| [WARNING|2024-11-12 02:01:07] logging.py:162 >> We recommend enable `upcast_layernorm` in quantized training. | |
| [INFO|2024-11-12 02:01:07] parser.py:355 >> Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: False, compute dtype: torch.bfloat16 | |
| [INFO|2024-11-12 02:01:07] configuration_utils.py:733 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/config.json | |
| [INFO|2024-11-12 02:01:07] configuration_utils.py:800 >> Model config MistralConfig { | |
| "_name_or_path": "mistralai/Mistral-7B-Instruct-v0.3", | |
| "architectures": [ | |
| "MistralForCausalLM" | |
| ], | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "head_dim": 128, | |
| "hidden_act": "silu", | |
| "hidden_size": 4096, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 14336, | |
| "max_position_embeddings": 32768, | |
| "model_type": "mistral", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 32, | |
| "num_key_value_heads": 8, | |
| "rms_norm_eps": 1e-05, | |
| "rope_theta": 1000000.0, | |
| "sliding_window": null, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.44.2", | |
| "use_cache": true, | |
| "vocab_size": 32768 | |
| } | |
| [INFO|2024-11-12 02:01:08] tokenization_utils_base.py:2269 >> loading file tokenizer.model from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/tokenizer.model | |
| [INFO|2024-11-12 02:01:08] tokenization_utils_base.py:2269 >> loading file tokenizer.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/tokenizer.json | |
| [INFO|2024-11-12 02:01:08] tokenization_utils_base.py:2269 >> loading file added_tokens.json from cache at None | |
| [INFO|2024-11-12 02:01:08] tokenization_utils_base.py:2269 >> loading file special_tokens_map.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/special_tokens_map.json | |
| [INFO|2024-11-12 02:01:08] tokenization_utils_base.py:2269 >> loading file tokenizer_config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/tokenizer_config.json | |
| [INFO|2024-11-12 02:01:09] configuration_utils.py:733 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/config.json | |
| [INFO|2024-11-12 02:01:09] configuration_utils.py:800 >> Model config MistralConfig { | |
| "_name_or_path": "mistralai/Mistral-7B-Instruct-v0.3", | |
| "architectures": [ | |
| "MistralForCausalLM" | |
| ], | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "head_dim": 128, | |
| "hidden_act": "silu", | |
| "hidden_size": 4096, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 14336, | |
| "max_position_embeddings": 32768, | |
| "model_type": "mistral", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 32, | |
| "num_key_value_heads": 8, | |
| "rms_norm_eps": 1e-05, | |
| "rope_theta": 1000000.0, | |
| "sliding_window": null, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.44.2", | |
| "use_cache": true, | |
| "vocab_size": 32768 | |
| } | |
| [INFO|2024-11-12 02:01:09] tokenization_utils_base.py:2269 >> loading file tokenizer.model from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/tokenizer.model | |
| [INFO|2024-11-12 02:01:09] tokenization_utils_base.py:2269 >> loading file tokenizer.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/tokenizer.json | |
| [INFO|2024-11-12 02:01:09] tokenization_utils_base.py:2269 >> loading file added_tokens.json from cache at None | |
| [INFO|2024-11-12 02:01:09] tokenization_utils_base.py:2269 >> loading file special_tokens_map.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/special_tokens_map.json | |
| [INFO|2024-11-12 02:01:09] tokenization_utils_base.py:2269 >> loading file tokenizer_config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/tokenizer_config.json | |
| [INFO|2024-11-12 02:01:10] logging.py:157 >> Add pad token: </s> | |
| [INFO|2024-11-12 02:01:10] logging.py:157 >> Loading dataset treino_pt_rde.json... | |
| [INFO|2024-11-12 02:01:15] configuration_utils.py:733 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/config.json | |
| [INFO|2024-11-12 02:01:15] configuration_utils.py:800 >> Model config MistralConfig { | |
| "_name_or_path": "mistralai/Mistral-7B-Instruct-v0.3", | |
| "architectures": [ | |
| "MistralForCausalLM" | |
| ], | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "head_dim": 128, | |
| "hidden_act": "silu", | |
| "hidden_size": 4096, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 14336, | |
| "max_position_embeddings": 32768, | |
| "model_type": "mistral", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 32, | |
| "num_key_value_heads": 8, | |
| "rms_norm_eps": 1e-05, | |
| "rope_theta": 1000000.0, | |
| "sliding_window": null, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.44.2", | |
| "use_cache": true, | |
| "vocab_size": 32768 | |
| } | |
| [INFO|2024-11-12 02:01:15] logging.py:157 >> Quantizing model to 4 bit with bitsandbytes. | |
| [INFO|2024-11-12 02:01:15] modeling_utils.py:3678 >> loading weights file model.safetensors from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/model.safetensors.index.json | |
| [INFO|2024-11-12 02:07:01] modeling_utils.py:1606 >> Instantiating MistralForCausalLM model under default dtype torch.bfloat16. | |
| [INFO|2024-11-12 02:07:01] configuration_utils.py:1038 >> Generate config GenerationConfig { | |
| "bos_token_id": 1, | |
| "eos_token_id": 2 | |
| } | |
| [INFO|2024-11-12 02:08:04] modeling_utils.py:4507 >> All model checkpoint weights were used when initializing MistralForCausalLM. | |
| [INFO|2024-11-12 02:08:04] modeling_utils.py:4515 >> All the weights of MistralForCausalLM were initialized from the model checkpoint at mistralai/Mistral-7B-Instruct-v0.3. | |
| If your task is similar to the task the model of the checkpoint was trained on, you can already use MistralForCausalLM for predictions without further training. | |
| [INFO|2024-11-12 02:08:05] configuration_utils.py:993 >> loading configuration file generation_config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/generation_config.json | |
| [INFO|2024-11-12 02:08:05] configuration_utils.py:1038 >> Generate config GenerationConfig { | |
| "bos_token_id": 1, | |
| "eos_token_id": 2 | |
| } | |
| [INFO|2024-11-12 02:08:05] logging.py:157 >> Gradient checkpointing enabled. | |
| [INFO|2024-11-12 02:08:05] logging.py:157 >> Using torch SDPA for faster training and inference. | |
| [INFO|2024-11-12 02:08:05] logging.py:157 >> Upcasting trainable params to float32. | |
| [INFO|2024-11-12 02:08:05] logging.py:157 >> Fine-tuning method: LoRA | |
| [INFO|2024-11-12 02:08:05] logging.py:157 >> Found linear modules: o_proj,q_proj,v_proj,down_proj,gate_proj,k_proj,up_proj | |
| [INFO|2024-11-12 02:08:06] logging.py:157 >> trainable params: 20,971,520 || all params: 7,268,995,072 || trainable%: 0.2885 | |
| [INFO|2024-11-12 02:08:06] trainer.py:648 >> Using auto half precision backend | |
| [INFO|2024-11-12 02:08:06] trainer.py:2134 >> ***** Running training ***** | |
| [INFO|2024-11-12 02:08:06] trainer.py:2135 >> Num examples = 300 | |
| [INFO|2024-11-12 02:08:06] trainer.py:2136 >> Num Epochs = 3 | |
| [INFO|2024-11-12 02:08:06] trainer.py:2137 >> Instantaneous batch size per device = 2 | |
| [INFO|2024-11-12 02:08:06] trainer.py:2140 >> Total train batch size (w. parallel, distributed & accumulation) = 8 | |
| [INFO|2024-11-12 02:08:06] trainer.py:2141 >> Gradient Accumulation steps = 4 | |
| [INFO|2024-11-12 02:08:06] trainer.py:2142 >> Total optimization steps = 111 | |
| [INFO|2024-11-12 02:08:06] trainer.py:2143 >> Number of trainable parameters = 20,971,520 | |
| [INFO|2024-11-12 02:20:35] logging.py:157 >> {'loss': 0.4778, 'learning_rate': 4.9005e-05, 'epoch': 0.27, 'throughput': 44.35} | |
| [INFO|2024-11-12 02:32:47] logging.py:157 >> {'loss': 0.3431, 'learning_rate': 4.6101e-05, 'epoch': 0.53, 'throughput': 44.24} | |
| [INFO|2024-11-12 02:44:52] logging.py:157 >> {'loss': 0.3180, 'learning_rate': 4.1517e-05, 'epoch': 0.80, 'throughput': 44.19} | |
| [INFO|2024-11-12 02:56:44] logging.py:157 >> {'loss': 0.4028, 'learning_rate': 3.5619e-05, 'epoch': 1.07, 'throughput': 44.18} | |
| [INFO|2024-11-12 03:08:54] logging.py:157 >> {'loss': 0.1864, 'learning_rate': 2.8876e-05, 'epoch': 1.33, 'throughput': 44.20} | |
| [INFO|2024-11-12 03:21:09] logging.py:157 >> {'loss': 0.2180, 'learning_rate': 2.1825e-05, 'epoch': 1.60, 'throughput': 44.20} | |
| [INFO|2024-11-12 03:33:22] logging.py:157 >> {'loss': 0.2257, 'learning_rate': 1.5026e-05, 'epoch': 1.87, 'throughput': 44.18} | |
| [INFO|2024-11-12 03:45:39] logging.py:157 >> {'loss': 0.2099, 'learning_rate': 9.0208e-06, 'epoch': 2.13, 'throughput': 44.19} | |
| [INFO|2024-11-12 03:57:47] logging.py:157 >> {'loss': 0.1456, 'learning_rate': 4.2873e-06, 'epoch': 2.40, 'throughput': 44.18} | |
| [INFO|2024-11-12 04:10:11] logging.py:157 >> {'loss': 0.0777, 'learning_rate': 1.2018e-06, 'epoch': 2.67, 'throughput': 44.18} | |
| [INFO|2024-11-12 04:22:13] logging.py:157 >> {'loss': 0.0711, 'learning_rate': 1.0012e-08, 'epoch': 2.93, 'throughput': 44.18} | |
| [INFO|2024-11-12 04:23:24] trainer.py:3503 >> Saving model checkpoint to saves/Mistral-7B-Instruct-v0.3/lora/train_mistral/checkpoint-111 | |
| [INFO|2024-11-12 04:23:24] configuration_utils.py:733 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/config.json | |
| [INFO|2024-11-12 04:23:24] configuration_utils.py:800 >> Model config MistralConfig { | |
| "architectures": [ | |
| "MistralForCausalLM" | |
| ], | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "head_dim": 128, | |
| "hidden_act": "silu", | |
| "hidden_size": 4096, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 14336, | |
| "max_position_embeddings": 32768, | |
| "model_type": "mistral", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 32, | |
| "num_key_value_heads": 8, | |
| "rms_norm_eps": 1e-05, | |
| "rope_theta": 1000000.0, | |
| "sliding_window": null, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.44.2", | |
| "use_cache": true, | |
| "vocab_size": 32768 | |
| } | |
| [INFO|2024-11-12 04:23:25] tokenization_utils_base.py:2684 >> tokenizer config file saved in saves/Mistral-7B-Instruct-v0.3/lora/train_mistral/checkpoint-111/tokenizer_config.json | |
| [INFO|2024-11-12 04:23:25] tokenization_utils_base.py:2693 >> Special tokens file saved in saves/Mistral-7B-Instruct-v0.3/lora/train_mistral/checkpoint-111/special_tokens_map.json | |
| [INFO|2024-11-12 04:23:25] trainer.py:2394 >> | |
| Training completed. Do not forget to share your model on huggingface.co/models =) | |
| [INFO|2024-11-12 04:23:25] trainer.py:3503 >> Saving model checkpoint to saves/Mistral-7B-Instruct-v0.3/lora/train_mistral | |
| [INFO|2024-11-12 04:23:26] configuration_utils.py:733 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/e0bc86c23ce5aae1db576c8cca6f06f1f73af2db/config.json | |
| [INFO|2024-11-12 04:23:26] configuration_utils.py:800 >> Model config MistralConfig { | |
| "architectures": [ | |
| "MistralForCausalLM" | |
| ], | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "head_dim": 128, | |
| "hidden_act": "silu", | |
| "hidden_size": 4096, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 14336, | |
| "max_position_embeddings": 32768, | |
| "model_type": "mistral", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 32, | |
| "num_key_value_heads": 8, | |
| "rms_norm_eps": 1e-05, | |
| "rope_theta": 1000000.0, | |
| "sliding_window": null, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.44.2", | |
| "use_cache": true, | |
| "vocab_size": 32768 | |
| } | |
| [INFO|2024-11-12 04:23:26] tokenization_utils_base.py:2684 >> tokenizer config file saved in saves/Mistral-7B-Instruct-v0.3/lora/train_mistral/tokenizer_config.json | |
| [INFO|2024-11-12 04:23:26] tokenization_utils_base.py:2693 >> Special tokens file saved in saves/Mistral-7B-Instruct-v0.3/lora/train_mistral/special_tokens_map.json | |
| [WARNING|2024-11-12 04:23:26] logging.py:162 >> No metric eval_loss to plot. | |
| [WARNING|2024-11-12 04:23:26] logging.py:162 >> No metric eval_accuracy to plot. | |
| [INFO|2024-11-12 04:23:26] modelcard.py:449 >> Dropping the following result as it does not have all the necessary fields: | |
| {'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}} | |