Yi-6B-ruozhiba-5e-5-50 / trainer_state.json
yyx123's picture
Model save
42e8bcb verified
raw
history blame
19.7 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 10.0,
"eval_steps": 500,
"global_step": 550,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02,
"learning_rate": 9.09090909090909e-07,
"loss": 2.7431,
"step": 1
},
{
"epoch": 0.07,
"learning_rate": 3.636363636363636e-06,
"loss": 2.889,
"step": 4
},
{
"epoch": 0.15,
"learning_rate": 7.272727272727272e-06,
"loss": 2.7193,
"step": 8
},
{
"epoch": 0.22,
"learning_rate": 1.0909090909090909e-05,
"loss": 2.6675,
"step": 12
},
{
"epoch": 0.29,
"learning_rate": 1.4545454545454545e-05,
"loss": 2.608,
"step": 16
},
{
"epoch": 0.36,
"learning_rate": 1.8181818181818182e-05,
"loss": 2.6623,
"step": 20
},
{
"epoch": 0.44,
"learning_rate": 2.1818181818181818e-05,
"loss": 2.7249,
"step": 24
},
{
"epoch": 0.51,
"learning_rate": 2.5454545454545454e-05,
"loss": 2.7954,
"step": 28
},
{
"epoch": 0.58,
"learning_rate": 2.909090909090909e-05,
"loss": 2.6101,
"step": 32
},
{
"epoch": 0.65,
"learning_rate": 3.272727272727273e-05,
"loss": 2.4391,
"step": 36
},
{
"epoch": 0.73,
"learning_rate": 3.6363636363636364e-05,
"loss": 2.2331,
"step": 40
},
{
"epoch": 0.8,
"learning_rate": 4e-05,
"loss": 2.2489,
"step": 44
},
{
"epoch": 0.87,
"learning_rate": 4.3636363636363636e-05,
"loss": 2.1787,
"step": 48
},
{
"epoch": 0.95,
"learning_rate": 4.7272727272727275e-05,
"loss": 2.2181,
"step": 52
},
{
"epoch": 1.02,
"learning_rate": 4.999949650182266e-05,
"loss": 2.2997,
"step": 56
},
{
"epoch": 1.09,
"learning_rate": 4.9987413559579636e-05,
"loss": 2.0619,
"step": 60
},
{
"epoch": 1.16,
"learning_rate": 4.995922759815339e-05,
"loss": 2.1645,
"step": 64
},
{
"epoch": 1.24,
"learning_rate": 4.991495678185202e-05,
"loss": 1.9819,
"step": 68
},
{
"epoch": 1.31,
"learning_rate": 4.985462964079137e-05,
"loss": 2.0363,
"step": 72
},
{
"epoch": 1.38,
"learning_rate": 4.977828505250903e-05,
"loss": 1.9152,
"step": 76
},
{
"epoch": 1.45,
"learning_rate": 4.968597221690986e-05,
"loss": 1.9794,
"step": 80
},
{
"epoch": 1.53,
"learning_rate": 4.957775062455933e-05,
"loss": 1.9805,
"step": 84
},
{
"epoch": 1.6,
"learning_rate": 4.9453690018345144e-05,
"loss": 1.9941,
"step": 88
},
{
"epoch": 1.67,
"learning_rate": 4.931387034853173e-05,
"loss": 1.9186,
"step": 92
},
{
"epoch": 1.75,
"learning_rate": 4.915838172123671e-05,
"loss": 1.9123,
"step": 96
},
{
"epoch": 1.82,
"learning_rate": 4.898732434036244e-05,
"loss": 1.965,
"step": 100
},
{
"epoch": 1.89,
"learning_rate": 4.880080844302004e-05,
"loss": 1.8973,
"step": 104
},
{
"epoch": 1.96,
"learning_rate": 4.859895422848767e-05,
"loss": 1.8473,
"step": 108
},
{
"epoch": 2.0,
"pls_score": 58.0,
"std": 4.354308211415448,
"step": 110
},
{
"epoch": 2.0,
"eval_loss": 1.9843206405639648,
"eval_runtime": 4.904,
"eval_samples_per_second": 4.69,
"eval_steps_per_second": 1.223,
"step": 110
},
{
"epoch": 2.04,
"learning_rate": 4.838189178074867e-05,
"loss": 2.0581,
"step": 112
},
{
"epoch": 2.11,
"learning_rate": 4.8149760984659506e-05,
"loss": 1.8467,
"step": 116
},
{
"epoch": 2.18,
"learning_rate": 4.790271143580174e-05,
"loss": 1.9691,
"step": 120
},
{
"epoch": 2.25,
"learning_rate": 4.764090234407577e-05,
"loss": 1.8183,
"step": 124
},
{
"epoch": 2.33,
"learning_rate": 4.7364502431098844e-05,
"loss": 1.7403,
"step": 128
},
{
"epoch": 2.4,
"learning_rate": 4.707368982147318e-05,
"loss": 1.7877,
"step": 132
},
{
"epoch": 2.47,
"learning_rate": 4.6768651927994434e-05,
"loss": 1.7595,
"step": 136
},
{
"epoch": 2.55,
"learning_rate": 4.644958533087443e-05,
"loss": 1.7273,
"step": 140
},
{
"epoch": 2.62,
"learning_rate": 4.611669565105596e-05,
"loss": 1.669,
"step": 144
},
{
"epoch": 2.69,
"learning_rate": 4.5770197417701365e-05,
"loss": 1.742,
"step": 148
},
{
"epoch": 2.76,
"learning_rate": 4.5410313929940244e-05,
"loss": 1.6976,
"step": 152
},
{
"epoch": 2.84,
"learning_rate": 4.503727711296538e-05,
"loss": 1.8632,
"step": 156
},
{
"epoch": 2.91,
"learning_rate": 4.465132736856969e-05,
"loss": 1.9186,
"step": 160
},
{
"epoch": 2.98,
"learning_rate": 4.425271342022039e-05,
"loss": 1.7645,
"step": 164
},
{
"epoch": 3.0,
"pls_score": 64.0,
"std": 3.7523326078587433,
"step": 165
},
{
"epoch": 3.0,
"eval_loss": 1.968030333518982,
"eval_runtime": 4.9015,
"eval_samples_per_second": 4.692,
"eval_steps_per_second": 1.224,
"step": 165
},
{
"epoch": 3.05,
"learning_rate": 4.384169215277041e-05,
"loss": 1.4812,
"step": 168
},
{
"epoch": 3.13,
"learning_rate": 4.341852844691012e-05,
"loss": 1.6649,
"step": 172
},
{
"epoch": 3.2,
"learning_rate": 4.2983495008466276e-05,
"loss": 1.5835,
"step": 176
},
{
"epoch": 3.27,
"learning_rate": 4.2536872192658036e-05,
"loss": 1.6401,
"step": 180
},
{
"epoch": 3.35,
"learning_rate": 4.2078947823423364e-05,
"loss": 1.573,
"step": 184
},
{
"epoch": 3.42,
"learning_rate": 4.161001700793231e-05,
"loss": 1.4984,
"step": 188
},
{
"epoch": 3.49,
"learning_rate": 4.113038194640658e-05,
"loss": 1.5315,
"step": 192
},
{
"epoch": 3.56,
"learning_rate": 4.064035173736804e-05,
"loss": 1.541,
"step": 196
},
{
"epoch": 3.64,
"learning_rate": 4.014024217844167e-05,
"loss": 1.5492,
"step": 200
},
{
"epoch": 3.71,
"learning_rate": 3.9630375562841295e-05,
"loss": 1.6265,
"step": 204
},
{
"epoch": 3.78,
"learning_rate": 3.911108047166924e-05,
"loss": 1.5753,
"step": 208
},
{
"epoch": 3.85,
"learning_rate": 3.858269156216383e-05,
"loss": 1.4876,
"step": 212
},
{
"epoch": 3.93,
"learning_rate": 3.804554935203115e-05,
"loss": 1.5884,
"step": 216
},
{
"epoch": 4.0,
"learning_rate": 3.7500000000000003e-05,
"loss": 1.4795,
"step": 220
},
{
"epoch": 4.0,
"pls_score": 60.8,
"std": 3.750626614313932,
"step": 220
},
{
"epoch": 4.0,
"eval_loss": 2.076031446456909,
"eval_runtime": 4.9029,
"eval_samples_per_second": 4.691,
"eval_steps_per_second": 1.224,
"step": 220
},
{
"epoch": 4.07,
"learning_rate": 3.694639508274158e-05,
"loss": 1.3002,
"step": 224
},
{
"epoch": 4.15,
"learning_rate": 3.638509136829758e-05,
"loss": 1.344,
"step": 228
},
{
"epoch": 4.22,
"learning_rate": 3.581645058616271e-05,
"loss": 1.209,
"step": 232
},
{
"epoch": 4.29,
"learning_rate": 3.5240839194169885e-05,
"loss": 1.3386,
"step": 236
},
{
"epoch": 4.36,
"learning_rate": 3.465862814232822e-05,
"loss": 1.3247,
"step": 240
},
{
"epoch": 4.44,
"learning_rate": 3.4070192633766025e-05,
"loss": 1.2247,
"step": 244
},
{
"epoch": 4.51,
"learning_rate": 3.3475911882933015e-05,
"loss": 1.2683,
"step": 248
},
{
"epoch": 4.58,
"learning_rate": 3.2876168871217325e-05,
"loss": 1.2856,
"step": 252
},
{
"epoch": 4.65,
"learning_rate": 3.2271350100134975e-05,
"loss": 1.271,
"step": 256
},
{
"epoch": 4.73,
"learning_rate": 3.166184534225087e-05,
"loss": 1.207,
"step": 260
},
{
"epoch": 4.8,
"learning_rate": 3.104804738999169e-05,
"loss": 1.2884,
"step": 264
},
{
"epoch": 4.87,
"learning_rate": 3.0430351802512698e-05,
"loss": 1.2375,
"step": 268
},
{
"epoch": 4.95,
"learning_rate": 2.9809156650781528e-05,
"loss": 1.2467,
"step": 272
},
{
"epoch": 5.0,
"pls_score": 62.8,
"std": 4.15730682052696,
"step": 275
},
{
"epoch": 5.0,
"eval_loss": 2.2714545726776123,
"eval_runtime": 4.904,
"eval_samples_per_second": 4.69,
"eval_steps_per_second": 1.224,
"step": 275
},
{
"epoch": 5.02,
"learning_rate": 2.918486226104327e-05,
"loss": 1.1868,
"step": 276
},
{
"epoch": 5.09,
"learning_rate": 2.8557870956832132e-05,
"loss": 1.099,
"step": 280
},
{
"epoch": 5.16,
"learning_rate": 2.792858679969596e-05,
"loss": 1.0657,
"step": 284
},
{
"epoch": 5.24,
"learning_rate": 2.7297415328800692e-05,
"loss": 1.0596,
"step": 288
},
{
"epoch": 5.31,
"learning_rate": 2.6664763299582602e-05,
"loss": 0.9201,
"step": 292
},
{
"epoch": 5.38,
"learning_rate": 2.6031038421616683e-05,
"loss": 1.0683,
"step": 296
},
{
"epoch": 5.45,
"learning_rate": 2.5396649095870202e-05,
"loss": 1.0069,
"step": 300
},
{
"epoch": 5.53,
"learning_rate": 2.4762004151510584e-05,
"loss": 0.9486,
"step": 304
},
{
"epoch": 5.6,
"learning_rate": 2.4127512582437485e-05,
"loss": 0.937,
"step": 308
},
{
"epoch": 5.67,
"learning_rate": 2.349358328370854e-05,
"loss": 1.0102,
"step": 312
},
{
"epoch": 5.75,
"learning_rate": 2.2860624788029013e-05,
"loss": 1.0058,
"step": 316
},
{
"epoch": 5.82,
"learning_rate": 2.222904500247473e-05,
"loss": 1.0564,
"step": 320
},
{
"epoch": 5.89,
"learning_rate": 2.1599250945618402e-05,
"loss": 0.9122,
"step": 324
},
{
"epoch": 5.96,
"learning_rate": 2.09716484852284e-05,
"loss": 1.0034,
"step": 328
},
{
"epoch": 6.04,
"learning_rate": 2.034664207670925e-05,
"loss": 0.8258,
"step": 332
},
{
"epoch": 6.11,
"learning_rate": 1.972463450245226e-05,
"loss": 0.8152,
"step": 336
},
{
"epoch": 6.18,
"learning_rate": 1.9106026612264316e-05,
"loss": 0.7462,
"step": 340
},
{
"epoch": 6.25,
"learning_rate": 1.84912170650422e-05,
"loss": 0.7995,
"step": 344
},
{
"epoch": 6.33,
"learning_rate": 1.7880602071858692e-05,
"loss": 0.7714,
"step": 348
},
{
"epoch": 6.4,
"learning_rate": 1.7274575140626318e-05,
"loss": 0.8581,
"step": 352
},
{
"epoch": 6.47,
"learning_rate": 1.667352682250298e-05,
"loss": 0.7631,
"step": 356
},
{
"epoch": 6.55,
"learning_rate": 1.6077844460203206e-05,
"loss": 0.8117,
"step": 360
},
{
"epoch": 6.62,
"learning_rate": 1.5487911938376924e-05,
"loss": 0.7725,
"step": 364
},
{
"epoch": 6.69,
"learning_rate": 1.4904109436216884e-05,
"loss": 0.8519,
"step": 368
},
{
"epoch": 6.76,
"learning_rate": 1.4326813182453958e-05,
"loss": 0.8037,
"step": 372
},
{
"epoch": 6.84,
"learning_rate": 1.3756395212898359e-05,
"loss": 0.848,
"step": 376
},
{
"epoch": 6.91,
"learning_rate": 1.3193223130682936e-05,
"loss": 0.797,
"step": 380
},
{
"epoch": 6.98,
"learning_rate": 1.2637659869363083e-05,
"loss": 0.7614,
"step": 384
},
{
"epoch": 7.0,
"pls_score": 57.2,
"std": 4.3085032203771165,
"step": 385
},
{
"epoch": 7.0,
"eval_loss": 2.7865102291107178,
"eval_runtime": 4.9204,
"eval_samples_per_second": 4.674,
"eval_steps_per_second": 1.219,
"step": 385
},
{
"epoch": 7.05,
"learning_rate": 1.2090063459025955e-05,
"loss": 0.8261,
"step": 388
},
{
"epoch": 7.13,
"learning_rate": 1.155078679555969e-05,
"loss": 0.639,
"step": 392
},
{
"epoch": 7.2,
"learning_rate": 1.1020177413231334e-05,
"loss": 0.748,
"step": 396
},
{
"epoch": 7.27,
"learning_rate": 1.049857726072005e-05,
"loss": 0.6049,
"step": 400
},
{
"epoch": 7.35,
"learning_rate": 9.986322480749927e-06,
"loss": 0.5971,
"step": 404
},
{
"epoch": 7.42,
"learning_rate": 9.483743193464408e-06,
"loss": 0.6648,
"step": 408
},
{
"epoch": 7.49,
"learning_rate": 8.991163283681944e-06,
"loss": 0.6513,
"step": 412
},
{
"epoch": 7.56,
"learning_rate": 8.508900192169964e-06,
"loss": 0.5917,
"step": 416
},
{
"epoch": 7.64,
"learning_rate": 8.0372647110717e-06,
"loss": 0.6839,
"step": 420
},
{
"epoch": 7.71,
"learning_rate": 7.576560783617668e-06,
"loss": 0.6658,
"step": 424
},
{
"epoch": 7.78,
"learning_rate": 7.127085308250914e-06,
"loss": 0.691,
"step": 428
},
{
"epoch": 7.85,
"learning_rate": 6.689127947292231e-06,
"loss": 0.7198,
"step": 432
},
{
"epoch": 7.93,
"learning_rate": 6.2629709402686535e-06,
"loss": 0.6533,
"step": 436
},
{
"epoch": 8.0,
"learning_rate": 5.848888922025553e-06,
"loss": 0.6237,
"step": 440
},
{
"epoch": 8.0,
"pls_score": 60.4,
"std": 4.507416111254875,
"step": 440
},
{
"epoch": 8.0,
"eval_loss": 2.958909511566162,
"eval_runtime": 4.9181,
"eval_samples_per_second": 4.677,
"eval_steps_per_second": 1.22,
"step": 440
},
{
"epoch": 8.07,
"learning_rate": 5.4471487457395225e-06,
"loss": 0.5912,
"step": 444
},
{
"epoch": 8.15,
"learning_rate": 5.058009310946119e-06,
"loss": 0.6374,
"step": 448
},
{
"epoch": 8.22,
"learning_rate": 4.681721396693303e-06,
"loss": 0.5975,
"step": 452
},
{
"epoch": 8.29,
"learning_rate": 4.318527499928074e-06,
"loss": 0.5448,
"step": 456
},
{
"epoch": 8.36,
"learning_rate": 3.968661679220468e-06,
"loss": 0.5574,
"step": 460
},
{
"epoch": 8.44,
"learning_rate": 3.632349403925664e-06,
"loss": 0.6206,
"step": 464
},
{
"epoch": 8.51,
"learning_rate": 3.3098074088812686e-06,
"loss": 0.6207,
"step": 468
},
{
"epoch": 8.58,
"learning_rate": 3.0012435547336737e-06,
"loss": 0.5411,
"step": 472
},
{
"epoch": 8.65,
"learning_rate": 2.7068566939831645e-06,
"loss": 0.6231,
"step": 476
},
{
"epoch": 8.73,
"learning_rate": 2.4268365428344736e-06,
"loss": 0.5944,
"step": 480
},
{
"epoch": 8.8,
"learning_rate": 2.1613635589349756e-06,
"loss": 0.6279,
"step": 484
},
{
"epoch": 8.87,
"learning_rate": 1.9106088250797267e-06,
"loss": 0.561,
"step": 488
},
{
"epoch": 8.95,
"learning_rate": 1.674733938957873e-06,
"loss": 0.5711,
"step": 492
},
{
"epoch": 9.0,
"pls_score": 55.2,
"std": 4.365684367885522,
"step": 495
},
{
"epoch": 9.0,
"eval_loss": 3.058058977127075,
"eval_runtime": 4.9251,
"eval_samples_per_second": 4.67,
"eval_steps_per_second": 1.218,
"step": 495
},
{
"epoch": 9.02,
"learning_rate": 1.4538909090118846e-06,
"loss": 0.5386,
"step": 496
},
{
"epoch": 9.09,
"learning_rate": 1.248222056476367e-06,
"loss": 0.6031,
"step": 500
},
{
"epoch": 9.16,
"learning_rate": 1.0578599236598707e-06,
"loss": 0.5676,
"step": 504
},
{
"epoch": 9.24,
"learning_rate": 8.829271885286094e-07,
"loss": 0.5298,
"step": 508
},
{
"epoch": 9.31,
"learning_rate": 7.235365856472442e-07,
"loss": 0.5428,
"step": 512
},
{
"epoch": 9.38,
"learning_rate": 5.797908335276214e-07,
"loss": 0.5811,
"step": 516
},
{
"epoch": 9.45,
"learning_rate": 4.517825684323324e-07,
"loss": 0.5703,
"step": 520
},
{
"epoch": 9.53,
"learning_rate": 3.395942846757066e-07,
"loss": 0.6147,
"step": 524
},
{
"epoch": 9.6,
"learning_rate": 2.4329828146074095e-07,
"loss": 0.5451,
"step": 528
},
{
"epoch": 9.67,
"learning_rate": 1.6295661628624447e-07,
"loss": 0.5303,
"step": 532
},
{
"epoch": 9.75,
"learning_rate": 9.862106495415469e-08,
"loss": 0.5629,
"step": 536
},
{
"epoch": 9.82,
"learning_rate": 5.033308820289184e-08,
"loss": 0.5472,
"step": 540
},
{
"epoch": 9.89,
"learning_rate": 1.812380498815991e-08,
"loss": 0.5334,
"step": 544
},
{
"epoch": 9.96,
"learning_rate": 2.0139724285161977e-09,
"loss": 0.5884,
"step": 548
},
{
"epoch": 10.0,
"pls_score": 60.0,
"std": 4.195235392680606,
"step": 550
},
{
"epoch": 10.0,
"eval_loss": 3.072557210922241,
"eval_runtime": 4.9196,
"eval_samples_per_second": 4.675,
"eval_steps_per_second": 1.22,
"step": 550
},
{
"epoch": 10.0,
"step": 550,
"total_flos": 1.870637420231885e+16,
"train_loss": 0.2617128190127286,
"train_runtime": 10832.3445,
"train_samples_per_second": 0.2,
"train_steps_per_second": 0.051
}
],
"logging_steps": 4,
"max_steps": 550,
"num_input_tokens_seen": 0,
"num_train_epochs": 10,
"save_steps": 55,
"total_flos": 1.870637420231885e+16,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}