {'loss': 2.8306, 'learning_rate': 1.9028e-05, 'epoch': 0.28} 05/05/2024 11:12:48 - INFO - llmtuner.extras.callbacks - {'loss': 2.9983, 'learning_rate': 1.8294e-05, 'epoch': 0.37} 05/05/2024 11:13:37 - INFO - llmtuner.extras.callbacks - {'loss': 3.0059, 'learning_rate': 1.7378e-05, 'epoch': 0.47} 05/05/2024 11:14:26 - INFO - llmtuner.extras.callbacks - {'loss': 2.8450, 'learning_rate': 1.6301e-05, 'epoch': 0.56} 05/05/2024 11:15:16 - INFO - llmtuner.extras.callbacks - {'loss': 2.9230, 'learning_rate': 1.5085e-05, 'epoch': 0.65} 05/05/2024 11:16:05 - INFO - llmtuner.extras.callbacks - {'loss': 2.9725, 'learning_rate': 1.3758e-05, 'epoch': 0.75} 05/05/2024 11:16:55 - INFO - llmtuner.extras.callbacks - {'loss': 2.9861, 'learning_rate': 1.2349e-05, 'epoch': 0.84} 05/05/2024 11:17:45 - INFO - llmtuner.extras.callbacks - {'loss': 2.9897, 'learning_rate': 1.0888e-05, 'epoch': 0.93} 05/05/2024 11:18:35 - INFO - llmtuner.extras.callbacks - {'loss': 2.9677, 'learning_rate': 9.4076e-06, 'epoch': 1.03} 05/05/2024 11:19:24 - INFO - llmtuner.extras.callbacks - {'loss': 2.9620, 'learning_rate': 7.9402e-06, 'epoch': 1.12} 05/05/2024 11:20:14 - INFO - llmtuner.extras.callbacks - {'loss': 2.9335, 'learning_rate': 6.5180e-06, 'epoch': 1.21} 05/05/2024 11:21:04 - INFO - llmtuner.extras.callbacks - {'loss': 2.9682, 'learning_rate': 5.1721e-06, 'epoch': 1.31} 05/05/2024 11:21:54 - INFO - llmtuner.extras.callbacks - {'loss': 2.9475, 'learning_rate': 3.9320e-06, 'epoch': 1.40} 05/05/2024 11:22:43 - INFO - llmtuner.extras.callbacks - {'loss': 2.9373, 'learning_rate': 2.8249e-06, 'epoch': 1.50} 05/05/2024 11:23:34 - INFO - llmtuner.extras.callbacks - {'loss': 2.9115, 'learning_rate': 1.8751e-06, 'epoch': 1.59} 05/05/2024 11:24:24 - INFO - llmtuner.extras.callbacks - {'loss': 2.9237, 'learning_rate': 1.1034e-06, 'epoch': 1.68} 05/05/2024 11:25:13 - INFO - llmtuner.extras.callbacks - {'loss': 2.8901, 'learning_rate': 5.2674e-07, 'epoch': 1.78} 05/05/2024 11:26:03 - INFO - llmtuner.extras.callbacks - {'loss': 2.8905, 'learning_rate': 1.5769e-07, 'epoch': 1.87} 05/05/2024 11:26:03 - INFO - transformers.trainer - Saving model checkpoint to saves/LLaMA-tiny/lora/train_2024-05-05-10-57-50/checkpoint-100 05/05/2024 11:26:03 - INFO - transformers.configuration_utils - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--TinyLlama--TinyLlama-1.1B-Chat-v1.0/snapshots/fe8a4ea1ffedaf415f4da2f062534de366a451e6/config.json 05/05/2024 11:26:03 - INFO - transformers.configuration_utils - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "intermediate_size": 5632, "max_position_embeddings": 2048, "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": "bfloat16", "transformers_version": "4.40.1", "use_cache": true, "vocab_size": 32000 } 05/05/2024 11:26:03 - INFO - transformers.tokenization_utils_base - tokenizer config file saved in saves/LLaMA-tiny/lora/train_2024-05-05-10-57-50/checkpoint-100/tokenizer_config.json 05/05/2024 11:26:03 - INFO - transformers.tokenization_utils_base - Special tokens file saved in saves/LLaMA-tiny/lora/train_2024-05-05-10-57-50/checkpoint-100/special_tokens_map.json 05/05/2024 11:26:53 - INFO - llmtuner.extras.callbacks - {'loss': 2.9548, 'learning_rate': 4.3916e-09, 'epoch': 1.96} 05/05/2024 11:27:03 - INFO - transformers.trainer - Training completed. Do not forget to share your model on huggingface.co/models =) 05/05/2024 11:27:03 - INFO - transformers.trainer - Saving model checkpoint to saves/LLaMA-tiny/lora/train_2024-05-05-10-57-50 05/05/2024 11:27:03 - INFO - transformers.configuration_utils - loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--TinyLlama--TinyLlama-1.1B-Chat-v1.0/snapshots/fe8a4ea1ffedaf415f4da2f062534de366a451e6/config.json 05/05/2024 11:27:03 - INFO - transformers.configuration_utils - Model config LlamaConfig { "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "intermediate_size": 5632, "max_position_embeddings": 2048, "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": "bfloat16", "transformers_version": "4.40.1", "use_cache": true, "vocab_size": 32000 } 05/05/2024 11:27:03 - INFO - transformers.tokenization_utils_base - tokenizer config file saved in saves/LLaMA-tiny/lora/train_2024-05-05-10-57-50/tokenizer_config.json 05/05/2024 11:27:03 - INFO - transformers.tokenization_utils_base - Special tokens file saved in saves/LLaMA-tiny/lora/train_2024-05-05-10-57-50/special_tokens_map.json 05/05/2024 11:27:03 - INFO - transformers.modelcard - Dropping the following result as it does not have all the necessary fields: {'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}