{
"best_metric": 0.06558172404766083,
"best_model_checkpoint": "./beans_outputs/checkpoint-520",
"epoch": 5.0,
"eval_steps": 500,
"global_step": 650,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.07692307692307693,
"grad_norm": 2.198366403579712,
"learning_rate": 1.9692307692307696e-05,
"loss": 1.0245,
"step": 10
},
{
"epoch": 0.15384615384615385,
"grad_norm": 1.8460220098495483,
"learning_rate": 1.9384615384615386e-05,
"loss": 0.9453,
"step": 20
},
{
"epoch": 0.23076923076923078,
"grad_norm": 2.117300271987915,
"learning_rate": 1.907692307692308e-05,
"loss": 0.8406,
"step": 30
},
{
"epoch": 0.3076923076923077,
"grad_norm": 2.645448684692383,
"learning_rate": 1.876923076923077e-05,
"loss": 0.6951,
"step": 40
},
{
"epoch": 0.38461538461538464,
"grad_norm": 4.24210262298584,
"learning_rate": 1.8461538461538465e-05,
"loss": 0.662,
"step": 50
},
{
"epoch": 0.46153846153846156,
"grad_norm": 2.095567464828491,
"learning_rate": 1.8153846153846155e-05,
"loss": 0.5811,
"step": 60
},
{
"epoch": 0.5384615384615384,
"grad_norm": 2.079636573791504,
"learning_rate": 1.784615384615385e-05,
"loss": 0.5193,
"step": 70
},
{
"epoch": 0.6153846153846154,
"grad_norm": 1.218340277671814,
"learning_rate": 1.753846153846154e-05,
"loss": 0.3945,
"step": 80
},
{
"epoch": 0.6923076923076923,
"grad_norm": 2.4416184425354004,
"learning_rate": 1.7230769230769234e-05,
"loss": 0.3528,
"step": 90
},
{
"epoch": 0.7692307692307693,
"grad_norm": 2.8852720260620117,
"learning_rate": 1.6923076923076924e-05,
"loss": 0.3689,
"step": 100
},
{
"epoch": 0.8461538461538461,
"grad_norm": 1.9503966569900513,
"learning_rate": 1.6615384615384618e-05,
"loss": 0.3368,
"step": 110
},
{
"epoch": 0.9230769230769231,
"grad_norm": 2.335829496383667,
"learning_rate": 1.630769230769231e-05,
"loss": 0.2468,
"step": 120
},
{
"epoch": 1.0,
"grad_norm": 7.0948638916015625,
"learning_rate": 1.6000000000000003e-05,
"loss": 0.2859,
"step": 130
},
{
"epoch": 1.0,
"eval_accuracy": 0.9624060150375939,
"eval_loss": 0.21892516314983368,
"eval_runtime": 0.9855,
"eval_samples_per_second": 134.957,
"eval_steps_per_second": 17.25,
"step": 130
},
{
"epoch": 1.0769230769230769,
"grad_norm": 1.765810251235962,
"learning_rate": 1.5692307692307693e-05,
"loss": 0.2006,
"step": 140
},
{
"epoch": 1.1538461538461537,
"grad_norm": 3.1403791904449463,
"learning_rate": 1.5384615384615387e-05,
"loss": 0.2243,
"step": 150
},
{
"epoch": 1.2307692307692308,
"grad_norm": 1.4933677911758423,
"learning_rate": 1.5076923076923078e-05,
"loss": 0.2035,
"step": 160
},
{
"epoch": 1.3076923076923077,
"grad_norm": 0.5171657800674438,
"learning_rate": 1.4769230769230772e-05,
"loss": 0.2208,
"step": 170
},
{
"epoch": 1.3846153846153846,
"grad_norm": 0.600874662399292,
"learning_rate": 1.4461538461538462e-05,
"loss": 0.211,
"step": 180
},
{
"epoch": 1.4615384615384617,
"grad_norm": 3.5654282569885254,
"learning_rate": 1.4153846153846156e-05,
"loss": 0.1607,
"step": 190
},
{
"epoch": 1.5384615384615383,
"grad_norm": 0.47061973810195923,
"learning_rate": 1.3846153846153847e-05,
"loss": 0.2526,
"step": 200
},
{
"epoch": 1.6153846153846154,
"grad_norm": 0.6342479586601257,
"learning_rate": 1.353846153846154e-05,
"loss": 0.1804,
"step": 210
},
{
"epoch": 1.6923076923076923,
"grad_norm": 2.391136884689331,
"learning_rate": 1.3230769230769231e-05,
"loss": 0.2121,
"step": 220
},
{
"epoch": 1.7692307692307692,
"grad_norm": 1.8535972833633423,
"learning_rate": 1.2923076923076925e-05,
"loss": 0.1678,
"step": 230
},
{
"epoch": 1.8461538461538463,
"grad_norm": 11.256393432617188,
"learning_rate": 1.2615384615384616e-05,
"loss": 0.2574,
"step": 240
},
{
"epoch": 1.9230769230769231,
"grad_norm": 0.4076194167137146,
"learning_rate": 1.230769230769231e-05,
"loss": 0.1442,
"step": 250
},
{
"epoch": 2.0,
"grad_norm": 0.3490784168243408,
"learning_rate": 1.2e-05,
"loss": 0.1316,
"step": 260
},
{
"epoch": 2.0,
"eval_accuracy": 0.9699248120300752,
"eval_loss": 0.13337133824825287,
"eval_runtime": 0.9938,
"eval_samples_per_second": 133.827,
"eval_steps_per_second": 17.106,
"step": 260
},
{
"epoch": 2.076923076923077,
"grad_norm": 0.5328232645988464,
"learning_rate": 1.1692307692307694e-05,
"loss": 0.0897,
"step": 270
},
{
"epoch": 2.1538461538461537,
"grad_norm": 0.3375113904476166,
"learning_rate": 1.1384615384615385e-05,
"loss": 0.1475,
"step": 280
},
{
"epoch": 2.230769230769231,
"grad_norm": 7.195676326751709,
"learning_rate": 1.1076923076923079e-05,
"loss": 0.1321,
"step": 290
},
{
"epoch": 2.3076923076923075,
"grad_norm": 8.226975440979004,
"learning_rate": 1.076923076923077e-05,
"loss": 0.1375,
"step": 300
},
{
"epoch": 2.3846153846153846,
"grad_norm": 5.051274299621582,
"learning_rate": 1.0461538461538463e-05,
"loss": 0.1755,
"step": 310
},
{
"epoch": 2.4615384615384617,
"grad_norm": 0.30559489130973816,
"learning_rate": 1.0153846153846154e-05,
"loss": 0.1542,
"step": 320
},
{
"epoch": 2.5384615384615383,
"grad_norm": 6.551879405975342,
"learning_rate": 9.846153846153848e-06,
"loss": 0.1872,
"step": 330
},
{
"epoch": 2.6153846153846154,
"grad_norm": 1.2321629524230957,
"learning_rate": 9.53846153846154e-06,
"loss": 0.0714,
"step": 340
},
{
"epoch": 2.6923076923076925,
"grad_norm": 0.3684569299221039,
"learning_rate": 9.230769230769232e-06,
"loss": 0.1325,
"step": 350
},
{
"epoch": 2.769230769230769,
"grad_norm": 9.62742805480957,
"learning_rate": 8.923076923076925e-06,
"loss": 0.1203,
"step": 360
},
{
"epoch": 2.8461538461538463,
"grad_norm": 0.3335096836090088,
"learning_rate": 8.615384615384617e-06,
"loss": 0.154,
"step": 370
},
{
"epoch": 2.9230769230769234,
"grad_norm": 1.1392183303833008,
"learning_rate": 8.307692307692309e-06,
"loss": 0.1306,
"step": 380
},
{
"epoch": 3.0,
"grad_norm": 0.29587361216545105,
"learning_rate": 8.000000000000001e-06,
"loss": 0.1438,
"step": 390
},
{
"epoch": 3.0,
"eval_accuracy": 0.9699248120300752,
"eval_loss": 0.09808181971311569,
"eval_runtime": 0.9838,
"eval_samples_per_second": 135.192,
"eval_steps_per_second": 17.28,
"step": 390
},
{
"epoch": 3.076923076923077,
"grad_norm": 2.2563395500183105,
"learning_rate": 7.692307692307694e-06,
"loss": 0.1325,
"step": 400
},
{
"epoch": 3.1538461538461537,
"grad_norm": 6.758065223693848,
"learning_rate": 7.384615384615386e-06,
"loss": 0.154,
"step": 410
},
{
"epoch": 3.230769230769231,
"grad_norm": 8.712298393249512,
"learning_rate": 7.076923076923078e-06,
"loss": 0.1422,
"step": 420
},
{
"epoch": 3.3076923076923075,
"grad_norm": 0.23161759972572327,
"learning_rate": 6.76923076923077e-06,
"loss": 0.0768,
"step": 430
},
{
"epoch": 3.3846153846153846,
"grad_norm": 5.164632797241211,
"learning_rate": 6.461538461538463e-06,
"loss": 0.1136,
"step": 440
},
{
"epoch": 3.4615384615384617,
"grad_norm": 12.590032577514648,
"learning_rate": 6.153846153846155e-06,
"loss": 0.0854,
"step": 450
},
{
"epoch": 3.5384615384615383,
"grad_norm": 4.483827590942383,
"learning_rate": 5.846153846153847e-06,
"loss": 0.098,
"step": 460
},
{
"epoch": 3.6153846153846154,
"grad_norm": 7.863719940185547,
"learning_rate": 5.538461538461539e-06,
"loss": 0.0778,
"step": 470
},
{
"epoch": 3.6923076923076925,
"grad_norm": 0.21892791986465454,
"learning_rate": 5.230769230769232e-06,
"loss": 0.0957,
"step": 480
},
{
"epoch": 3.769230769230769,
"grad_norm": 0.2107536494731903,
"learning_rate": 4.923076923076924e-06,
"loss": 0.0775,
"step": 490
},
{
"epoch": 3.8461538461538463,
"grad_norm": 0.9603075981140137,
"learning_rate": 4.615384615384616e-06,
"loss": 0.1357,
"step": 500
},
{
"epoch": 3.9230769230769234,
"grad_norm": 0.21214883029460907,
"learning_rate": 4.307692307692308e-06,
"loss": 0.1225,
"step": 510
},
{
"epoch": 4.0,
"grad_norm": 0.2996317744255066,
"learning_rate": 4.000000000000001e-06,
"loss": 0.0833,
"step": 520
},
{
"epoch": 4.0,
"eval_accuracy": 0.9924812030075187,
"eval_loss": 0.06558172404766083,
"eval_runtime": 1.0412,
"eval_samples_per_second": 127.735,
"eval_steps_per_second": 16.327,
"step": 520
},
{
"epoch": 4.076923076923077,
"grad_norm": 9.374862670898438,
"learning_rate": 3.692307692307693e-06,
"loss": 0.1009,
"step": 530
},
{
"epoch": 4.153846153846154,
"grad_norm": 1.5623271465301514,
"learning_rate": 3.384615384615385e-06,
"loss": 0.1658,
"step": 540
},
{
"epoch": 4.230769230769231,
"grad_norm": 0.3211674094200134,
"learning_rate": 3.0769230769230774e-06,
"loss": 0.1534,
"step": 550
},
{
"epoch": 4.3076923076923075,
"grad_norm": 1.208674430847168,
"learning_rate": 2.7692307692307697e-06,
"loss": 0.0504,
"step": 560
},
{
"epoch": 4.384615384615385,
"grad_norm": 0.3158506751060486,
"learning_rate": 2.461538461538462e-06,
"loss": 0.1345,
"step": 570
},
{
"epoch": 4.461538461538462,
"grad_norm": 0.2690170705318451,
"learning_rate": 2.153846153846154e-06,
"loss": 0.0816,
"step": 580
},
{
"epoch": 4.538461538461538,
"grad_norm": 9.395990371704102,
"learning_rate": 1.8461538461538465e-06,
"loss": 0.0714,
"step": 590
},
{
"epoch": 4.615384615384615,
"grad_norm": 3.067613124847412,
"learning_rate": 1.5384615384615387e-06,
"loss": 0.0892,
"step": 600
},
{
"epoch": 4.6923076923076925,
"grad_norm": 5.167181491851807,
"learning_rate": 1.230769230769231e-06,
"loss": 0.087,
"step": 610
},
{
"epoch": 4.769230769230769,
"grad_norm": 0.2610948383808136,
"learning_rate": 9.230769230769232e-07,
"loss": 0.0807,
"step": 620
},
{
"epoch": 4.846153846153846,
"grad_norm": 0.22187209129333496,
"learning_rate": 6.153846153846155e-07,
"loss": 0.0668,
"step": 630
},
{
"epoch": 4.923076923076923,
"grad_norm": 0.23086942732334137,
"learning_rate": 3.0769230769230774e-07,
"loss": 0.0784,
"step": 640
},
{
"epoch": 5.0,
"grad_norm": 0.6445406675338745,
"learning_rate": 0.0,
"loss": 0.1107,
"step": 650
},
{
"epoch": 5.0,
"eval_accuracy": 0.9774436090225563,
"eval_loss": 0.0816693976521492,
"eval_runtime": 1.9096,
"eval_samples_per_second": 69.648,
"eval_steps_per_second": 8.902,
"step": 650
},
{
"epoch": 5.0,
"step": 650,
"total_flos": 4.006371770595533e+17,
"train_loss": 0.0,
"train_runtime": 3.5904,
"train_samples_per_second": 1439.946,
"train_steps_per_second": 90.519
}
],
"logging_steps": 10,
"max_steps": 325,
"num_input_tokens_seen": 0,
"num_train_epochs": 5,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 4.006371770595533e+17,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}