File size: 17,194 Bytes
dba0ad4 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 |
[INFO|parser.py:325] 2024-09-19 12:17:17,636 >> Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: False, compute dtype: torch.bfloat16
[INFO|tokenization_utils_base.py:2287] 2024-09-19 12:17:17,638 >> loading file tokenizer.json
[INFO|tokenization_utils_base.py:2287] 2024-09-19 12:17:17,638 >> loading file added_tokens.json
[INFO|tokenization_utils_base.py:2287] 2024-09-19 12:17:17,638 >> loading file special_tokens_map.json
[INFO|tokenization_utils_base.py:2287] 2024-09-19 12:17:17,638 >> loading file tokenizer_config.json
[INFO|tokenization_utils_base.py:2533] 2024-09-19 12:17:17,897 >> Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
[INFO|template.py:270] 2024-09-19 12:17:17,897 >> Replace eos token: <|eot_id|>
[INFO|template.py:372] 2024-09-19 12:17:17,897 >> Add pad token: <|eot_id|>
[INFO|loader.py:50] 2024-09-19 12:17:17,897 >> Loading dataset Judge.json...
[INFO|configuration_utils.py:731] 2024-09-19 12:17:21,971 >> loading configuration file /home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693/config.json
[INFO|configuration_utils.py:800] 2024-09-19 12:17:21,972 >> Model config LlamaConfig {
"_name_or_path": "/home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693",
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 128000,
"eos_token_id": [
128001,
128008,
128009
],
"hidden_act": "silu",
"hidden_size": 4096,
"initializer_range": 0.02,
"intermediate_size": 14336,
"max_position_embeddings": 131072,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 32,
"num_hidden_layers": 32,
"num_key_value_heads": 8,
"pretraining_tp": 1,
"rms_norm_eps": 1e-05,
"rope_scaling": {
"factor": 8.0,
"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": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.43.2",
"use_cache": true,
"vocab_size": 128256
}
[WARNING|rope.py:57] 2024-09-19 12:17:21,972 >> Input length is smaller than max length. Consider increase input length.
[INFO|rope.py:63] 2024-09-19 12:17:21,972 >> Using linear scaling strategy and setting scaling factor to 1.0
[INFO|configuration_utils.py:731] 2024-09-19 12:17:23,421 >> loading configuration file /home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693/config.json
[INFO|configuration_utils.py:800] 2024-09-19 12:17:23,422 >> Model config LlamaConfig {
"_name_or_path": "/home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693",
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 128000,
"eos_token_id": [
128001,
128008,
128009
],
"hidden_act": "silu",
"hidden_size": 4096,
"initializer_range": 0.02,
"intermediate_size": 14336,
"max_position_embeddings": 131072,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 32,
"num_hidden_layers": 32,
"num_key_value_heads": 8,
"pretraining_tp": 1,
"rms_norm_eps": 1e-05,
"rope_scaling": {
"factor": 8.0,
"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": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.43.2",
"use_cache": true,
"vocab_size": 128256
}
[INFO|configuration_utils.py:733] 2024-09-19 12:17:24,156 >> loading configuration file config.json from cache at /home/marl/.cache/huggingface/hub/models--unslothai--other/snapshots/43d9e0f2f19a5d7836895f648dc0e762816acf77/config.json
[INFO|configuration_utils.py:800] 2024-09-19 12:17:24,157 >> Model config LlamaConfig {
"_name_or_path": "unslothai/other",
"architectures": [
"LlamaModel"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 0,
"initializer_range": 0.02,
"intermediate_size": 0,
"max_position_embeddings": 0,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 0,
"num_hidden_layers": 0,
"num_key_value_heads": 0,
"pretraining_tp": 1,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.43.2",
"use_cache": true,
"vocab_size": 0
}
[INFO|modeling_utils.py:3634] 2024-09-19 12:17:24,659 >> loading weights file model.safetensors from cache at /home/marl/.cache/huggingface/hub/models--unslothai--other/snapshots/43d9e0f2f19a5d7836895f648dc0e762816acf77/model.safetensors
[INFO|configuration_utils.py:1038] 2024-09-19 12:17:24,670 >> Generate config GenerationConfig {
"bos_token_id": 1,
"eos_token_id": 2
}
[INFO|configuration_utils.py:733] 2024-09-19 12:17:24,820 >> loading configuration file config.json from cache at /home/marl/.cache/huggingface/hub/models--unslothai--repeat/snapshots/7c48478c02f84ed89f149b0815cc0216ee831fb0/config.json
[INFO|configuration_utils.py:800] 2024-09-19 12:17:24,821 >> Model config LlamaConfig {
"_name_or_path": "unslothai/repeat",
"architectures": [
"LlamaModel"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 0,
"initializer_range": 0.02,
"intermediate_size": 0,
"max_position_embeddings": 0,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 0,
"num_hidden_layers": 0,
"num_key_value_heads": 0,
"pretraining_tp": 1,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.43.2",
"use_cache": true,
"vocab_size": 0
}
[INFO|modeling_utils.py:3634] 2024-09-19 12:17:24,823 >> loading weights file model.safetensors from cache at /home/marl/.cache/huggingface/hub/models--unslothai--repeat/snapshots/7c48478c02f84ed89f149b0815cc0216ee831fb0/model.safetensors
[INFO|configuration_utils.py:1038] 2024-09-19 12:17:24,824 >> Generate config GenerationConfig {
"bos_token_id": 1,
"eos_token_id": 2
}
[INFO|configuration_utils.py:733] 2024-09-19 12:17:25,466 >> loading configuration file config.json from cache at /home/marl/.cache/huggingface/hub/models--unslothai--vram-24/snapshots/61324ceeacd75b2b31f7a789a9c9d82058e6118c/config.json
[INFO|configuration_utils.py:800] 2024-09-19 12:17:25,467 >> Model config LlamaConfig {
"_name_or_path": "unslothai/vram-24",
"architectures": [
"LlamaModel"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 0,
"initializer_range": 0.02,
"intermediate_size": 0,
"max_position_embeddings": 0,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 0,
"num_hidden_layers": 0,
"num_key_value_heads": 0,
"pretraining_tp": 1,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.43.2",
"use_cache": true,
"vocab_size": 0
}
[INFO|modeling_utils.py:3634] 2024-09-19 12:17:25,853 >> loading weights file model.safetensors from cache at /home/marl/.cache/huggingface/hub/models--unslothai--vram-24/snapshots/61324ceeacd75b2b31f7a789a9c9d82058e6118c/model.safetensors
[INFO|configuration_utils.py:1038] 2024-09-19 12:17:25,854 >> Generate config GenerationConfig {
"bos_token_id": 1,
"eos_token_id": 2
}
[INFO|configuration_utils.py:733] 2024-09-19 12:17:26,452 >> loading configuration file config.json from cache at /home/marl/.cache/huggingface/hub/models--unslothai--1/snapshots/7ec782b7604cd9ea0781c23a4270f031650f5617/config.json
[INFO|configuration_utils.py:800] 2024-09-19 12:17:26,453 >> Model config LlamaConfig {
"_name_or_path": "unslothai/1",
"architectures": [
"LlamaModel"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 0,
"initializer_range": 0.02,
"intermediate_size": 0,
"max_position_embeddings": 2048,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 0,
"num_hidden_layers": 0,
"num_key_value_heads": 0,
"pretraining_tp": 1,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.43.2",
"use_cache": true,
"vocab_size": 0
}
[INFO|modeling_utils.py:3634] 2024-09-19 12:17:26,819 >> loading weights file model.safetensors from cache at /home/marl/.cache/huggingface/hub/models--unslothai--1/snapshots/7ec782b7604cd9ea0781c23a4270f031650f5617/model.safetensors
[INFO|configuration_utils.py:1038] 2024-09-19 12:17:26,821 >> Generate config GenerationConfig {
"bos_token_id": 1,
"eos_token_id": 2
}
[INFO|configuration_utils.py:731] 2024-09-19 12:17:26,821 >> loading configuration file /home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693/config.json
[INFO|configuration_utils.py:800] 2024-09-19 12:17:26,822 >> Model config LlamaConfig {
"_name_or_path": "/home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693",
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 128000,
"eos_token_id": [
128001,
128008,
128009
],
"hidden_act": "silu",
"hidden_size": 4096,
"initializer_range": 0.02,
"intermediate_size": 14336,
"max_position_embeddings": 131072,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 32,
"num_hidden_layers": 32,
"num_key_value_heads": 8,
"pretraining_tp": 1,
"rms_norm_eps": 1e-05,
"rope_scaling": {
"factor": 8.0,
"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": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.43.2",
"use_cache": true,
"vocab_size": 128256
}
[INFO|configuration_utils.py:731] 2024-09-19 12:17:26,836 >> loading configuration file /home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693/config.json
[INFO|configuration_utils.py:800] 2024-09-19 12:17:26,836 >> Model config LlamaConfig {
"_name_or_path": "/home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693",
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 128000,
"eos_token_id": [
128001,
128008,
128009
],
"hidden_act": "silu",
"hidden_size": 4096,
"initializer_range": 0.02,
"intermediate_size": 14336,
"max_position_embeddings": 131072,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 32,
"num_hidden_layers": 32,
"num_key_value_heads": 8,
"pretraining_tp": 1,
"rms_norm_eps": 1e-05,
"rope_scaling": {
"factor": 8.0,
"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": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.43.2",
"use_cache": true,
"vocab_size": 128256
}
[INFO|modeling_utils.py:3631] 2024-09-19 12:17:26,837 >> loading weights file /home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693/model.safetensors.index.json
[INFO|modeling_utils.py:1572] 2024-09-19 12:17:26,838 >> Instantiating LlamaForCausalLM model under default dtype torch.bfloat16.
[INFO|configuration_utils.py:1038] 2024-09-19 12:17:26,839 >> Generate config GenerationConfig {
"bos_token_id": 128000,
"eos_token_id": [
128001,
128008,
128009
]
}
[INFO|modeling_utils.py:4463] 2024-09-19 12:17:30,294 >> All model checkpoint weights were used when initializing LlamaForCausalLM.
[INFO|modeling_utils.py:4471] 2024-09-19 12:17:30,294 >> All the weights of LlamaForCausalLM were initialized from the model checkpoint at /home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693.
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|configuration_utils.py:991] 2024-09-19 12:17:30,297 >> loading configuration file /home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693/generation_config.json
[INFO|configuration_utils.py:1038] 2024-09-19 12:17:30,297 >> Generate config GenerationConfig {
"bos_token_id": 128000,
"do_sample": true,
"eos_token_id": [
128001,
128008,
128009
],
"temperature": 0.6,
"top_p": 0.9
}
[WARNING|logging.py:328] 2024-09-19 12:17:30,609 >> Unsloth: We successfully patched the tokenizer to add a {% if add_generation_prompt %} to the chat_template.
This is not a bug, but please notify the Unsloth maintainers - thanks!
[WARNING|logging.py:328] 2024-09-19 12:17:30,609 >> /home/marl/.cache/huggingface/hub/models--meta-llama--Meta-Llama-3.1-8B-Instruct/snapshots/07eb05b21d191a58c577b4a45982fe0c049d0693 does not have a padding token! Will use pad_token = <|finetune_right_pad_id|>.
[INFO|checkpointing.py:103] 2024-09-19 12:17:31,099 >> Gradient checkpointing enabled.
[INFO|adapter.py:302] 2024-09-19 12:17:31,099 >> Upcasting trainable params to float32.
[INFO|adapter.py:158] 2024-09-19 12:17:31,099 >> Fine-tuning method: DoRA
[INFO|misc.py:51] 2024-09-19 12:17:31,100 >> Found linear modules: v_proj,k_proj,o_proj,down_proj,q_proj,gate_proj,up_proj
[WARNING|logging.py:328] 2024-09-19 12:17:32,743 >> Not an error, but Unsloth cannot patch MLP layers with our manual autograd engine since either LoRA adapters
are not enabled or a bias term (like in Qwen) is used.
[WARNING|logging.py:328] 2024-09-19 12:17:32,744 >> Not an error, but Unsloth cannot patch Attention layers with our manual autograd engine since either LoRA adapters
are not enabled or a bias term (like in Qwen) is used.
[WARNING|logging.py:328] 2024-09-19 12:17:32,744 >> Not an error, but Unsloth cannot patch O projection layer with our manual autograd engine since either LoRA adapters
are not enabled or a bias term (like in Qwen) is used.
[WARNING|logging.py:328] 2024-09-19 12:17:32,745 >> Unsloth 2024.9 patched 32 layers with 0 QKV layers, 0 O layers and 0 MLP layers.
[INFO|loader.py:196] 2024-09-19 12:17:33,762 >> trainable params: 43,319,296 || all params: 8,073,580,544 || trainable%: 0.5366
[INFO|trainer.py:648] 2024-09-19 12:17:33,779 >> Using auto half precision backend
[WARNING|<string>:213] 2024-09-19 12:17:34,063 >> ==((====))== Unsloth - 2x faster free finetuning | Num GPUs = 1
\\ /| Num examples = 30 | Num Epochs = 10
O^O/ \_/ \ Batch size per device = 2 | Gradient Accumulation steps = 8
\ / Total batch size = 16 | Total steps = 10
"-____-" Number of trainable parameters = 43,319,296
[INFO|callbacks.py:137] 2024-09-19 12:17:34,578 >> Initial PiSSA adatper will be saved at: saves/LLaMA3-8B-Chat/lora/JudgeDredd/pissa_init.
[INFO|callbacks.py:310] 2024-09-19 12:18:40,414 >> {'loss': 1.2688, 'learning_rate': 2.5000e-05, 'epoch': 2.67, 'throughput': 228.45}
[INFO|callbacks.py:310] 2024-09-19 12:19:44,899 >> {'loss': 0.3328, 'learning_rate': 0.0000e+00, 'epoch': 5.33, 'throughput': 230.69}
[INFO|trainer.py:3503] 2024-09-19 12:19:44,900 >> Saving model checkpoint to saves/LLaMA3-8B-Chat/lora/JudgeDredd/checkpoint-10
[INFO|<string>:478] 2024-09-19 12:19:45,579 >>
Training completed. Do not forget to share your model on huggingface.co/models =)
[INFO|callbacks.py:153] 2024-09-19 12:19:45,581 >> Converted PiSSA adapter will be saved at: saves/LLaMA3-8B-Chat/lora/JudgeDredd/pissa_converted.
[INFO|trainer.py:3503] 2024-09-19 12:19:47,418 >> Saving model checkpoint to saves/LLaMA3-8B-Chat/lora/JudgeDredd
[WARNING|ploting.py:89] 2024-09-19 12:19:47,819 >> No metric eval_loss to plot.
[WARNING|ploting.py:89] 2024-09-19 12:19:47,819 >> No metric eval_accuracy to plot.
[INFO|modelcard.py:449] 2024-09-19 12:19:47,820 >> Dropping the following result as it does not have all the necessary fields:
{'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}
|