| [INFO|2024-12-18 09:25:29] configuration_utils.py:679 >> loading configuration file config.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/config.json | |
| [INFO|2024-12-18 09:25:29] configuration_utils.py:746 >> Model config LlamaConfig { | |
| "_name_or_path": "meta-llama/Llama-3.2-1B", | |
| "architectures": [ | |
| "LlamaForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 128000, | |
| "eos_token_id": 128001, | |
| "head_dim": 64, | |
| "hidden_act": "silu", | |
| "hidden_size": 2048, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 8192, | |
| "max_position_embeddings": 131072, | |
| "mlp_bias": false, | |
| "model_type": "llama", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 16, | |
| "num_key_value_heads": 8, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": { | |
| "factor": 32.0, | |
| "high_freq_factor": 4.0, | |
| "low_freq_factor": 1.0, | |
| "original_max_position_embeddings": 8192, | |
| "rope_type": "llama3" | |
| }, | |
| "rope_theta": 500000.0, | |
| "tie_word_embeddings": true, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.46.1", | |
| "use_cache": true, | |
| "vocab_size": 128256 | |
| } | |
| [INFO|2024-12-18 09:25:30] parser.py:355 >> Process rank: 1, device: cuda:1, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16 | |
| [INFO|2024-12-18 09:25:29] tokenization_utils_base.py:2211 >> loading file tokenizer.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/tokenizer.json | |
| [INFO|2024-12-18 09:25:29] tokenization_utils_base.py:2211 >> loading file tokenizer.model from cache at None | |
| [INFO|2024-12-18 09:25:29] tokenization_utils_base.py:2211 >> loading file added_tokens.json from cache at None | |
| [INFO|2024-12-18 09:25:29] tokenization_utils_base.py:2211 >> loading file special_tokens_map.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/special_tokens_map.json | |
| [INFO|2024-12-18 09:25:29] tokenization_utils_base.py:2211 >> loading file tokenizer_config.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/tokenizer_config.json | |
| [INFO|2024-12-18 09:25:30] tokenization_utils_base.py:2475 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. | |
| [INFO|2024-12-18 09:25:30] parser.py:355 >> Process rank: 2, device: cuda:2, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16 | |
| [INFO|2024-12-18 09:25:30] parser.py:355 >> Process rank: 3, device: cuda:3, n_gpu: 1, distributed training: True, compute dtype: torch.bfloat16 | |
| [INFO|2024-12-18 09:25:31] configuration_utils.py:679 >> loading configuration file config.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/config.json | |
| [INFO|2024-12-18 09:25:31] configuration_utils.py:746 >> Model config LlamaConfig { | |
| "_name_or_path": "meta-llama/Llama-3.2-1B", | |
| "architectures": [ | |
| "LlamaForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 128000, | |
| "eos_token_id": 128001, | |
| "head_dim": 64, | |
| "hidden_act": "silu", | |
| "hidden_size": 2048, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 8192, | |
| "max_position_embeddings": 131072, | |
| "mlp_bias": false, | |
| "model_type": "llama", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 16, | |
| "num_key_value_heads": 8, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": { | |
| "factor": 32.0, | |
| "high_freq_factor": 4.0, | |
| "low_freq_factor": 1.0, | |
| "original_max_position_embeddings": 8192, | |
| "rope_type": "llama3" | |
| }, | |
| "rope_theta": 500000.0, | |
| "tie_word_embeddings": true, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.46.1", | |
| "use_cache": true, | |
| "vocab_size": 128256 | |
| } | |
| [INFO|2024-12-18 09:25:31] tokenization_utils_base.py:2211 >> loading file tokenizer.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/tokenizer.json | |
| [INFO|2024-12-18 09:25:31] tokenization_utils_base.py:2211 >> loading file tokenizer.model from cache at None | |
| [INFO|2024-12-18 09:25:31] tokenization_utils_base.py:2211 >> loading file added_tokens.json from cache at None | |
| [INFO|2024-12-18 09:25:31] tokenization_utils_base.py:2211 >> loading file special_tokens_map.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/special_tokens_map.json | |
| [INFO|2024-12-18 09:25:31] tokenization_utils_base.py:2211 >> loading file tokenizer_config.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/tokenizer_config.json | |
| [INFO|2024-12-18 09:25:31] tokenization_utils_base.py:2475 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. | |
| [INFO|2024-12-18 09:25:31] logging.py:157 >> Add pad token: <|end_of_text|> | |
| [INFO|2024-12-18 09:25:31] logging.py:157 >> Loading dataset Harish-as-harry/openledger... | |
| [INFO|2024-12-18 09:25:45] configuration_utils.py:679 >> loading configuration file config.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/config.json | |
| [INFO|2024-12-18 09:25:45] configuration_utils.py:746 >> Model config LlamaConfig { | |
| "_name_or_path": "meta-llama/Llama-3.2-1B", | |
| "architectures": [ | |
| "LlamaForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 128000, | |
| "eos_token_id": 128001, | |
| "head_dim": 64, | |
| "hidden_act": "silu", | |
| "hidden_size": 2048, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 8192, | |
| "max_position_embeddings": 131072, | |
| "mlp_bias": false, | |
| "model_type": "llama", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 16, | |
| "num_key_value_heads": 8, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": { | |
| "factor": 32.0, | |
| "high_freq_factor": 4.0, | |
| "low_freq_factor": 1.0, | |
| "original_max_position_embeddings": 8192, | |
| "rope_type": "llama3" | |
| }, | |
| "rope_theta": 500000.0, | |
| "tie_word_embeddings": true, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.46.1", | |
| "use_cache": true, | |
| "vocab_size": 128256 | |
| } | |
| [INFO|2024-12-18 09:25:45] modeling_utils.py:3937 >> loading weights file model.safetensors from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/model.safetensors | |
| [INFO|2024-12-18 09:25:45] modeling_utils.py:1670 >> Instantiating LlamaForCausalLM model under default dtype torch.bfloat16. | |
| [INFO|2024-12-18 09:25:45] configuration_utils.py:1096 >> Generate config GenerationConfig { | |
| "bos_token_id": 128000, | |
| "eos_token_id": 128001 | |
| } | |
| [INFO|2024-12-18 09:25:46] modeling_utils.py:4800 >> All model checkpoint weights were used when initializing LlamaForCausalLM. | |
| [INFO|2024-12-18 09:25:46] modeling_utils.py:4808 >> All the weights of LlamaForCausalLM were initialized from the model checkpoint at meta-llama/Llama-3.2-1B. | |
| 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. | |
| [INFO|2024-12-18 09:25:46] configuration_utils.py:1051 >> loading configuration file generation_config.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/generation_config.json | |
| [INFO|2024-12-18 09:25:46] configuration_utils.py:1096 >> Generate config GenerationConfig { | |
| "bos_token_id": 128000, | |
| "do_sample": true, | |
| "eos_token_id": 128001, | |
| "temperature": 0.6, | |
| "top_p": 0.9 | |
| } | |
| [INFO|2024-12-18 09:25:46] logging.py:157 >> Gradient checkpointing enabled. | |
| [INFO|2024-12-18 09:25:46] logging.py:157 >> Using torch SDPA for faster training and inference. | |
| [INFO|2024-12-18 09:25:46] logging.py:157 >> Upcasting trainable params to float32. | |
| [INFO|2024-12-18 09:25:46] logging.py:157 >> Fine-tuning method: LoRA | |
| [INFO|2024-12-18 09:25:46] logging.py:157 >> Found linear modules: q_proj,down_proj,o_proj,v_proj,up_proj,gate_proj,k_proj | |
| [INFO|2024-12-18 09:25:47] logging.py:157 >> trainable params: 11,272,192 || all params: 1,247,086,592 || trainable%: 0.9039 | |
| [INFO|2024-12-18 09:25:47] trainer.py:698 >> Using auto half precision backend | |
| [INFO|2024-12-18 09:25:47] trainer.py:2313 >> ***** Running training ***** | |
| [INFO|2024-12-18 09:25:47] trainer.py:2314 >> Num examples = 153 | |
| [INFO|2024-12-18 09:25:47] trainer.py:2315 >> Num Epochs = 30 | |
| [INFO|2024-12-18 09:25:47] trainer.py:2316 >> Instantaneous batch size per device = 2 | |
| [INFO|2024-12-18 09:25:47] trainer.py:2319 >> Total train batch size (w. parallel, distributed & accumulation) = 64 | |
| [INFO|2024-12-18 09:25:47] trainer.py:2320 >> Gradient Accumulation steps = 8 | |
| [INFO|2024-12-18 09:25:47] trainer.py:2321 >> Total optimization steps = 60 | |
| [INFO|2024-12-18 09:25:47] trainer.py:2322 >> Number of trainable parameters = 11,272,192 | |
| [INFO|2024-12-18 09:25:53] logging.py:157 >> {'loss': 3.3709, 'learning_rate': 1.9659e-04, 'epoch': 2.30} | |
| [INFO|2024-12-18 09:25:58] logging.py:157 >> {'loss': 2.7270, 'learning_rate': 1.8660e-04, 'epoch': 4.60} | |
| [INFO|2024-12-18 09:26:03] logging.py:157 >> {'loss': 2.6508, 'learning_rate': 1.7071e-04, 'epoch': 7.05} | |
| [INFO|2024-12-18 09:26:09] logging.py:157 >> {'loss': 1.9384, 'learning_rate': 1.5000e-04, 'epoch': 9.35} | |
| [INFO|2024-12-18 09:26:14] logging.py:157 >> {'loss': 1.3815, 'learning_rate': 1.2588e-04, 'epoch': 11.65} | |
| [INFO|2024-12-18 09:26:19] logging.py:157 >> {'loss': 1.2123, 'learning_rate': 1.0000e-04, 'epoch': 14.10} | |
| [INFO|2024-12-18 09:26:24] logging.py:157 >> {'loss': 0.7774, 'learning_rate': 7.4118e-05, 'epoch': 16.40} | |
| [INFO|2024-12-18 09:26:30] logging.py:157 >> {'loss': 0.5555, 'learning_rate': 5.0000e-05, 'epoch': 18.70} | |
| [INFO|2024-12-18 09:26:35] logging.py:157 >> {'loss': 0.4299, 'learning_rate': 2.9289e-05, 'epoch': 21.15} | |
| [INFO|2024-12-18 09:26:40] logging.py:157 >> {'loss': 0.2811, 'learning_rate': 1.3397e-05, 'epoch': 23.45} | |
| [INFO|2024-12-18 09:26:45] logging.py:157 >> {'loss': 0.2445, 'learning_rate': 3.4074e-06, 'epoch': 25.75} | |
| [INFO|2024-12-18 09:26:51] logging.py:157 >> {'loss': 0.2564, 'learning_rate': 0.0000e+00, 'epoch': 28.20} | |
| [INFO|2024-12-18 09:26:51] trainer.py:3801 >> Saving model checkpoint to saves/Llama-3.2-1B/lora/openledger-18-12-2024--try-01/checkpoint-60 | |
| [INFO|2024-12-18 09:26:51] configuration_utils.py:679 >> loading configuration file config.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/config.json | |
| [INFO|2024-12-18 09:26:51] configuration_utils.py:746 >> Model config LlamaConfig { | |
| "architectures": [ | |
| "LlamaForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 128000, | |
| "eos_token_id": 128001, | |
| "head_dim": 64, | |
| "hidden_act": "silu", | |
| "hidden_size": 2048, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 8192, | |
| "max_position_embeddings": 131072, | |
| "mlp_bias": false, | |
| "model_type": "llama", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 16, | |
| "num_key_value_heads": 8, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": { | |
| "factor": 32.0, | |
| "high_freq_factor": 4.0, | |
| "low_freq_factor": 1.0, | |
| "original_max_position_embeddings": 8192, | |
| "rope_type": "llama3" | |
| }, | |
| "rope_theta": 500000.0, | |
| "tie_word_embeddings": true, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.46.1", | |
| "use_cache": true, | |
| "vocab_size": 128256 | |
| } | |
| [INFO|2024-12-18 09:26:51] tokenization_utils_base.py:2646 >> tokenizer config file saved in saves/Llama-3.2-1B/lora/openledger-18-12-2024--try-01/checkpoint-60/tokenizer_config.json | |
| [INFO|2024-12-18 09:26:51] tokenization_utils_base.py:2655 >> Special tokens file saved in saves/Llama-3.2-1B/lora/openledger-18-12-2024--try-01/checkpoint-60/special_tokens_map.json | |
| [INFO|2024-12-18 09:26:52] trainer.py:2584 >> | |
| Training completed. Do not forget to share your model on huggingface.co/models =) | |
| [INFO|2024-12-18 09:26:52] trainer.py:3801 >> Saving model checkpoint to saves/Llama-3.2-1B/lora/openledger-18-12-2024--try-01 | |
| [INFO|2024-12-18 09:26:52] configuration_utils.py:679 >> loading configuration file config.json from cache at /home/ubuntu/.cache/huggingface/hub/models--meta-llama--Llama-3.2-1B/snapshots/4e20de362430cd3b72f300e6b0f18e50e7166e08/config.json | |
| [INFO|2024-12-18 09:26:52] configuration_utils.py:746 >> Model config LlamaConfig { | |
| "architectures": [ | |
| "LlamaForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 128000, | |
| "eos_token_id": 128001, | |
| "head_dim": 64, | |
| "hidden_act": "silu", | |
| "hidden_size": 2048, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 8192, | |
| "max_position_embeddings": 131072, | |
| "mlp_bias": false, | |
| "model_type": "llama", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 16, | |
| "num_key_value_heads": 8, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": { | |
| "factor": 32.0, | |
| "high_freq_factor": 4.0, | |
| "low_freq_factor": 1.0, | |
| "original_max_position_embeddings": 8192, | |
| "rope_type": "llama3" | |
| }, | |
| "rope_theta": 500000.0, | |
| "tie_word_embeddings": true, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.46.1", | |
| "use_cache": true, | |
| "vocab_size": 128256 | |
| } | |
| [INFO|2024-12-18 09:26:52] tokenization_utils_base.py:2646 >> tokenizer config file saved in saves/Llama-3.2-1B/lora/openledger-18-12-2024--try-01/tokenizer_config.json | |
| [INFO|2024-12-18 09:26:52] tokenization_utils_base.py:2655 >> Special tokens file saved in saves/Llama-3.2-1B/lora/openledger-18-12-2024--try-01/special_tokens_map.json | |
| [WARNING|2024-12-18 09:26:52] logging.py:162 >> No metric eval_loss to plot. | |
| [WARNING|2024-12-18 09:26:52] logging.py:162 >> No metric eval_accuracy to plot. | |
| [INFO|2024-12-18 09:26:52] trainer.py:4117 >> | |
| ***** Running Evaluation ***** | |
| [INFO|2024-12-18 09:26:52] trainer.py:4119 >> Num examples = 9 | |
| [INFO|2024-12-18 09:26:52] trainer.py:4122 >> Batch size = 2 | |
| [INFO|2024-12-18 09:26:52] modelcard.py:449 >> Dropping the following result as it does not have all the necessary fields: | |
| {'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}} | |