Text Classification
Transformers
PyTorch
English
electra
reward-model
reward_model
RLHF
Inference Endpoints
theblackcat102's picture
Upload 12 files
ae6162c
raw
history blame
62.4 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.8507242605603262,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 1.62e-07,
"loss": 0.6957,
"step": 10
},
{
"epoch": 0.01,
"learning_rate": 3.42e-07,
"loss": 0.6958,
"step": 20
},
{
"epoch": 0.01,
"learning_rate": 5.22e-07,
"loss": 0.6916,
"step": 30
},
{
"epoch": 0.02,
"learning_rate": 7.02e-07,
"loss": 0.6959,
"step": 40
},
{
"epoch": 0.02,
"learning_rate": 8.820000000000001e-07,
"loss": 0.6927,
"step": 50
},
{
"epoch": 0.03,
"learning_rate": 1.062e-06,
"loss": 0.6921,
"step": 60
},
{
"epoch": 0.03,
"learning_rate": 1.2420000000000001e-06,
"loss": 0.691,
"step": 70
},
{
"epoch": 0.04,
"learning_rate": 1.4220000000000001e-06,
"loss": 0.6892,
"step": 80
},
{
"epoch": 0.04,
"learning_rate": 1.6019999999999999e-06,
"loss": 0.6881,
"step": 90
},
{
"epoch": 0.05,
"learning_rate": 1.782e-06,
"loss": 0.6857,
"step": 100
},
{
"epoch": 0.05,
"learning_rate": 1.962e-06,
"loss": 0.6836,
"step": 110
},
{
"epoch": 0.06,
"learning_rate": 2.142e-06,
"loss": 0.6802,
"step": 120
},
{
"epoch": 0.06,
"learning_rate": 2.322e-06,
"loss": 0.6714,
"step": 130
},
{
"epoch": 0.06,
"learning_rate": 2.4840000000000003e-06,
"loss": 0.6661,
"step": 140
},
{
"epoch": 0.07,
"learning_rate": 2.664e-06,
"loss": 0.6599,
"step": 150
},
{
"epoch": 0.07,
"learning_rate": 2.826e-06,
"loss": 0.6524,
"step": 160
},
{
"epoch": 0.08,
"learning_rate": 3.006e-06,
"loss": 0.6362,
"step": 170
},
{
"epoch": 0.08,
"learning_rate": 3.168e-06,
"loss": 0.632,
"step": 180
},
{
"epoch": 0.09,
"learning_rate": 3.348e-06,
"loss": 0.6301,
"step": 190
},
{
"epoch": 0.09,
"learning_rate": 3.5280000000000004e-06,
"loss": 0.6272,
"step": 200
},
{
"epoch": 0.1,
"learning_rate": 3.708e-06,
"loss": 0.6172,
"step": 210
},
{
"epoch": 0.1,
"learning_rate": 3.87e-06,
"loss": 0.6146,
"step": 220
},
{
"epoch": 0.11,
"learning_rate": 4.05e-06,
"loss": 0.606,
"step": 230
},
{
"epoch": 0.11,
"learning_rate": 4.23e-06,
"loss": 0.6194,
"step": 240
},
{
"epoch": 0.12,
"learning_rate": 4.392e-06,
"loss": 0.5863,
"step": 250
},
{
"epoch": 0.12,
"eval_webgpt_accuracy": 0.547752808988764,
"eval_webgpt_loss": 0.7199520468711853,
"eval_webgpt_runtime": 152.9037,
"eval_webgpt_samples_per_second": 25.611,
"eval_webgpt_steps_per_second": 2.564,
"step": 250
},
{
"epoch": 0.12,
"eval_hfsummary_accuracy": 0.630263277211861,
"eval_hfsummary_loss": 0.6598580479621887,
"eval_hfsummary_runtime": 2471.5424,
"eval_hfsummary_samples_per_second": 13.386,
"eval_hfsummary_steps_per_second": 1.339,
"step": 250
},
{
"epoch": 0.12,
"eval_gptsynthetic_accuracy": 0.9978883861236802,
"eval_gptsynthetic_loss": 0.09524902701377869,
"eval_gptsynthetic_runtime": 116.7014,
"eval_gptsynthetic_samples_per_second": 28.406,
"eval_gptsynthetic_steps_per_second": 2.845,
"step": 250
},
{
"epoch": 0.12,
"learning_rate": 4.5720000000000004e-06,
"loss": 0.5818,
"step": 260
},
{
"epoch": 0.12,
"learning_rate": 4.752e-06,
"loss": 0.5817,
"step": 270
},
{
"epoch": 0.13,
"learning_rate": 4.932e-06,
"loss": 0.5713,
"step": 280
},
{
"epoch": 0.13,
"learning_rate": 5.1119999999999995e-06,
"loss": 0.5787,
"step": 290
},
{
"epoch": 0.14,
"learning_rate": 5.2919999999999995e-06,
"loss": 0.566,
"step": 300
},
{
"epoch": 0.14,
"learning_rate": 5.472e-06,
"loss": 0.5765,
"step": 310
},
{
"epoch": 0.15,
"learning_rate": 5.652e-06,
"loss": 0.5286,
"step": 320
},
{
"epoch": 0.15,
"learning_rate": 5.814000000000001e-06,
"loss": 0.5644,
"step": 330
},
{
"epoch": 0.16,
"learning_rate": 5.9940000000000005e-06,
"loss": 0.5278,
"step": 340
},
{
"epoch": 0.16,
"learning_rate": 6.1740000000000005e-06,
"loss": 0.5376,
"step": 350
},
{
"epoch": 0.17,
"learning_rate": 6.354e-06,
"loss": 0.5396,
"step": 360
},
{
"epoch": 0.17,
"learning_rate": 6.534e-06,
"loss": 0.517,
"step": 370
},
{
"epoch": 0.18,
"learning_rate": 6.7140000000000004e-06,
"loss": 0.5001,
"step": 380
},
{
"epoch": 0.18,
"learning_rate": 6.894e-06,
"loss": 0.5069,
"step": 390
},
{
"epoch": 0.19,
"learning_rate": 7.074e-06,
"loss": 0.5102,
"step": 400
},
{
"epoch": 0.19,
"learning_rate": 7.254e-06,
"loss": 0.5105,
"step": 410
},
{
"epoch": 0.19,
"learning_rate": 7.4339999999999995e-06,
"loss": 0.5332,
"step": 420
},
{
"epoch": 0.2,
"learning_rate": 7.614e-06,
"loss": 0.5076,
"step": 430
},
{
"epoch": 0.2,
"learning_rate": 7.794e-06,
"loss": 0.5088,
"step": 440
},
{
"epoch": 0.21,
"learning_rate": 7.974e-06,
"loss": 0.4968,
"step": 450
},
{
"epoch": 0.21,
"learning_rate": 8.154e-06,
"loss": 0.4902,
"step": 460
},
{
"epoch": 0.22,
"learning_rate": 8.334e-06,
"loss": 0.5097,
"step": 470
},
{
"epoch": 0.22,
"learning_rate": 8.514e-06,
"loss": 0.5038,
"step": 480
},
{
"epoch": 0.23,
"learning_rate": 8.694e-06,
"loss": 0.5082,
"step": 490
},
{
"epoch": 0.23,
"learning_rate": 8.874e-06,
"loss": 0.4887,
"step": 500
},
{
"epoch": 0.23,
"eval_webgpt_accuracy": 0.5623084780388151,
"eval_webgpt_loss": 0.7031316161155701,
"eval_webgpt_runtime": 152.8509,
"eval_webgpt_samples_per_second": 25.62,
"eval_webgpt_steps_per_second": 2.565,
"step": 500
},
{
"epoch": 0.23,
"eval_hfsummary_accuracy": 0.6744249312335641,
"eval_hfsummary_loss": 0.6107630729675293,
"eval_hfsummary_runtime": 2471.6908,
"eval_hfsummary_samples_per_second": 13.385,
"eval_hfsummary_steps_per_second": 1.339,
"step": 500
},
{
"epoch": 0.23,
"eval_gptsynthetic_accuracy": 0.9987933634992459,
"eval_gptsynthetic_loss": 0.023942505940794945,
"eval_gptsynthetic_runtime": 116.8138,
"eval_gptsynthetic_samples_per_second": 28.379,
"eval_gptsynthetic_steps_per_second": 2.842,
"step": 500
},
{
"epoch": 0.24,
"learning_rate": 8.99293563579278e-06,
"loss": 0.4875,
"step": 510
},
{
"epoch": 0.24,
"learning_rate": 8.969387755102041e-06,
"loss": 0.483,
"step": 520
},
{
"epoch": 0.25,
"learning_rate": 8.945839874411304e-06,
"loss": 0.5057,
"step": 530
},
{
"epoch": 0.25,
"learning_rate": 8.922291993720566e-06,
"loss": 0.4923,
"step": 540
},
{
"epoch": 0.25,
"learning_rate": 8.898744113029827e-06,
"loss": 0.4623,
"step": 550
},
{
"epoch": 0.26,
"learning_rate": 8.87519623233909e-06,
"loss": 0.4735,
"step": 560
},
{
"epoch": 0.26,
"learning_rate": 8.851648351648352e-06,
"loss": 0.4516,
"step": 570
},
{
"epoch": 0.27,
"learning_rate": 8.828100470957614e-06,
"loss": 0.4542,
"step": 580
},
{
"epoch": 0.27,
"learning_rate": 8.804552590266875e-06,
"loss": 0.465,
"step": 590
},
{
"epoch": 0.28,
"learning_rate": 8.781004709576138e-06,
"loss": 0.4577,
"step": 600
},
{
"epoch": 0.28,
"learning_rate": 8.7574568288854e-06,
"loss": 0.4994,
"step": 610
},
{
"epoch": 0.29,
"learning_rate": 8.733908948194663e-06,
"loss": 0.4565,
"step": 620
},
{
"epoch": 0.29,
"learning_rate": 8.710361067503925e-06,
"loss": 0.4617,
"step": 630
},
{
"epoch": 0.3,
"learning_rate": 8.686813186813188e-06,
"loss": 0.4598,
"step": 640
},
{
"epoch": 0.3,
"learning_rate": 8.663265306122449e-06,
"loss": 0.4441,
"step": 650
},
{
"epoch": 0.31,
"learning_rate": 8.639717425431711e-06,
"loss": 0.4935,
"step": 660
},
{
"epoch": 0.31,
"learning_rate": 8.616169544740974e-06,
"loss": 0.4646,
"step": 670
},
{
"epoch": 0.31,
"learning_rate": 8.592621664050236e-06,
"loss": 0.4771,
"step": 680
},
{
"epoch": 0.32,
"learning_rate": 8.569073783359497e-06,
"loss": 0.4754,
"step": 690
},
{
"epoch": 0.32,
"learning_rate": 8.54552590266876e-06,
"loss": 0.4684,
"step": 700
},
{
"epoch": 0.33,
"learning_rate": 8.521978021978022e-06,
"loss": 0.4691,
"step": 710
},
{
"epoch": 0.33,
"learning_rate": 8.498430141287285e-06,
"loss": 0.4807,
"step": 720
},
{
"epoch": 0.34,
"learning_rate": 8.474882260596547e-06,
"loss": 0.4597,
"step": 730
},
{
"epoch": 0.34,
"learning_rate": 8.45133437990581e-06,
"loss": 0.455,
"step": 740
},
{
"epoch": 0.35,
"learning_rate": 8.42778649921507e-06,
"loss": 0.4561,
"step": 750
},
{
"epoch": 0.35,
"eval_webgpt_accuracy": 0.5702247191011236,
"eval_webgpt_loss": 0.6994287371635437,
"eval_webgpt_runtime": 152.9386,
"eval_webgpt_samples_per_second": 25.605,
"eval_webgpt_steps_per_second": 2.563,
"step": 750
},
{
"epoch": 0.35,
"eval_hfsummary_accuracy": 0.6827373575552399,
"eval_hfsummary_loss": 0.60725337266922,
"eval_hfsummary_runtime": 2470.4498,
"eval_hfsummary_samples_per_second": 13.391,
"eval_hfsummary_steps_per_second": 1.339,
"step": 750
},
{
"epoch": 0.35,
"eval_gptsynthetic_accuracy": 0.9975867269984917,
"eval_gptsynthetic_loss": 0.01110268384218216,
"eval_gptsynthetic_runtime": 116.8749,
"eval_gptsynthetic_samples_per_second": 28.364,
"eval_gptsynthetic_steps_per_second": 2.841,
"step": 750
},
{
"epoch": 0.35,
"learning_rate": 8.404238618524333e-06,
"loss": 0.4578,
"step": 760
},
{
"epoch": 0.36,
"learning_rate": 8.380690737833595e-06,
"loss": 0.4798,
"step": 770
},
{
"epoch": 0.36,
"learning_rate": 8.359497645211932e-06,
"loss": 0.4402,
"step": 780
},
{
"epoch": 0.37,
"learning_rate": 8.335949764521193e-06,
"loss": 0.4632,
"step": 790
},
{
"epoch": 0.37,
"learning_rate": 8.312401883830455e-06,
"loss": 0.4505,
"step": 800
},
{
"epoch": 0.37,
"learning_rate": 8.288854003139718e-06,
"loss": 0.4641,
"step": 810
},
{
"epoch": 0.38,
"learning_rate": 8.26530612244898e-06,
"loss": 0.4373,
"step": 820
},
{
"epoch": 0.38,
"learning_rate": 8.241758241758241e-06,
"loss": 0.4192,
"step": 830
},
{
"epoch": 0.39,
"learning_rate": 8.218210361067504e-06,
"loss": 0.4671,
"step": 840
},
{
"epoch": 0.39,
"learning_rate": 8.194662480376766e-06,
"loss": 0.4481,
"step": 850
},
{
"epoch": 0.4,
"learning_rate": 8.171114599686029e-06,
"loss": 0.4481,
"step": 860
},
{
"epoch": 0.4,
"learning_rate": 8.147566718995291e-06,
"loss": 0.4317,
"step": 870
},
{
"epoch": 0.41,
"learning_rate": 8.124018838304554e-06,
"loss": 0.4517,
"step": 880
},
{
"epoch": 0.41,
"learning_rate": 8.100470957613814e-06,
"loss": 0.4282,
"step": 890
},
{
"epoch": 0.42,
"learning_rate": 8.076923076923077e-06,
"loss": 0.4543,
"step": 900
},
{
"epoch": 0.42,
"learning_rate": 8.05337519623234e-06,
"loss": 0.4303,
"step": 910
},
{
"epoch": 0.43,
"learning_rate": 8.029827315541602e-06,
"loss": 0.4423,
"step": 920
},
{
"epoch": 0.43,
"learning_rate": 8.006279434850863e-06,
"loss": 0.4624,
"step": 930
},
{
"epoch": 0.43,
"learning_rate": 7.982731554160125e-06,
"loss": 0.4318,
"step": 940
},
{
"epoch": 0.44,
"learning_rate": 7.959183673469388e-06,
"loss": 0.4374,
"step": 950
},
{
"epoch": 0.44,
"learning_rate": 7.93563579277865e-06,
"loss": 0.4271,
"step": 960
},
{
"epoch": 0.45,
"learning_rate": 7.912087912087913e-06,
"loss": 0.4535,
"step": 970
},
{
"epoch": 0.45,
"learning_rate": 7.888540031397175e-06,
"loss": 0.4393,
"step": 980
},
{
"epoch": 0.46,
"learning_rate": 7.864992150706436e-06,
"loss": 0.4513,
"step": 990
},
{
"epoch": 0.46,
"learning_rate": 7.841444270015699e-06,
"loss": 0.4259,
"step": 1000
},
{
"epoch": 0.46,
"eval_webgpt_accuracy": 0.5720122574055159,
"eval_webgpt_loss": 0.6880369782447815,
"eval_webgpt_runtime": 153.0376,
"eval_webgpt_samples_per_second": 25.588,
"eval_webgpt_steps_per_second": 2.561,
"step": 1000
},
{
"epoch": 0.46,
"eval_hfsummary_accuracy": 0.6844905238339933,
"eval_hfsummary_loss": 0.6170333027839661,
"eval_hfsummary_runtime": 2474.9308,
"eval_hfsummary_samples_per_second": 13.367,
"eval_hfsummary_steps_per_second": 1.337,
"step": 1000
},
{
"epoch": 0.46,
"eval_gptsynthetic_accuracy": 0.9978883861236802,
"eval_gptsynthetic_loss": 0.014577150344848633,
"eval_gptsynthetic_runtime": 116.8198,
"eval_gptsynthetic_samples_per_second": 28.377,
"eval_gptsynthetic_steps_per_second": 2.842,
"step": 1000
},
{
"epoch": 0.47,
"learning_rate": 7.817896389324961e-06,
"loss": 0.449,
"step": 1010
},
{
"epoch": 0.47,
"learning_rate": 7.794348508634224e-06,
"loss": 0.4274,
"step": 1020
},
{
"epoch": 0.48,
"learning_rate": 7.770800627943485e-06,
"loss": 0.4014,
"step": 1030
},
{
"epoch": 0.48,
"learning_rate": 7.747252747252747e-06,
"loss": 0.4356,
"step": 1040
},
{
"epoch": 0.49,
"learning_rate": 7.72370486656201e-06,
"loss": 0.4378,
"step": 1050
},
{
"epoch": 0.49,
"learning_rate": 7.700156985871272e-06,
"loss": 0.4298,
"step": 1060
},
{
"epoch": 0.49,
"learning_rate": 7.676609105180535e-06,
"loss": 0.4514,
"step": 1070
},
{
"epoch": 0.5,
"learning_rate": 7.653061224489796e-06,
"loss": 0.4526,
"step": 1080
},
{
"epoch": 0.5,
"learning_rate": 7.629513343799058e-06,
"loss": 0.4192,
"step": 1090
},
{
"epoch": 0.51,
"learning_rate": 7.60596546310832e-06,
"loss": 0.4433,
"step": 1100
},
{
"epoch": 0.51,
"learning_rate": 7.582417582417582e-06,
"loss": 0.4199,
"step": 1110
},
{
"epoch": 0.52,
"learning_rate": 7.558869701726845e-06,
"loss": 0.4419,
"step": 1120
},
{
"epoch": 0.52,
"learning_rate": 7.535321821036107e-06,
"loss": 0.4163,
"step": 1130
},
{
"epoch": 0.53,
"learning_rate": 7.511773940345369e-06,
"loss": 0.411,
"step": 1140
},
{
"epoch": 0.53,
"learning_rate": 7.4882260596546314e-06,
"loss": 0.4492,
"step": 1150
},
{
"epoch": 0.54,
"learning_rate": 7.464678178963893e-06,
"loss": 0.454,
"step": 1160
},
{
"epoch": 0.54,
"learning_rate": 7.441130298273156e-06,
"loss": 0.4311,
"step": 1170
},
{
"epoch": 0.55,
"learning_rate": 7.417582417582417e-06,
"loss": 0.4368,
"step": 1180
},
{
"epoch": 0.55,
"learning_rate": 7.39403453689168e-06,
"loss": 0.4264,
"step": 1190
},
{
"epoch": 0.56,
"learning_rate": 7.3704866562009415e-06,
"loss": 0.4163,
"step": 1200
},
{
"epoch": 0.56,
"learning_rate": 7.346938775510204e-06,
"loss": 0.4214,
"step": 1210
},
{
"epoch": 0.56,
"learning_rate": 7.3233908948194665e-06,
"loss": 0.4263,
"step": 1220
},
{
"epoch": 0.57,
"learning_rate": 7.299843014128729e-06,
"loss": 0.4233,
"step": 1230
},
{
"epoch": 0.57,
"learning_rate": 7.276295133437991e-06,
"loss": 0.4331,
"step": 1240
},
{
"epoch": 0.58,
"learning_rate": 7.252747252747253e-06,
"loss": 0.4241,
"step": 1250
},
{
"epoch": 0.58,
"eval_webgpt_accuracy": 0.5806945863125639,
"eval_webgpt_loss": 0.6766389608383179,
"eval_webgpt_runtime": 153.0189,
"eval_webgpt_samples_per_second": 25.592,
"eval_webgpt_steps_per_second": 2.562,
"step": 1250
},
{
"epoch": 0.58,
"eval_hfsummary_accuracy": 0.6927122691412508,
"eval_hfsummary_loss": 0.6117472648620605,
"eval_hfsummary_runtime": 2473.4013,
"eval_hfsummary_samples_per_second": 13.376,
"eval_hfsummary_steps_per_second": 1.338,
"step": 1250
},
{
"epoch": 0.58,
"eval_gptsynthetic_accuracy": 0.9987933634992459,
"eval_gptsynthetic_loss": 0.007306403946131468,
"eval_gptsynthetic_runtime": 116.8034,
"eval_gptsynthetic_samples_per_second": 28.381,
"eval_gptsynthetic_steps_per_second": 2.842,
"step": 1250
},
{
"epoch": 0.58,
"learning_rate": 7.229199372056515e-06,
"loss": 0.3976,
"step": 1260
},
{
"epoch": 0.59,
"learning_rate": 7.205651491365777e-06,
"loss": 0.4221,
"step": 1270
},
{
"epoch": 0.59,
"learning_rate": 7.182103610675039e-06,
"loss": 0.4225,
"step": 1280
},
{
"epoch": 0.6,
"learning_rate": 7.158555729984302e-06,
"loss": 0.3961,
"step": 1290
},
{
"epoch": 0.6,
"learning_rate": 7.135007849293563e-06,
"loss": 0.4348,
"step": 1300
},
{
"epoch": 0.61,
"learning_rate": 7.111459968602826e-06,
"loss": 0.4146,
"step": 1310
},
{
"epoch": 0.61,
"learning_rate": 7.087912087912088e-06,
"loss": 0.4122,
"step": 1320
},
{
"epoch": 0.62,
"learning_rate": 7.064364207221351e-06,
"loss": 0.4126,
"step": 1330
},
{
"epoch": 0.62,
"learning_rate": 7.0408163265306125e-06,
"loss": 0.4,
"step": 1340
},
{
"epoch": 0.62,
"learning_rate": 7.017268445839875e-06,
"loss": 0.3939,
"step": 1350
},
{
"epoch": 0.63,
"learning_rate": 6.993720565149137e-06,
"loss": 0.4144,
"step": 1360
},
{
"epoch": 0.63,
"learning_rate": 6.970172684458399e-06,
"loss": 0.3895,
"step": 1370
},
{
"epoch": 0.64,
"learning_rate": 6.946624803767661e-06,
"loss": 0.431,
"step": 1380
},
{
"epoch": 0.64,
"learning_rate": 6.923076923076923e-06,
"loss": 0.4156,
"step": 1390
},
{
"epoch": 0.65,
"learning_rate": 6.899529042386185e-06,
"loss": 0.4317,
"step": 1400
},
{
"epoch": 0.65,
"learning_rate": 6.8759811616954476e-06,
"loss": 0.4142,
"step": 1410
},
{
"epoch": 0.66,
"learning_rate": 6.85243328100471e-06,
"loss": 0.393,
"step": 1420
},
{
"epoch": 0.66,
"learning_rate": 6.828885400313973e-06,
"loss": 0.4156,
"step": 1430
},
{
"epoch": 0.67,
"learning_rate": 6.805337519623234e-06,
"loss": 0.3995,
"step": 1440
},
{
"epoch": 0.67,
"learning_rate": 6.781789638932497e-06,
"loss": 0.4084,
"step": 1450
},
{
"epoch": 0.68,
"learning_rate": 6.7582417582417585e-06,
"loss": 0.4162,
"step": 1460
},
{
"epoch": 0.68,
"learning_rate": 6.734693877551021e-06,
"loss": 0.4167,
"step": 1470
},
{
"epoch": 0.68,
"learning_rate": 6.711145996860283e-06,
"loss": 0.3878,
"step": 1480
},
{
"epoch": 0.69,
"learning_rate": 6.687598116169544e-06,
"loss": 0.403,
"step": 1490
},
{
"epoch": 0.69,
"learning_rate": 6.664050235478807e-06,
"loss": 0.3988,
"step": 1500
},
{
"epoch": 0.69,
"eval_webgpt_accuracy": 0.5840143003064351,
"eval_webgpt_loss": 0.6755279302597046,
"eval_webgpt_runtime": 152.9789,
"eval_webgpt_samples_per_second": 25.598,
"eval_webgpt_steps_per_second": 2.562,
"step": 1500
},
{
"epoch": 0.69,
"eval_hfsummary_accuracy": 0.6905661518000181,
"eval_hfsummary_loss": 0.6140123605728149,
"eval_hfsummary_runtime": 2473.83,
"eval_hfsummary_samples_per_second": 13.373,
"eval_hfsummary_steps_per_second": 1.338,
"step": 1500
},
{
"epoch": 0.69,
"eval_gptsynthetic_accuracy": 0.9978883861236802,
"eval_gptsynthetic_loss": 0.008705263957381248,
"eval_gptsynthetic_runtime": 116.82,
"eval_gptsynthetic_samples_per_second": 28.377,
"eval_gptsynthetic_steps_per_second": 2.842,
"step": 1500
},
{
"epoch": 0.7,
"learning_rate": 6.6405023547880685e-06,
"loss": 0.4004,
"step": 1510
},
{
"epoch": 0.7,
"learning_rate": 6.616954474097332e-06,
"loss": 0.3856,
"step": 1520
},
{
"epoch": 0.71,
"learning_rate": 6.5934065934065935e-06,
"loss": 0.4168,
"step": 1530
},
{
"epoch": 0.71,
"learning_rate": 6.569858712715856e-06,
"loss": 0.401,
"step": 1540
},
{
"epoch": 0.72,
"learning_rate": 6.546310832025118e-06,
"loss": 0.4038,
"step": 1550
},
{
"epoch": 0.72,
"learning_rate": 6.52276295133438e-06,
"loss": 0.3804,
"step": 1560
},
{
"epoch": 0.73,
"learning_rate": 6.499215070643642e-06,
"loss": 0.3787,
"step": 1570
},
{
"epoch": 0.73,
"learning_rate": 6.4756671899529044e-06,
"loss": 0.4003,
"step": 1580
},
{
"epoch": 0.74,
"learning_rate": 6.452119309262166e-06,
"loss": 0.3954,
"step": 1590
},
{
"epoch": 0.74,
"learning_rate": 6.428571428571429e-06,
"loss": 0.4011,
"step": 1600
},
{
"epoch": 0.74,
"learning_rate": 6.40502354788069e-06,
"loss": 0.3882,
"step": 1610
},
{
"epoch": 0.75,
"learning_rate": 6.381475667189954e-06,
"loss": 0.3954,
"step": 1620
},
{
"epoch": 0.75,
"learning_rate": 6.357927786499215e-06,
"loss": 0.3788,
"step": 1630
},
{
"epoch": 0.76,
"learning_rate": 6.334379905808478e-06,
"loss": 0.3921,
"step": 1640
},
{
"epoch": 0.76,
"learning_rate": 6.3108320251177395e-06,
"loss": 0.39,
"step": 1650
},
{
"epoch": 0.77,
"learning_rate": 6.287284144427002e-06,
"loss": 0.4137,
"step": 1660
},
{
"epoch": 0.77,
"learning_rate": 6.263736263736264e-06,
"loss": 0.4032,
"step": 1670
},
{
"epoch": 0.78,
"learning_rate": 6.240188383045526e-06,
"loss": 0.3865,
"step": 1680
},
{
"epoch": 0.78,
"learning_rate": 6.216640502354788e-06,
"loss": 0.3971,
"step": 1690
},
{
"epoch": 0.79,
"learning_rate": 6.19309262166405e-06,
"loss": 0.3585,
"step": 1700
},
{
"epoch": 0.79,
"learning_rate": 6.169544740973312e-06,
"loss": 0.3933,
"step": 1710
},
{
"epoch": 0.8,
"learning_rate": 6.1459968602825754e-06,
"loss": 0.4003,
"step": 1720
},
{
"epoch": 0.8,
"learning_rate": 6.122448979591837e-06,
"loss": 0.3931,
"step": 1730
},
{
"epoch": 0.8,
"learning_rate": 6.0989010989011e-06,
"loss": 0.3751,
"step": 1740
},
{
"epoch": 0.81,
"learning_rate": 6.075353218210361e-06,
"loss": 0.3793,
"step": 1750
},
{
"epoch": 0.81,
"eval_webgpt_accuracy": 0.593973442288049,
"eval_webgpt_loss": 0.6751279830932617,
"eval_webgpt_runtime": 152.9792,
"eval_webgpt_samples_per_second": 25.598,
"eval_webgpt_steps_per_second": 2.562,
"step": 1750
},
{
"epoch": 0.81,
"eval_hfsummary_accuracy": 0.6901127467279268,
"eval_hfsummary_loss": 0.655200183391571,
"eval_hfsummary_runtime": 2476.3756,
"eval_hfsummary_samples_per_second": 13.359,
"eval_hfsummary_steps_per_second": 1.336,
"step": 1750
},
{
"epoch": 0.81,
"eval_gptsynthetic_accuracy": 0.9978883861236802,
"eval_gptsynthetic_loss": 0.007934121415019035,
"eval_gptsynthetic_runtime": 116.9338,
"eval_gptsynthetic_samples_per_second": 28.349,
"eval_gptsynthetic_steps_per_second": 2.839,
"step": 1750
},
{
"epoch": 0.81,
"learning_rate": 6.051805337519624e-06,
"loss": 0.3883,
"step": 1760
},
{
"epoch": 0.82,
"learning_rate": 6.0282574568288855e-06,
"loss": 0.3918,
"step": 1770
},
{
"epoch": 0.82,
"learning_rate": 6.004709576138148e-06,
"loss": 0.4086,
"step": 1780
},
{
"epoch": 0.83,
"learning_rate": 5.98116169544741e-06,
"loss": 0.3876,
"step": 1790
},
{
"epoch": 0.83,
"learning_rate": 5.957613814756672e-06,
"loss": 0.3613,
"step": 1800
},
{
"epoch": 0.84,
"learning_rate": 5.934065934065934e-06,
"loss": 0.3715,
"step": 1810
},
{
"epoch": 0.84,
"learning_rate": 5.910518053375197e-06,
"loss": 0.3879,
"step": 1820
},
{
"epoch": 0.85,
"learning_rate": 5.886970172684459e-06,
"loss": 0.3893,
"step": 1830
},
{
"epoch": 0.85,
"learning_rate": 5.863422291993721e-06,
"loss": 0.34,
"step": 1840
},
{
"epoch": 0.86,
"learning_rate": 5.839874411302983e-06,
"loss": 0.3938,
"step": 1850
},
{
"epoch": 0.86,
"learning_rate": 5.816326530612246e-06,
"loss": 0.3802,
"step": 1860
},
{
"epoch": 0.87,
"learning_rate": 5.792778649921507e-06,
"loss": 0.3869,
"step": 1870
},
{
"epoch": 0.87,
"learning_rate": 5.76923076923077e-06,
"loss": 0.4023,
"step": 1880
},
{
"epoch": 0.87,
"learning_rate": 5.7456828885400314e-06,
"loss": 0.3694,
"step": 1890
},
{
"epoch": 0.88,
"learning_rate": 5.722135007849293e-06,
"loss": 0.358,
"step": 1900
},
{
"epoch": 0.88,
"learning_rate": 5.698587127158556e-06,
"loss": 0.3939,
"step": 1910
},
{
"epoch": 0.89,
"learning_rate": 5.675039246467817e-06,
"loss": 0.3805,
"step": 1920
},
{
"epoch": 0.89,
"learning_rate": 5.651491365777081e-06,
"loss": 0.3754,
"step": 1930
},
{
"epoch": 0.9,
"learning_rate": 5.627943485086342e-06,
"loss": 0.3454,
"step": 1940
},
{
"epoch": 0.9,
"learning_rate": 5.604395604395605e-06,
"loss": 0.4109,
"step": 1950
},
{
"epoch": 0.91,
"learning_rate": 5.5808477237048665e-06,
"loss": 0.364,
"step": 1960
},
{
"epoch": 0.91,
"learning_rate": 5.557299843014129e-06,
"loss": 0.3869,
"step": 1970
},
{
"epoch": 0.92,
"learning_rate": 5.533751962323391e-06,
"loss": 0.3858,
"step": 1980
},
{
"epoch": 0.92,
"learning_rate": 5.510204081632653e-06,
"loss": 0.3634,
"step": 1990
},
{
"epoch": 0.93,
"learning_rate": 5.486656200941915e-06,
"loss": 0.3444,
"step": 2000
},
{
"epoch": 0.93,
"eval_webgpt_accuracy": 0.5916751787538305,
"eval_webgpt_loss": 0.6755998134613037,
"eval_webgpt_runtime": 153.0377,
"eval_webgpt_samples_per_second": 25.588,
"eval_webgpt_steps_per_second": 2.561,
"step": 2000
},
{
"epoch": 0.93,
"eval_hfsummary_accuracy": 0.6894477526221927,
"eval_hfsummary_loss": 0.6544569730758667,
"eval_hfsummary_runtime": 2482.2617,
"eval_hfsummary_samples_per_second": 13.328,
"eval_hfsummary_steps_per_second": 1.333,
"step": 2000
},
{
"epoch": 0.93,
"eval_gptsynthetic_accuracy": 0.9975867269984917,
"eval_gptsynthetic_loss": 0.00854982528835535,
"eval_gptsynthetic_runtime": 117.0015,
"eval_gptsynthetic_samples_per_second": 28.333,
"eval_gptsynthetic_steps_per_second": 2.838,
"step": 2000
},
{
"epoch": 0.93,
"learning_rate": 5.463108320251177e-06,
"loss": 0.3548,
"step": 2010
},
{
"epoch": 0.93,
"learning_rate": 5.439560439560439e-06,
"loss": 0.3849,
"step": 2020
},
{
"epoch": 0.94,
"learning_rate": 5.4160125588697024e-06,
"loss": 0.3732,
"step": 2030
},
{
"epoch": 0.94,
"learning_rate": 5.392464678178964e-06,
"loss": 0.3589,
"step": 2040
},
{
"epoch": 0.95,
"learning_rate": 5.368916797488227e-06,
"loss": 0.3595,
"step": 2050
},
{
"epoch": 0.95,
"learning_rate": 5.345368916797488e-06,
"loss": 0.3704,
"step": 2060
},
{
"epoch": 0.96,
"learning_rate": 5.324175824175825e-06,
"loss": 0.3846,
"step": 2070
},
{
"epoch": 0.96,
"learning_rate": 5.300627943485086e-06,
"loss": 0.3862,
"step": 2080
},
{
"epoch": 0.97,
"learning_rate": 5.277080062794349e-06,
"loss": 0.3956,
"step": 2090
},
{
"epoch": 0.97,
"learning_rate": 5.253532182103611e-06,
"loss": 0.3629,
"step": 2100
},
{
"epoch": 0.98,
"learning_rate": 5.229984301412873e-06,
"loss": 0.3674,
"step": 2110
},
{
"epoch": 0.98,
"learning_rate": 5.206436420722135e-06,
"loss": 0.3827,
"step": 2120
},
{
"epoch": 0.99,
"learning_rate": 5.182888540031397e-06,
"loss": 0.3751,
"step": 2130
},
{
"epoch": 0.99,
"learning_rate": 5.159340659340659e-06,
"loss": 0.3834,
"step": 2140
},
{
"epoch": 0.99,
"learning_rate": 5.1357927786499215e-06,
"loss": 0.3425,
"step": 2150
},
{
"epoch": 1.0,
"learning_rate": 5.112244897959183e-06,
"loss": 0.3731,
"step": 2160
},
{
"epoch": 1.0,
"learning_rate": 5.0886970172684465e-06,
"loss": 0.3724,
"step": 2170
},
{
"epoch": 1.01,
"learning_rate": 5.065149136577708e-06,
"loss": 0.3842,
"step": 2180
},
{
"epoch": 1.01,
"learning_rate": 5.041601255886971e-06,
"loss": 0.3661,
"step": 2190
},
{
"epoch": 1.02,
"learning_rate": 5.018053375196232e-06,
"loss": 0.3795,
"step": 2200
},
{
"epoch": 1.02,
"learning_rate": 4.994505494505495e-06,
"loss": 0.3766,
"step": 2210
},
{
"epoch": 1.03,
"learning_rate": 4.9709576138147566e-06,
"loss": 0.3641,
"step": 2220
},
{
"epoch": 1.03,
"learning_rate": 4.947409733124019e-06,
"loss": 0.355,
"step": 2230
},
{
"epoch": 1.04,
"learning_rate": 4.923861852433281e-06,
"loss": 0.3481,
"step": 2240
},
{
"epoch": 1.04,
"learning_rate": 4.900313971742543e-06,
"loss": 0.312,
"step": 2250
},
{
"epoch": 1.04,
"eval_webgpt_accuracy": 0.5970377936670072,
"eval_webgpt_loss": 0.6743035912513733,
"eval_webgpt_runtime": 153.349,
"eval_webgpt_samples_per_second": 25.537,
"eval_webgpt_steps_per_second": 2.556,
"step": 2250
},
{
"epoch": 1.04,
"eval_hfsummary_accuracy": 0.6958558776410846,
"eval_hfsummary_loss": 0.6904739141464233,
"eval_hfsummary_runtime": 2483.5335,
"eval_hfsummary_samples_per_second": 13.321,
"eval_hfsummary_steps_per_second": 1.332,
"step": 2250
},
{
"epoch": 1.04,
"eval_gptsynthetic_accuracy": 0.9972850678733032,
"eval_gptsynthetic_loss": 0.011290821246802807,
"eval_gptsynthetic_runtime": 117.0341,
"eval_gptsynthetic_samples_per_second": 28.325,
"eval_gptsynthetic_steps_per_second": 2.837,
"step": 2250
},
{
"epoch": 1.05,
"learning_rate": 4.876766091051805e-06,
"loss": 0.3593,
"step": 2260
},
{
"epoch": 1.05,
"learning_rate": 4.853218210361068e-06,
"loss": 0.3378,
"step": 2270
},
{
"epoch": 1.06,
"learning_rate": 4.82967032967033e-06,
"loss": 0.3522,
"step": 2280
},
{
"epoch": 1.06,
"learning_rate": 4.8061224489795925e-06,
"loss": 0.3732,
"step": 2290
},
{
"epoch": 1.06,
"learning_rate": 4.782574568288854e-06,
"loss": 0.3616,
"step": 2300
},
{
"epoch": 1.07,
"learning_rate": 4.759026687598117e-06,
"loss": 0.3478,
"step": 2310
},
{
"epoch": 1.07,
"learning_rate": 4.735478806907378e-06,
"loss": 0.3721,
"step": 2320
},
{
"epoch": 1.08,
"learning_rate": 4.711930926216641e-06,
"loss": 0.3314,
"step": 2330
},
{
"epoch": 1.08,
"learning_rate": 4.6883830455259025e-06,
"loss": 0.3413,
"step": 2340
},
{
"epoch": 1.09,
"learning_rate": 4.664835164835165e-06,
"loss": 0.3266,
"step": 2350
},
{
"epoch": 1.09,
"learning_rate": 4.641287284144427e-06,
"loss": 0.3579,
"step": 2360
},
{
"epoch": 1.1,
"learning_rate": 4.61773940345369e-06,
"loss": 0.3549,
"step": 2370
},
{
"epoch": 1.1,
"learning_rate": 4.594191522762952e-06,
"loss": 0.347,
"step": 2380
},
{
"epoch": 1.11,
"learning_rate": 4.570643642072214e-06,
"loss": 0.3453,
"step": 2390
},
{
"epoch": 1.11,
"learning_rate": 4.547095761381476e-06,
"loss": 0.3411,
"step": 2400
},
{
"epoch": 1.12,
"learning_rate": 4.5235478806907384e-06,
"loss": 0.3411,
"step": 2410
},
{
"epoch": 1.12,
"learning_rate": 4.5e-06,
"loss": 0.361,
"step": 2420
},
{
"epoch": 1.12,
"learning_rate": 4.476452119309263e-06,
"loss": 0.3576,
"step": 2430
},
{
"epoch": 1.13,
"learning_rate": 4.452904238618524e-06,
"loss": 0.3476,
"step": 2440
},
{
"epoch": 1.13,
"learning_rate": 4.429356357927787e-06,
"loss": 0.3382,
"step": 2450
},
{
"epoch": 1.14,
"learning_rate": 4.4058084772370485e-06,
"loss": 0.3485,
"step": 2460
},
{
"epoch": 1.14,
"learning_rate": 4.382260596546311e-06,
"loss": 0.3237,
"step": 2470
},
{
"epoch": 1.15,
"learning_rate": 4.3587127158555735e-06,
"loss": 0.3385,
"step": 2480
},
{
"epoch": 1.15,
"learning_rate": 4.335164835164835e-06,
"loss": 0.3124,
"step": 2490
},
{
"epoch": 1.16,
"learning_rate": 4.311616954474098e-06,
"loss": 0.3385,
"step": 2500
},
{
"epoch": 1.16,
"eval_webgpt_accuracy": 0.589632277834525,
"eval_webgpt_loss": 0.6730017066001892,
"eval_webgpt_runtime": 153.2835,
"eval_webgpt_samples_per_second": 25.547,
"eval_webgpt_steps_per_second": 2.557,
"step": 2500
},
{
"epoch": 1.16,
"eval_hfsummary_accuracy": 0.6939515763383006,
"eval_hfsummary_loss": 0.6676614880561829,
"eval_hfsummary_runtime": 2484.0871,
"eval_hfsummary_samples_per_second": 13.318,
"eval_hfsummary_steps_per_second": 1.332,
"step": 2500
},
{
"epoch": 1.16,
"eval_gptsynthetic_accuracy": 0.9987933634992459,
"eval_gptsynthetic_loss": 0.008312013931572437,
"eval_gptsynthetic_runtime": 117.0521,
"eval_gptsynthetic_samples_per_second": 28.321,
"eval_gptsynthetic_steps_per_second": 2.836,
"step": 2500
},
{
"epoch": 1.16,
"learning_rate": 4.288069073783359e-06,
"loss": 0.3575,
"step": 2510
},
{
"epoch": 1.17,
"learning_rate": 4.264521193092622e-06,
"loss": 0.3562,
"step": 2520
},
{
"epoch": 1.17,
"learning_rate": 4.240973312401884e-06,
"loss": 0.3296,
"step": 2530
},
{
"epoch": 1.18,
"learning_rate": 4.217425431711146e-06,
"loss": 0.3447,
"step": 2540
},
{
"epoch": 1.18,
"learning_rate": 4.193877551020409e-06,
"loss": 0.3373,
"step": 2550
},
{
"epoch": 1.18,
"learning_rate": 4.17032967032967e-06,
"loss": 0.3394,
"step": 2560
},
{
"epoch": 1.19,
"learning_rate": 4.146781789638933e-06,
"loss": 0.3294,
"step": 2570
},
{
"epoch": 1.19,
"learning_rate": 4.123233908948195e-06,
"loss": 0.3715,
"step": 2580
},
{
"epoch": 1.2,
"learning_rate": 4.099686028257457e-06,
"loss": 0.3356,
"step": 2590
},
{
"epoch": 1.2,
"learning_rate": 4.0761381475667195e-06,
"loss": 0.3497,
"step": 2600
},
{
"epoch": 1.21,
"learning_rate": 4.052590266875981e-06,
"loss": 0.3189,
"step": 2610
},
{
"epoch": 1.21,
"learning_rate": 4.029042386185244e-06,
"loss": 0.3473,
"step": 2620
},
{
"epoch": 1.22,
"learning_rate": 4.005494505494505e-06,
"loss": 0.3259,
"step": 2630
},
{
"epoch": 1.22,
"learning_rate": 3.981946624803768e-06,
"loss": 0.3294,
"step": 2640
},
{
"epoch": 1.23,
"learning_rate": 3.9583987441130295e-06,
"loss": 0.3294,
"step": 2650
},
{
"epoch": 1.23,
"learning_rate": 3.934850863422292e-06,
"loss": 0.3221,
"step": 2660
},
{
"epoch": 1.24,
"learning_rate": 3.911302982731554e-06,
"loss": 0.3698,
"step": 2670
},
{
"epoch": 1.24,
"learning_rate": 3.887755102040816e-06,
"loss": 0.3514,
"step": 2680
},
{
"epoch": 1.24,
"learning_rate": 3.864207221350079e-06,
"loss": 0.3461,
"step": 2690
},
{
"epoch": 1.25,
"learning_rate": 3.84065934065934e-06,
"loss": 0.3321,
"step": 2700
},
{
"epoch": 1.25,
"learning_rate": 3.817111459968603e-06,
"loss": 0.3241,
"step": 2710
},
{
"epoch": 1.26,
"learning_rate": 3.793563579277865e-06,
"loss": 0.3466,
"step": 2720
},
{
"epoch": 1.26,
"learning_rate": 3.770015698587127e-06,
"loss": 0.3188,
"step": 2730
},
{
"epoch": 1.27,
"learning_rate": 3.7464678178963892e-06,
"loss": 0.32,
"step": 2740
},
{
"epoch": 1.27,
"learning_rate": 3.7229199372056513e-06,
"loss": 0.3383,
"step": 2750
},
{
"epoch": 1.27,
"eval_webgpt_accuracy": 0.5776302349336058,
"eval_webgpt_loss": 0.6770792603492737,
"eval_webgpt_runtime": 153.2414,
"eval_webgpt_samples_per_second": 25.554,
"eval_webgpt_steps_per_second": 2.558,
"step": 2750
},
{
"epoch": 1.27,
"eval_hfsummary_accuracy": 0.6819212284254753,
"eval_hfsummary_loss": 0.7081930041313171,
"eval_hfsummary_runtime": 2484.801,
"eval_hfsummary_samples_per_second": 13.314,
"eval_hfsummary_steps_per_second": 1.332,
"step": 2750
},
{
"epoch": 1.27,
"eval_gptsynthetic_accuracy": 0.9984917043740573,
"eval_gptsynthetic_loss": 0.006183410994708538,
"eval_gptsynthetic_runtime": 117.1145,
"eval_gptsynthetic_samples_per_second": 28.306,
"eval_gptsynthetic_steps_per_second": 2.835,
"step": 2750
},
{
"epoch": 1.28,
"learning_rate": 3.699372056514914e-06,
"loss": 0.3136,
"step": 2760
},
{
"epoch": 1.28,
"learning_rate": 3.675824175824176e-06,
"loss": 0.3358,
"step": 2770
},
{
"epoch": 1.29,
"learning_rate": 3.652276295133438e-06,
"loss": 0.3612,
"step": 2780
},
{
"epoch": 1.29,
"learning_rate": 3.6287284144427e-06,
"loss": 0.319,
"step": 2790
},
{
"epoch": 1.3,
"learning_rate": 3.605180533751962e-06,
"loss": 0.3039,
"step": 2800
},
{
"epoch": 1.3,
"learning_rate": 3.5816326530612247e-06,
"loss": 0.3383,
"step": 2810
},
{
"epoch": 1.3,
"learning_rate": 3.558084772370487e-06,
"loss": 0.3096,
"step": 2820
},
{
"epoch": 1.31,
"learning_rate": 3.534536891679749e-06,
"loss": 0.3377,
"step": 2830
},
{
"epoch": 1.31,
"learning_rate": 3.510989010989011e-06,
"loss": 0.314,
"step": 2840
},
{
"epoch": 1.32,
"learning_rate": 3.487441130298273e-06,
"loss": 0.344,
"step": 2850
},
{
"epoch": 1.32,
"learning_rate": 3.4638932496075356e-06,
"loss": 0.3158,
"step": 2860
},
{
"epoch": 1.33,
"learning_rate": 3.4403453689167977e-06,
"loss": 0.3277,
"step": 2870
},
{
"epoch": 1.33,
"learning_rate": 3.41679748822606e-06,
"loss": 0.3292,
"step": 2880
},
{
"epoch": 1.34,
"learning_rate": 3.393249607535322e-06,
"loss": 0.3479,
"step": 2890
},
{
"epoch": 1.34,
"learning_rate": 3.369701726844584e-06,
"loss": 0.338,
"step": 2900
},
{
"epoch": 1.35,
"learning_rate": 3.3461538461538465e-06,
"loss": 0.3388,
"step": 2910
},
{
"epoch": 1.35,
"learning_rate": 3.3226059654631086e-06,
"loss": 0.2873,
"step": 2920
},
{
"epoch": 1.36,
"learning_rate": 3.2990580847723707e-06,
"loss": 0.3147,
"step": 2930
},
{
"epoch": 1.36,
"learning_rate": 3.2755102040816328e-06,
"loss": 0.3423,
"step": 2940
},
{
"epoch": 1.36,
"learning_rate": 3.251962323390895e-06,
"loss": 0.3052,
"step": 2950
},
{
"epoch": 1.37,
"learning_rate": 3.2284144427001574e-06,
"loss": 0.3308,
"step": 2960
},
{
"epoch": 1.37,
"learning_rate": 3.2048665620094195e-06,
"loss": 0.3484,
"step": 2970
},
{
"epoch": 1.38,
"learning_rate": 3.1813186813186816e-06,
"loss": 0.3287,
"step": 2980
},
{
"epoch": 1.38,
"learning_rate": 3.1577708006279437e-06,
"loss": 0.2954,
"step": 2990
},
{
"epoch": 1.39,
"learning_rate": 3.1342229199372058e-06,
"loss": 0.3284,
"step": 3000
},
{
"epoch": 1.39,
"eval_webgpt_accuracy": 0.5832482124616956,
"eval_webgpt_loss": 0.6810373067855835,
"eval_webgpt_runtime": 153.3795,
"eval_webgpt_samples_per_second": 25.531,
"eval_webgpt_steps_per_second": 2.556,
"step": 3000
},
{
"epoch": 1.39,
"eval_hfsummary_accuracy": 0.6889641205452952,
"eval_hfsummary_loss": 0.7132372856140137,
"eval_hfsummary_runtime": 2481.1392,
"eval_hfsummary_samples_per_second": 13.334,
"eval_hfsummary_steps_per_second": 1.334,
"step": 3000
},
{
"epoch": 1.39,
"eval_gptsynthetic_accuracy": 0.9987933634992459,
"eval_gptsynthetic_loss": 0.005898704752326012,
"eval_gptsynthetic_runtime": 116.9777,
"eval_gptsynthetic_samples_per_second": 28.339,
"eval_gptsynthetic_steps_per_second": 2.838,
"step": 3000
},
{
"epoch": 1.39,
"learning_rate": 3.1106750392464683e-06,
"loss": 0.3138,
"step": 3010
},
{
"epoch": 1.4,
"learning_rate": 3.0871271585557304e-06,
"loss": 0.3364,
"step": 3020
},
{
"epoch": 1.4,
"learning_rate": 3.063579277864992e-06,
"loss": 0.3439,
"step": 3030
},
{
"epoch": 1.41,
"learning_rate": 3.040031397174254e-06,
"loss": 0.3489,
"step": 3040
},
{
"epoch": 1.41,
"learning_rate": 3.0164835164835162e-06,
"loss": 0.3419,
"step": 3050
},
{
"epoch": 1.42,
"learning_rate": 2.9929356357927783e-06,
"loss": 0.3052,
"step": 3060
},
{
"epoch": 1.42,
"learning_rate": 2.969387755102041e-06,
"loss": 0.3126,
"step": 3070
},
{
"epoch": 1.43,
"learning_rate": 2.945839874411303e-06,
"loss": 0.3475,
"step": 3080
},
{
"epoch": 1.43,
"learning_rate": 2.922291993720565e-06,
"loss": 0.3444,
"step": 3090
},
{
"epoch": 1.43,
"learning_rate": 2.898744113029827e-06,
"loss": 0.3279,
"step": 3100
},
{
"epoch": 1.44,
"learning_rate": 2.8751962323390892e-06,
"loss": 0.3315,
"step": 3110
},
{
"epoch": 1.44,
"learning_rate": 2.8516483516483517e-06,
"loss": 0.3238,
"step": 3120
},
{
"epoch": 1.45,
"learning_rate": 2.828100470957614e-06,
"loss": 0.3291,
"step": 3130
},
{
"epoch": 1.45,
"learning_rate": 2.804552590266876e-06,
"loss": 0.3014,
"step": 3140
},
{
"epoch": 1.46,
"learning_rate": 2.781004709576138e-06,
"loss": 0.3397,
"step": 3150
},
{
"epoch": 1.46,
"learning_rate": 2.7574568288854e-06,
"loss": 0.3143,
"step": 3160
},
{
"epoch": 1.47,
"learning_rate": 2.7339089481946626e-06,
"loss": 0.3317,
"step": 3170
},
{
"epoch": 1.47,
"learning_rate": 2.7103610675039247e-06,
"loss": 0.3522,
"step": 3180
},
{
"epoch": 1.48,
"learning_rate": 2.686813186813187e-06,
"loss": 0.3324,
"step": 3190
},
{
"epoch": 1.48,
"learning_rate": 2.663265306122449e-06,
"loss": 0.3043,
"step": 3200
},
{
"epoch": 1.49,
"learning_rate": 2.639717425431711e-06,
"loss": 0.3424,
"step": 3210
},
{
"epoch": 1.49,
"learning_rate": 2.6161695447409735e-06,
"loss": 0.3461,
"step": 3220
},
{
"epoch": 1.49,
"learning_rate": 2.5926216640502356e-06,
"loss": 0.3289,
"step": 3230
},
{
"epoch": 1.5,
"learning_rate": 2.5690737833594977e-06,
"loss": 0.3057,
"step": 3240
},
{
"epoch": 1.5,
"learning_rate": 2.54552590266876e-06,
"loss": 0.3118,
"step": 3250
},
{
"epoch": 1.5,
"eval_webgpt_accuracy": 0.5847803881511746,
"eval_webgpt_loss": 0.6773096919059753,
"eval_webgpt_runtime": 153.2012,
"eval_webgpt_samples_per_second": 25.561,
"eval_webgpt_steps_per_second": 2.559,
"step": 3250
},
{
"epoch": 1.5,
"eval_hfsummary_accuracy": 0.6834628056705861,
"eval_hfsummary_loss": 0.7264418005943298,
"eval_hfsummary_runtime": 2480.0981,
"eval_hfsummary_samples_per_second": 13.339,
"eval_hfsummary_steps_per_second": 1.334,
"step": 3250
},
{
"epoch": 1.5,
"eval_gptsynthetic_accuracy": 0.9984917043740573,
"eval_gptsynthetic_loss": 0.008459103293716908,
"eval_gptsynthetic_runtime": 116.9721,
"eval_gptsynthetic_samples_per_second": 28.34,
"eval_gptsynthetic_steps_per_second": 2.838,
"step": 3250
},
{
"epoch": 1.51,
"learning_rate": 2.521978021978022e-06,
"loss": 0.2776,
"step": 3260
},
{
"epoch": 1.51,
"learning_rate": 2.4984301412872844e-06,
"loss": 0.3286,
"step": 3270
},
{
"epoch": 1.52,
"learning_rate": 2.4748822605965465e-06,
"loss": 0.3118,
"step": 3280
},
{
"epoch": 1.52,
"learning_rate": 2.4513343799058086e-06,
"loss": 0.3444,
"step": 3290
},
{
"epoch": 1.53,
"learning_rate": 2.4277864992150707e-06,
"loss": 0.3078,
"step": 3300
},
{
"epoch": 1.53,
"learning_rate": 2.4042386185243328e-06,
"loss": 0.3139,
"step": 3310
},
{
"epoch": 1.54,
"learning_rate": 2.3806907378335953e-06,
"loss": 0.32,
"step": 3320
},
{
"epoch": 1.54,
"learning_rate": 2.3571428571428574e-06,
"loss": 0.3211,
"step": 3330
},
{
"epoch": 1.55,
"learning_rate": 2.3335949764521195e-06,
"loss": 0.306,
"step": 3340
},
{
"epoch": 1.55,
"learning_rate": 2.3100470957613816e-06,
"loss": 0.3045,
"step": 3350
},
{
"epoch": 1.55,
"learning_rate": 2.2864992150706437e-06,
"loss": 0.3123,
"step": 3360
},
{
"epoch": 1.56,
"learning_rate": 2.262951334379906e-06,
"loss": 0.3073,
"step": 3370
},
{
"epoch": 1.56,
"learning_rate": 2.239403453689168e-06,
"loss": 0.3061,
"step": 3380
},
{
"epoch": 1.57,
"learning_rate": 2.21585557299843e-06,
"loss": 0.297,
"step": 3390
},
{
"epoch": 1.57,
"learning_rate": 2.1923076923076925e-06,
"loss": 0.3467,
"step": 3400
},
{
"epoch": 1.58,
"learning_rate": 2.1687598116169546e-06,
"loss": 0.3183,
"step": 3410
},
{
"epoch": 1.58,
"learning_rate": 2.1452119309262166e-06,
"loss": 0.3063,
"step": 3420
},
{
"epoch": 1.59,
"learning_rate": 2.1216640502354787e-06,
"loss": 0.286,
"step": 3430
},
{
"epoch": 1.59,
"learning_rate": 2.098116169544741e-06,
"loss": 0.2963,
"step": 3440
},
{
"epoch": 1.6,
"learning_rate": 2.0745682888540034e-06,
"loss": 0.3189,
"step": 3450
},
{
"epoch": 1.6,
"learning_rate": 2.0510204081632654e-06,
"loss": 0.3128,
"step": 3460
},
{
"epoch": 1.61,
"learning_rate": 2.0274725274725275e-06,
"loss": 0.3463,
"step": 3470
},
{
"epoch": 1.61,
"learning_rate": 2.0039246467817896e-06,
"loss": 0.2914,
"step": 3480
},
{
"epoch": 1.61,
"learning_rate": 1.9803767660910517e-06,
"loss": 0.3021,
"step": 3490
},
{
"epoch": 1.62,
"learning_rate": 1.9568288854003142e-06,
"loss": 0.3168,
"step": 3500
},
{
"epoch": 1.62,
"eval_webgpt_accuracy": 0.5870786516853933,
"eval_webgpt_loss": 0.6775321364402771,
"eval_webgpt_runtime": 153.1117,
"eval_webgpt_samples_per_second": 25.576,
"eval_webgpt_steps_per_second": 2.56,
"step": 3500
},
{
"epoch": 1.62,
"eval_hfsummary_accuracy": 0.6833116706465556,
"eval_hfsummary_loss": 0.7361535429954529,
"eval_hfsummary_runtime": 2478.5992,
"eval_hfsummary_samples_per_second": 13.347,
"eval_hfsummary_steps_per_second": 1.335,
"step": 3500
},
{
"epoch": 1.62,
"eval_gptsynthetic_accuracy": 0.9984917043740573,
"eval_gptsynthetic_loss": 0.005919534247368574,
"eval_gptsynthetic_runtime": 116.9876,
"eval_gptsynthetic_samples_per_second": 28.336,
"eval_gptsynthetic_steps_per_second": 2.838,
"step": 3500
},
{
"epoch": 1.62,
"learning_rate": 1.9332810047095763e-06,
"loss": 0.3267,
"step": 3510
},
{
"epoch": 1.63,
"learning_rate": 1.9097331240188384e-06,
"loss": 0.3299,
"step": 3520
},
{
"epoch": 1.63,
"learning_rate": 1.8861852433281007e-06,
"loss": 0.3035,
"step": 3530
},
{
"epoch": 1.64,
"learning_rate": 1.8626373626373626e-06,
"loss": 0.2943,
"step": 3540
},
{
"epoch": 1.64,
"learning_rate": 1.8390894819466247e-06,
"loss": 0.287,
"step": 3550
},
{
"epoch": 1.65,
"learning_rate": 1.815541601255887e-06,
"loss": 0.3217,
"step": 3560
},
{
"epoch": 1.65,
"learning_rate": 1.7919937205651491e-06,
"loss": 0.3173,
"step": 3570
},
{
"epoch": 1.66,
"learning_rate": 1.7684458398744112e-06,
"loss": 0.319,
"step": 3580
},
{
"epoch": 1.66,
"learning_rate": 1.7448979591836735e-06,
"loss": 0.3236,
"step": 3590
},
{
"epoch": 1.67,
"learning_rate": 1.7213500784929356e-06,
"loss": 0.3055,
"step": 3600
},
{
"epoch": 1.67,
"learning_rate": 1.697802197802198e-06,
"loss": 0.3119,
"step": 3610
},
{
"epoch": 1.67,
"learning_rate": 1.67425431711146e-06,
"loss": 0.3349,
"step": 3620
},
{
"epoch": 1.68,
"learning_rate": 1.650706436420722e-06,
"loss": 0.2947,
"step": 3630
},
{
"epoch": 1.68,
"learning_rate": 1.6271585557299844e-06,
"loss": 0.3285,
"step": 3640
},
{
"epoch": 1.69,
"learning_rate": 1.6036106750392465e-06,
"loss": 0.3137,
"step": 3650
},
{
"epoch": 1.69,
"learning_rate": 1.5800627943485088e-06,
"loss": 0.2999,
"step": 3660
},
{
"epoch": 1.7,
"learning_rate": 1.5565149136577709e-06,
"loss": 0.3313,
"step": 3670
},
{
"epoch": 1.7,
"learning_rate": 1.532967032967033e-06,
"loss": 0.3295,
"step": 3680
},
{
"epoch": 1.71,
"learning_rate": 1.5094191522762953e-06,
"loss": 0.2978,
"step": 3690
},
{
"epoch": 1.71,
"learning_rate": 1.4858712715855574e-06,
"loss": 0.3036,
"step": 3700
},
{
"epoch": 1.72,
"learning_rate": 1.4623233908948197e-06,
"loss": 0.3072,
"step": 3710
},
{
"epoch": 1.72,
"learning_rate": 1.4387755102040818e-06,
"loss": 0.3102,
"step": 3720
},
{
"epoch": 1.73,
"learning_rate": 1.4152276295133439e-06,
"loss": 0.3215,
"step": 3730
},
{
"epoch": 1.73,
"learning_rate": 1.3916797488226062e-06,
"loss": 0.3018,
"step": 3740
},
{
"epoch": 1.74,
"learning_rate": 1.368131868131868e-06,
"loss": 0.312,
"step": 3750
},
{
"epoch": 1.74,
"eval_webgpt_accuracy": 0.5873340143003064,
"eval_webgpt_loss": 0.6725783348083496,
"eval_webgpt_runtime": 153.1463,
"eval_webgpt_samples_per_second": 25.57,
"eval_webgpt_steps_per_second": 2.56,
"step": 3750
},
{
"epoch": 1.74,
"eval_hfsummary_accuracy": 0.6811957803101291,
"eval_hfsummary_loss": 0.7275147438049316,
"eval_hfsummary_runtime": 2480.3805,
"eval_hfsummary_samples_per_second": 13.338,
"eval_hfsummary_steps_per_second": 1.334,
"step": 3750
},
{
"epoch": 1.74,
"eval_gptsynthetic_accuracy": 0.9984917043740573,
"eval_gptsynthetic_loss": 0.0056294966489076614,
"eval_gptsynthetic_runtime": 116.9973,
"eval_gptsynthetic_samples_per_second": 28.334,
"eval_gptsynthetic_steps_per_second": 2.838,
"step": 3750
},
{
"epoch": 1.74,
"learning_rate": 1.3445839874411302e-06,
"loss": 0.3263,
"step": 3760
},
{
"epoch": 1.74,
"learning_rate": 1.3210361067503925e-06,
"loss": 0.3066,
"step": 3770
},
{
"epoch": 1.75,
"learning_rate": 1.2974882260596545e-06,
"loss": 0.2967,
"step": 3780
},
{
"epoch": 1.75,
"learning_rate": 1.2739403453689169e-06,
"loss": 0.3409,
"step": 3790
},
{
"epoch": 1.76,
"learning_rate": 1.250392464678179e-06,
"loss": 0.3158,
"step": 3800
},
{
"epoch": 1.76,
"learning_rate": 1.226844583987441e-06,
"loss": 0.3123,
"step": 3810
},
{
"epoch": 1.77,
"learning_rate": 1.2032967032967033e-06,
"loss": 0.3299,
"step": 3820
},
{
"epoch": 1.77,
"learning_rate": 1.1797488226059654e-06,
"loss": 0.2974,
"step": 3830
},
{
"epoch": 1.78,
"learning_rate": 1.1562009419152277e-06,
"loss": 0.3047,
"step": 3840
},
{
"epoch": 1.78,
"learning_rate": 1.1326530612244898e-06,
"loss": 0.2854,
"step": 3850
},
{
"epoch": 1.79,
"learning_rate": 1.109105180533752e-06,
"loss": 0.3379,
"step": 3860
},
{
"epoch": 1.79,
"learning_rate": 1.0855572998430142e-06,
"loss": 0.3147,
"step": 3870
},
{
"epoch": 1.8,
"learning_rate": 1.0620094191522763e-06,
"loss": 0.2994,
"step": 3880
},
{
"epoch": 1.8,
"learning_rate": 1.0384615384615386e-06,
"loss": 0.3146,
"step": 3890
},
{
"epoch": 1.8,
"learning_rate": 1.0149136577708005e-06,
"loss": 0.3412,
"step": 3900
},
{
"epoch": 1.81,
"learning_rate": 9.913657770800628e-07,
"loss": 0.285,
"step": 3910
},
{
"epoch": 1.81,
"learning_rate": 9.67817896389325e-07,
"loss": 0.3162,
"step": 3920
},
{
"epoch": 1.82,
"learning_rate": 9.442700156985871e-07,
"loss": 0.3123,
"step": 3930
},
{
"epoch": 1.82,
"learning_rate": 9.207221350078493e-07,
"loss": 0.3323,
"step": 3940
},
{
"epoch": 1.83,
"learning_rate": 8.971742543171115e-07,
"loss": 0.3072,
"step": 3950
},
{
"epoch": 1.83,
"learning_rate": 8.736263736263737e-07,
"loss": 0.3104,
"step": 3960
},
{
"epoch": 1.84,
"learning_rate": 8.500784929356358e-07,
"loss": 0.3,
"step": 3970
},
{
"epoch": 1.84,
"learning_rate": 8.26530612244898e-07,
"loss": 0.3235,
"step": 3980
},
{
"epoch": 1.85,
"learning_rate": 8.029827315541602e-07,
"loss": 0.3237,
"step": 3990
},
{
"epoch": 1.85,
"learning_rate": 7.794348508634223e-07,
"loss": 0.3086,
"step": 4000
},
{
"epoch": 1.85,
"eval_webgpt_accuracy": 0.5875893769152196,
"eval_webgpt_loss": 0.6754750609397888,
"eval_webgpt_runtime": 153.2384,
"eval_webgpt_samples_per_second": 25.555,
"eval_webgpt_steps_per_second": 2.558,
"step": 4000
},
{
"epoch": 1.85,
"eval_hfsummary_accuracy": 0.6865761871656137,
"eval_hfsummary_loss": 0.7391700744628906,
"eval_hfsummary_runtime": 2482.99,
"eval_hfsummary_samples_per_second": 13.324,
"eval_hfsummary_steps_per_second": 1.333,
"step": 4000
},
{
"epoch": 1.85,
"eval_gptsynthetic_accuracy": 0.9984917043740573,
"eval_gptsynthetic_loss": 0.006686400156468153,
"eval_gptsynthetic_runtime": 116.9831,
"eval_gptsynthetic_samples_per_second": 28.337,
"eval_gptsynthetic_steps_per_second": 2.838,
"step": 4000
}
],
"max_steps": 4322,
"num_train_epochs": 2,
"total_flos": 7.590150027767578e+17,
"trial_name": null,
"trial_params": null
}