File size: 9,398 Bytes
33e0752 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 |
05/09/2024 19:59:23 - INFO - transformers.tokenization_utils_base - loading file vocab.json from cache at /home/.cache/huggingface/hub/models--microsoft--phi-1_5/snapshots/675aa382d814580b22651a30acb1a585d7c25963/vocab.json
05/09/2024 19:59:23 - INFO - transformers.tokenization_utils_base - loading file merges.txt from cache at /home/.cache/huggingface/hub/models--microsoft--phi-1_5/snapshots/675aa382d814580b22651a30acb1a585d7c25963/merges.txt
05/09/2024 19:59:23 - INFO - transformers.tokenization_utils_base - loading file tokenizer.json from cache at /home/.cache/huggingface/hub/models--microsoft--phi-1_5/snapshots/675aa382d814580b22651a30acb1a585d7c25963/tokenizer.json
05/09/2024 19:59:23 - INFO - transformers.tokenization_utils_base - loading file added_tokens.json from cache at /home/.cache/huggingface/hub/models--microsoft--phi-1_5/snapshots/675aa382d814580b22651a30acb1a585d7c25963/added_tokens.json
05/09/2024 19:59:23 - INFO - transformers.tokenization_utils_base - loading file special_tokens_map.json from cache at /home/.cache/huggingface/hub/models--microsoft--phi-1_5/snapshots/675aa382d814580b22651a30acb1a585d7c25963/special_tokens_map.json
05/09/2024 19:59:23 - INFO - transformers.tokenization_utils_base - loading file tokenizer_config.json from cache at /home/.cache/huggingface/hub/models--microsoft--phi-1_5/snapshots/675aa382d814580b22651a30acb1a585d7c25963/tokenizer_config.json
05/09/2024 19:59:23 - INFO - llmtuner.data.template - Add pad token: <|endoftext|>
05/09/2024 19:59:23 - INFO - llmtuner.data.loader - Loading dataset alpaca_data_en_52k.json...
05/09/2024 19:59:24 - INFO - transformers.configuration_utils - loading configuration file config.json from cache at /home/.cache/huggingface/hub/models--microsoft--phi-1_5/snapshots/675aa382d814580b22651a30acb1a585d7c25963/config.json
05/09/2024 19:59:24 - INFO - transformers.configuration_utils - Model config PhiConfig {
"_name_or_path": "microsoft/phi-1_5",
"architectures": [
"PhiForCausalLM"
],
"attention_dropout": 0.0,
"bos_token_id": null,
"embd_pdrop": 0.0,
"eos_token_id": null,
"hidden_act": "gelu_new",
"hidden_size": 2048,
"initializer_range": 0.02,
"intermediate_size": 8192,
"layer_norm_eps": 1e-05,
"max_position_embeddings": 2048,
"model_type": "phi",
"num_attention_heads": 32,
"num_hidden_layers": 24,
"num_key_value_heads": 32,
"partial_rotary_factor": 0.5,
"qk_layernorm": false,
"resid_pdrop": 0.0,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": false,
"torch_dtype": "float16",
"transformers_version": "4.40.1",
"use_cache": true,
"vocab_size": 51200
}
05/09/2024 20:08:34 - INFO - transformers.modeling_utils - loading weights file model.safetensors from cache at /home/.cache/huggingface/hub/models--microsoft--phi-1_5/snapshots/675aa382d814580b22651a30acb1a585d7c25963/model.safetensors
05/09/2024 20:08:34 - INFO - transformers.modeling_utils - Instantiating PhiForCausalLM model under default dtype torch.float16.
05/09/2024 20:08:34 - INFO - transformers.generation.configuration_utils - Generate config GenerationConfig {}
05/09/2024 20:08:35 - INFO - transformers.modeling_utils - All model checkpoint weights were used when initializing PhiForCausalLM.
05/09/2024 20:08:35 - INFO - transformers.modeling_utils - All the weights of PhiForCausalLM were initialized from the model checkpoint at microsoft/phi-1_5.
If your task is similar to the task the model of the checkpoint was trained on, you can already use PhiForCausalLM for predictions without further training.
05/09/2024 20:08:36 - INFO - transformers.generation.configuration_utils - loading configuration file generation_config.json from cache at /home/.cache/huggingface/hub/models--microsoft--phi-1_5/snapshots/675aa382d814580b22651a30acb1a585d7c25963/generation_config.json
05/09/2024 20:08:36 - INFO - transformers.generation.configuration_utils - Generate config GenerationConfig {}
05/09/2024 20:08:36 - INFO - llmtuner.model.utils.checkpointing - Gradient checkpointing enabled.
05/09/2024 20:08:36 - INFO - llmtuner.model.utils.attention - Using torch SDPA for faster training and inference.
05/09/2024 20:08:36 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
05/09/2024 20:08:36 - INFO - llmtuner.model.loader - trainable params: 1572864 || all params: 1419843584 || trainable%: 0.1108
05/09/2024 20:08:36 - INFO - transformers.trainer - Using auto half precision backend
05/09/2024 20:08:36 - INFO - transformers.trainer - ***** Running training *****
05/09/2024 20:08:36 - INFO - transformers.trainer - Num examples = 500
05/09/2024 20:08:36 - INFO - transformers.trainer - Num Epochs = 3
05/09/2024 20:08:36 - INFO - transformers.trainer - Instantaneous batch size per device = 2
05/09/2024 20:08:36 - INFO - transformers.trainer - Total train batch size (w. parallel, distributed & accumulation) = 16
05/09/2024 20:08:36 - INFO - transformers.trainer - Gradient Accumulation steps = 8
05/09/2024 20:08:36 - INFO - transformers.trainer - Total optimization steps = 93
05/09/2024 20:08:36 - INFO - transformers.trainer - Number of trainable parameters = 1,572,864
05/09/2024 20:08:47 - INFO - llmtuner.extras.callbacks - {'loss': 2.1283, 'learning_rate': 1.9858e-04, 'epoch': 0.16}
05/09/2024 20:08:57 - INFO - llmtuner.extras.callbacks - {'loss': 1.8000, 'learning_rate': 1.9435e-04, 'epoch': 0.32}
05/09/2024 20:09:07 - INFO - llmtuner.extras.callbacks - {'loss': 1.5595, 'learning_rate': 1.8743e-04, 'epoch': 0.48}
05/09/2024 20:09:18 - INFO - llmtuner.extras.callbacks - {'loss': 1.6097, 'learning_rate': 1.7803e-04, 'epoch': 0.64}
05/09/2024 20:09:29 - INFO - llmtuner.extras.callbacks - {'loss': 1.2942, 'learning_rate': 1.6641e-04, 'epoch': 0.80}
05/09/2024 20:09:38 - INFO - llmtuner.extras.callbacks - {'loss': 1.4620, 'learning_rate': 1.5290e-04, 'epoch': 0.96}
05/09/2024 20:09:48 - INFO - llmtuner.extras.callbacks - {'loss': 1.4432, 'learning_rate': 1.3788e-04, 'epoch': 1.12}
05/09/2024 20:09:59 - INFO - llmtuner.extras.callbacks - {'loss': 1.3561, 'learning_rate': 1.2178e-04, 'epoch': 1.28}
05/09/2024 20:10:09 - INFO - llmtuner.extras.callbacks - {'loss': 1.3299, 'learning_rate': 1.0506e-04, 'epoch': 1.44}
05/09/2024 20:10:19 - INFO - llmtuner.extras.callbacks - {'loss': 1.2208, 'learning_rate': 8.8204e-05, 'epoch': 1.60}
05/09/2024 20:10:30 - INFO - llmtuner.extras.callbacks - {'loss': 1.4574, 'learning_rate': 7.1679e-05, 'epoch': 1.76}
05/09/2024 20:10:41 - INFO - llmtuner.extras.callbacks - {'loss': 1.1423, 'learning_rate': 5.5961e-05, 'epoch': 1.92}
05/09/2024 20:10:50 - INFO - llmtuner.extras.callbacks - {'loss': 1.3134, 'learning_rate': 4.1495e-05, 'epoch': 2.08}
05/09/2024 20:11:01 - INFO - llmtuner.extras.callbacks - {'loss': 1.3384, 'learning_rate': 2.8695e-05, 'epoch': 2.24}
05/09/2024 20:11:12 - INFO - llmtuner.extras.callbacks - {'loss': 1.3873, 'learning_rate': 1.7924e-05, 'epoch': 2.40}
05/09/2024 20:11:22 - INFO - llmtuner.extras.callbacks - {'loss': 1.2214, 'learning_rate': 9.4885e-06, 'epoch': 2.56}
05/09/2024 20:11:32 - INFO - llmtuner.extras.callbacks - {'loss': 1.3198, 'learning_rate': 3.6294e-06, 'epoch': 2.72}
05/09/2024 20:11:42 - INFO - llmtuner.extras.callbacks - {'loss': 1.3832, 'learning_rate': 5.1307e-07, 'epoch': 2.88}
05/09/2024 20:11:49 - INFO - transformers.trainer -
Training completed. Do not forget to share your model on huggingface.co/models =)
05/09/2024 20:11:49 - INFO - transformers.trainer - Saving model checkpoint to saves/Phi-1.5-1.3B/lora/train_2024-05-09-19-57-19
05/09/2024 20:11:49 - INFO - transformers.configuration_utils - loading configuration file config.json from cache at /home/.cache/huggingface/hub/models--microsoft--phi-1_5/snapshots/675aa382d814580b22651a30acb1a585d7c25963/config.json
05/09/2024 20:11:49 - INFO - transformers.configuration_utils - Model config PhiConfig {
"_name_or_path": "microsoft/phi-1_5",
"architectures": [
"PhiForCausalLM"
],
"attention_dropout": 0.0,
"bos_token_id": null,
"embd_pdrop": 0.0,
"eos_token_id": null,
"hidden_act": "gelu_new",
"hidden_size": 2048,
"initializer_range": 0.02,
"intermediate_size": 8192,
"layer_norm_eps": 1e-05,
"max_position_embeddings": 2048,
"model_type": "phi",
"num_attention_heads": 32,
"num_hidden_layers": 24,
"num_key_value_heads": 32,
"partial_rotary_factor": 0.5,
"qk_layernorm": false,
"resid_pdrop": 0.0,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": false,
"torch_dtype": "float16",
"transformers_version": "4.40.1",
"use_cache": true,
"vocab_size": 51200
}
05/09/2024 20:11:49 - INFO - transformers.tokenization_utils_base - tokenizer config file saved in saves/Phi-1.5-1.3B/lora/train_2024-05-09-19-57-19/tokenizer_config.json
05/09/2024 20:11:49 - INFO - transformers.tokenization_utils_base - Special tokens file saved in saves/Phi-1.5-1.3B/lora/train_2024-05-09-19-57-19/special_tokens_map.json
05/09/2024 20:11:49 - INFO - transformers.modelcard - Dropping the following result as it does not have all the necessary fields:
{'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}
|