Text Classification
Transformers
PyTorch
English
deberta-v2
reward-model
reward_model
RLHF
Inference Endpoints
reward-model-deberta-v3-large / trainer_state.json
theblackcat102's picture
Upload 13 files
24b78a2
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.301327088212334,
"global_step": 2500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 1.8e-07,
"loss": 0.6935,
"step": 10
},
{
"epoch": 0.01,
"learning_rate": 3.8e-07,
"loss": 0.6974,
"step": 20
},
{
"epoch": 0.02,
"learning_rate": 5.800000000000001e-07,
"loss": 0.694,
"step": 30
},
{
"epoch": 0.02,
"learning_rate": 7.8e-07,
"loss": 0.6966,
"step": 40
},
{
"epoch": 0.03,
"learning_rate": 9.800000000000001e-07,
"loss": 0.6948,
"step": 50
},
{
"epoch": 0.03,
"learning_rate": 1.1800000000000001e-06,
"loss": 0.6957,
"step": 60
},
{
"epoch": 0.04,
"learning_rate": 1.3800000000000001e-06,
"loss": 0.6952,
"step": 70
},
{
"epoch": 0.04,
"learning_rate": 1.5800000000000001e-06,
"loss": 0.6972,
"step": 80
},
{
"epoch": 0.05,
"learning_rate": 1.7800000000000001e-06,
"loss": 0.6925,
"step": 90
},
{
"epoch": 0.05,
"learning_rate": 1.98e-06,
"loss": 0.6903,
"step": 100
},
{
"epoch": 0.06,
"learning_rate": 2.1800000000000003e-06,
"loss": 0.6837,
"step": 110
},
{
"epoch": 0.06,
"learning_rate": 2.38e-06,
"loss": 0.6754,
"step": 120
},
{
"epoch": 0.07,
"learning_rate": 2.5800000000000003e-06,
"loss": 0.6447,
"step": 130
},
{
"epoch": 0.07,
"learning_rate": 2.7800000000000005e-06,
"loss": 0.6391,
"step": 140
},
{
"epoch": 0.08,
"learning_rate": 2.9800000000000003e-06,
"loss": 0.6118,
"step": 150
},
{
"epoch": 0.08,
"learning_rate": 3.1800000000000005e-06,
"loss": 0.6071,
"step": 160
},
{
"epoch": 0.09,
"learning_rate": 3.3800000000000007e-06,
"loss": 0.6027,
"step": 170
},
{
"epoch": 0.09,
"learning_rate": 3.58e-06,
"loss": 0.5978,
"step": 180
},
{
"epoch": 0.1,
"learning_rate": 3.7800000000000002e-06,
"loss": 0.5974,
"step": 190
},
{
"epoch": 0.1,
"learning_rate": 3.980000000000001e-06,
"loss": 0.5929,
"step": 200
},
{
"epoch": 0.11,
"learning_rate": 4.18e-06,
"loss": 0.5853,
"step": 210
},
{
"epoch": 0.11,
"learning_rate": 4.38e-06,
"loss": 0.582,
"step": 220
},
{
"epoch": 0.12,
"learning_rate": 4.58e-06,
"loss": 0.5443,
"step": 230
},
{
"epoch": 0.12,
"learning_rate": 4.78e-06,
"loss": 0.5475,
"step": 240
},
{
"epoch": 0.13,
"learning_rate": 4.980000000000001e-06,
"loss": 0.5348,
"step": 250
},
{
"epoch": 0.13,
"eval_webgpt_accuracy": 0.5572012257405515,
"eval_webgpt_loss": 0.6882389783859253,
"eval_webgpt_runtime": 166.4816,
"eval_webgpt_samples_per_second": 23.522,
"eval_webgpt_steps_per_second": 4.709,
"step": 250
},
{
"epoch": 0.13,
"eval_hfsummary_accuracy": 0.6711301877096999,
"eval_hfsummary_loss": 0.6183204054832458,
"eval_hfsummary_runtime": 3223.0913,
"eval_hfsummary_samples_per_second": 10.264,
"eval_hfsummary_steps_per_second": 2.053,
"step": 250
},
{
"epoch": 0.13,
"eval_gptsynthetic_accuracy": 0.9993966817496229,
"eval_gptsynthetic_loss": 0.009097465313971043,
"eval_gptsynthetic_runtime": 101.5011,
"eval_gptsynthetic_samples_per_second": 32.66,
"eval_gptsynthetic_steps_per_second": 6.532,
"step": 250
},
{
"epoch": 0.14,
"learning_rate": 5.18e-06,
"loss": 0.537,
"step": 260
},
{
"epoch": 0.14,
"learning_rate": 5.380000000000001e-06,
"loss": 0.5286,
"step": 270
},
{
"epoch": 0.15,
"learning_rate": 5.580000000000001e-06,
"loss": 0.4957,
"step": 280
},
{
"epoch": 0.15,
"learning_rate": 5.78e-06,
"loss": 0.5117,
"step": 290
},
{
"epoch": 0.16,
"learning_rate": 5.98e-06,
"loss": 0.4977,
"step": 300
},
{
"epoch": 0.16,
"learning_rate": 6.18e-06,
"loss": 0.4801,
"step": 310
},
{
"epoch": 0.17,
"learning_rate": 6.380000000000001e-06,
"loss": 0.499,
"step": 320
},
{
"epoch": 0.17,
"learning_rate": 6.5800000000000005e-06,
"loss": 0.4702,
"step": 330
},
{
"epoch": 0.18,
"learning_rate": 6.780000000000001e-06,
"loss": 0.4715,
"step": 340
},
{
"epoch": 0.18,
"learning_rate": 6.98e-06,
"loss": 0.4686,
"step": 350
},
{
"epoch": 0.19,
"learning_rate": 7.180000000000001e-06,
"loss": 0.488,
"step": 360
},
{
"epoch": 0.19,
"learning_rate": 7.3800000000000005e-06,
"loss": 0.4671,
"step": 370
},
{
"epoch": 0.2,
"learning_rate": 7.58e-06,
"loss": 0.4889,
"step": 380
},
{
"epoch": 0.2,
"learning_rate": 7.78e-06,
"loss": 0.4718,
"step": 390
},
{
"epoch": 0.21,
"learning_rate": 7.980000000000002e-06,
"loss": 0.4706,
"step": 400
},
{
"epoch": 0.21,
"learning_rate": 8.18e-06,
"loss": 0.4839,
"step": 410
},
{
"epoch": 0.22,
"learning_rate": 8.380000000000001e-06,
"loss": 0.4576,
"step": 420
},
{
"epoch": 0.22,
"learning_rate": 8.580000000000001e-06,
"loss": 0.4865,
"step": 430
},
{
"epoch": 0.23,
"learning_rate": 8.78e-06,
"loss": 0.468,
"step": 440
},
{
"epoch": 0.23,
"learning_rate": 8.98e-06,
"loss": 0.4335,
"step": 450
},
{
"epoch": 0.24,
"learning_rate": 9.180000000000002e-06,
"loss": 0.4516,
"step": 460
},
{
"epoch": 0.24,
"learning_rate": 9.38e-06,
"loss": 0.4603,
"step": 470
},
{
"epoch": 0.25,
"learning_rate": 9.58e-06,
"loss": 0.467,
"step": 480
},
{
"epoch": 0.26,
"learning_rate": 9.780000000000001e-06,
"loss": 0.4678,
"step": 490
},
{
"epoch": 0.26,
"learning_rate": 9.980000000000001e-06,
"loss": 0.4456,
"step": 500
},
{
"epoch": 0.26,
"eval_webgpt_accuracy": 0.5893769152196119,
"eval_webgpt_loss": 0.6772251725196838,
"eval_webgpt_runtime": 166.9484,
"eval_webgpt_samples_per_second": 23.456,
"eval_webgpt_steps_per_second": 4.696,
"step": 500
},
{
"epoch": 0.26,
"eval_hfsummary_accuracy": 0.7065864643472478,
"eval_hfsummary_loss": 0.5687452554702759,
"eval_hfsummary_runtime": 3231.8075,
"eval_hfsummary_samples_per_second": 10.237,
"eval_hfsummary_steps_per_second": 2.047,
"step": 500
},
{
"epoch": 0.26,
"eval_gptsynthetic_accuracy": 0.9993966817496229,
"eval_gptsynthetic_loss": 0.005173402838408947,
"eval_gptsynthetic_runtime": 101.6573,
"eval_gptsynthetic_samples_per_second": 32.61,
"eval_gptsynthetic_steps_per_second": 6.522,
"step": 500
},
{
"epoch": 0.27,
"learning_rate": 9.999821166271659e-06,
"loss": 0.4375,
"step": 510
},
{
"epoch": 0.27,
"learning_rate": 9.999202992029537e-06,
"loss": 0.4608,
"step": 520
},
{
"epoch": 0.28,
"learning_rate": 9.998143324029546e-06,
"loss": 0.4674,
"step": 530
},
{
"epoch": 0.28,
"learning_rate": 9.996642255853966e-06,
"loss": 0.4468,
"step": 540
},
{
"epoch": 0.29,
"learning_rate": 9.99469992006637e-06,
"loss": 0.4466,
"step": 550
},
{
"epoch": 0.29,
"learning_rate": 9.992316488199927e-06,
"loss": 0.4608,
"step": 560
},
{
"epoch": 0.3,
"learning_rate": 9.989492170742248e-06,
"loss": 0.436,
"step": 570
},
{
"epoch": 0.3,
"learning_rate": 9.986227217116798e-06,
"loss": 0.4249,
"step": 580
},
{
"epoch": 0.31,
"learning_rate": 9.982521915660865e-06,
"loss": 0.4702,
"step": 590
},
{
"epoch": 0.31,
"learning_rate": 9.978376593600107e-06,
"loss": 0.4336,
"step": 600
},
{
"epoch": 0.32,
"learning_rate": 9.973791617019634e-06,
"loss": 0.4617,
"step": 610
},
{
"epoch": 0.32,
"learning_rate": 9.968767390831707e-06,
"loss": 0.4547,
"step": 620
},
{
"epoch": 0.33,
"learning_rate": 9.963304358739943e-06,
"loss": 0.4285,
"step": 630
},
{
"epoch": 0.33,
"learning_rate": 9.957403003200167e-06,
"loss": 0.4626,
"step": 640
},
{
"epoch": 0.34,
"learning_rate": 9.951063845377778e-06,
"loss": 0.4153,
"step": 650
},
{
"epoch": 0.34,
"learning_rate": 9.944287445101733e-06,
"loss": 0.4196,
"step": 660
},
{
"epoch": 0.35,
"learning_rate": 9.937074400815114e-06,
"loss": 0.4461,
"step": 670
},
{
"epoch": 0.35,
"learning_rate": 9.929425349522264e-06,
"loss": 0.4108,
"step": 680
},
{
"epoch": 0.36,
"learning_rate": 9.921340966732539e-06,
"loss": 0.4466,
"step": 690
},
{
"epoch": 0.36,
"learning_rate": 9.912821966400652e-06,
"loss": 0.4608,
"step": 700
},
{
"epoch": 0.37,
"learning_rate": 9.903869100863622e-06,
"loss": 0.4382,
"step": 710
},
{
"epoch": 0.37,
"learning_rate": 9.894483160774331e-06,
"loss": 0.4172,
"step": 720
},
{
"epoch": 0.38,
"learning_rate": 9.884664975031698e-06,
"loss": 0.4021,
"step": 730
},
{
"epoch": 0.39,
"learning_rate": 9.874415410707476e-06,
"loss": 0.4108,
"step": 740
},
{
"epoch": 0.39,
"learning_rate": 9.863735372969692e-06,
"loss": 0.4402,
"step": 750
},
{
"epoch": 0.39,
"eval_webgpt_accuracy": 0.5947395301327886,
"eval_webgpt_loss": 0.6755640506744385,
"eval_webgpt_runtime": 167.2694,
"eval_webgpt_samples_per_second": 23.411,
"eval_webgpt_steps_per_second": 4.687,
"step": 750
},
{
"epoch": 0.39,
"eval_hfsummary_accuracy": 0.7158661548227186,
"eval_hfsummary_loss": 0.5612313151359558,
"eval_hfsummary_runtime": 3314.4907,
"eval_hfsummary_samples_per_second": 9.981,
"eval_hfsummary_steps_per_second": 1.996,
"step": 750
},
{
"epoch": 0.39,
"eval_gptsynthetic_accuracy": 0.9996983408748115,
"eval_gptsynthetic_loss": 0.005162315908819437,
"eval_gptsynthetic_runtime": 105.2324,
"eval_gptsynthetic_samples_per_second": 31.502,
"eval_gptsynthetic_steps_per_second": 6.3,
"step": 750
},
{
"epoch": 0.4,
"learning_rate": 9.852625805002682e-06,
"loss": 0.4411,
"step": 760
},
{
"epoch": 0.4,
"learning_rate": 9.841087687923826e-06,
"loss": 0.4239,
"step": 770
},
{
"epoch": 0.41,
"learning_rate": 9.82912204069688e-06,
"loss": 0.4489,
"step": 780
},
{
"epoch": 0.41,
"learning_rate": 9.816729920041996e-06,
"loss": 0.4337,
"step": 790
},
{
"epoch": 0.42,
"learning_rate": 9.803912420342406e-06,
"loss": 0.4346,
"step": 800
},
{
"epoch": 0.42,
"learning_rate": 9.790670673547766e-06,
"loss": 0.3994,
"step": 810
},
{
"epoch": 0.43,
"learning_rate": 9.777005849074195e-06,
"loss": 0.4151,
"step": 820
},
{
"epoch": 0.43,
"learning_rate": 9.762919153700996e-06,
"loss": 0.4519,
"step": 830
},
{
"epoch": 0.44,
"learning_rate": 9.748411831464083e-06,
"loss": 0.4409,
"step": 840
},
{
"epoch": 0.44,
"learning_rate": 9.733485163546124e-06,
"loss": 0.4069,
"step": 850
},
{
"epoch": 0.45,
"learning_rate": 9.718140468163386e-06,
"loss": 0.3974,
"step": 860
},
{
"epoch": 0.45,
"learning_rate": 9.702379100449325e-06,
"loss": 0.4192,
"step": 870
},
{
"epoch": 0.46,
"learning_rate": 9.686202452334909e-06,
"loss": 0.4106,
"step": 880
},
{
"epoch": 0.46,
"learning_rate": 9.669611952425694e-06,
"loss": 0.4224,
"step": 890
},
{
"epoch": 0.47,
"learning_rate": 9.652609065875656e-06,
"loss": 0.4176,
"step": 900
},
{
"epoch": 0.47,
"learning_rate": 9.635195294257806e-06,
"loss": 0.3897,
"step": 910
},
{
"epoch": 0.48,
"learning_rate": 9.617372175431574e-06,
"loss": 0.4201,
"step": 920
},
{
"epoch": 0.48,
"learning_rate": 9.599141283407004e-06,
"loss": 0.3955,
"step": 930
},
{
"epoch": 0.49,
"learning_rate": 9.580504228205736e-06,
"loss": 0.3754,
"step": 940
},
{
"epoch": 0.49,
"learning_rate": 9.56146265571884e-06,
"loss": 0.407,
"step": 950
},
{
"epoch": 0.5,
"learning_rate": 9.542018247561442e-06,
"loss": 0.4225,
"step": 960
},
{
"epoch": 0.5,
"learning_rate": 9.522172720924236e-06,
"loss": 0.415,
"step": 970
},
{
"epoch": 0.51,
"learning_rate": 9.501927828421816e-06,
"loss": 0.4014,
"step": 980
},
{
"epoch": 0.52,
"learning_rate": 9.481285357937905e-06,
"loss": 0.4074,
"step": 990
},
{
"epoch": 0.52,
"learning_rate": 9.460247132467466e-06,
"loss": 0.3874,
"step": 1000
},
{
"epoch": 0.52,
"eval_webgpt_accuracy": 0.6021450459652706,
"eval_webgpt_loss": 0.6651596426963806,
"eval_webgpt_runtime": 167.2303,
"eval_webgpt_samples_per_second": 23.417,
"eval_webgpt_steps_per_second": 4.688,
"step": 1000
},
{
"epoch": 0.52,
"eval_hfsummary_accuracy": 0.714052534534353,
"eval_hfsummary_loss": 0.5728897452354431,
"eval_hfsummary_runtime": 3240.5714,
"eval_hfsummary_samples_per_second": 10.209,
"eval_hfsummary_steps_per_second": 2.042,
"step": 1000
},
{
"epoch": 0.52,
"eval_gptsynthetic_accuracy": 0.9990950226244344,
"eval_gptsynthetic_loss": 0.0032840222120285034,
"eval_gptsynthetic_runtime": 101.8215,
"eval_gptsynthetic_samples_per_second": 32.557,
"eval_gptsynthetic_steps_per_second": 6.511,
"step": 1000
},
{
"epoch": 0.53,
"learning_rate": 9.438815009955701e-06,
"loss": 0.386,
"step": 1010
},
{
"epoch": 0.53,
"learning_rate": 9.41699088313398e-06,
"loss": 0.401,
"step": 1020
},
{
"epoch": 0.54,
"learning_rate": 9.394776679352675e-06,
"loss": 0.4199,
"step": 1030
},
{
"epoch": 0.54,
"learning_rate": 9.372174360410963e-06,
"loss": 0.4093,
"step": 1040
},
{
"epoch": 0.55,
"learning_rate": 9.349185922383564e-06,
"loss": 0.4313,
"step": 1050
},
{
"epoch": 0.55,
"learning_rate": 9.325813395444472e-06,
"loss": 0.4164,
"step": 1060
},
{
"epoch": 0.56,
"learning_rate": 9.302058843687655e-06,
"loss": 0.3871,
"step": 1070
},
{
"epoch": 0.56,
"learning_rate": 9.277924364944774e-06,
"loss": 0.4115,
"step": 1080
},
{
"epoch": 0.57,
"learning_rate": 9.253412090599923e-06,
"loss": 0.4094,
"step": 1090
},
{
"epoch": 0.57,
"learning_rate": 9.228524185401384e-06,
"loss": 0.3957,
"step": 1100
},
{
"epoch": 0.58,
"learning_rate": 9.203262847270469e-06,
"loss": 0.3952,
"step": 1110
},
{
"epoch": 0.58,
"learning_rate": 9.177630307107408e-06,
"loss": 0.4011,
"step": 1120
},
{
"epoch": 0.59,
"learning_rate": 9.15162882859433e-06,
"loss": 0.4042,
"step": 1130
},
{
"epoch": 0.59,
"learning_rate": 9.125260707995356e-06,
"loss": 0.404,
"step": 1140
},
{
"epoch": 0.6,
"learning_rate": 9.098528273953807e-06,
"loss": 0.3904,
"step": 1150
},
{
"epoch": 0.6,
"learning_rate": 9.071433887286551e-06,
"loss": 0.4011,
"step": 1160
},
{
"epoch": 0.61,
"learning_rate": 9.04397994077552e-06,
"loss": 0.3806,
"step": 1170
},
{
"epoch": 0.61,
"learning_rate": 9.01616885895639e-06,
"loss": 0.3878,
"step": 1180
},
{
"epoch": 0.62,
"learning_rate": 8.988003097904472e-06,
"loss": 0.384,
"step": 1190
},
{
"epoch": 0.62,
"learning_rate": 8.959485145017794e-06,
"loss": 0.4023,
"step": 1200
},
{
"epoch": 0.63,
"learning_rate": 8.930617518797448e-06,
"loss": 0.3609,
"step": 1210
},
{
"epoch": 0.63,
"learning_rate": 8.901402768625169e-06,
"loss": 0.3879,
"step": 1220
},
{
"epoch": 0.64,
"learning_rate": 8.871843474538182e-06,
"loss": 0.4003,
"step": 1230
},
{
"epoch": 0.65,
"learning_rate": 8.841942247001368e-06,
"loss": 0.3974,
"step": 1240
},
{
"epoch": 0.65,
"learning_rate": 8.811701726676711e-06,
"loss": 0.4067,
"step": 1250
},
{
"epoch": 0.65,
"eval_webgpt_accuracy": 0.5967824310520939,
"eval_webgpt_loss": 0.6645339727401733,
"eval_webgpt_runtime": 167.52,
"eval_webgpt_samples_per_second": 23.376,
"eval_webgpt_steps_per_second": 4.68,
"step": 1250
},
{
"epoch": 0.65,
"eval_hfsummary_accuracy": 0.7167125109572893,
"eval_hfsummary_loss": 0.559929370880127,
"eval_hfsummary_runtime": 3290.2163,
"eval_hfsummary_samples_per_second": 10.055,
"eval_hfsummary_steps_per_second": 2.011,
"step": 1250
},
{
"epoch": 0.65,
"eval_gptsynthetic_accuracy": 0.9993966817496229,
"eval_gptsynthetic_loss": 0.002522215712815523,
"eval_gptsynthetic_runtime": 116.8538,
"eval_gptsynthetic_samples_per_second": 28.369,
"eval_gptsynthetic_steps_per_second": 5.674,
"step": 1250
},
{
"epoch": 0.66,
"learning_rate": 8.78112458419011e-06,
"loss": 0.3581,
"step": 1260
},
{
"epoch": 0.66,
"learning_rate": 8.750213519895511e-06,
"loss": 0.3628,
"step": 1270
},
{
"epoch": 0.67,
"learning_rate": 8.718971263636441e-06,
"loss": 0.3799,
"step": 1280
},
{
"epoch": 0.67,
"learning_rate": 8.687400574504929e-06,
"loss": 0.4083,
"step": 1290
},
{
"epoch": 0.68,
"learning_rate": 8.655504240597841e-06,
"loss": 0.3771,
"step": 1300
},
{
"epoch": 0.68,
"learning_rate": 8.623285078770649e-06,
"loss": 0.4016,
"step": 1310
},
{
"epoch": 0.69,
"learning_rate": 8.590745934388675e-06,
"loss": 0.3878,
"step": 1320
},
{
"epoch": 0.69,
"learning_rate": 8.557889681075808e-06,
"loss": 0.3568,
"step": 1330
},
{
"epoch": 0.7,
"learning_rate": 8.524719220460718e-06,
"loss": 0.3914,
"step": 1340
},
{
"epoch": 0.7,
"learning_rate": 8.491237481920616e-06,
"loss": 0.4001,
"step": 1350
},
{
"epoch": 0.71,
"learning_rate": 8.457447422322545e-06,
"loss": 0.3683,
"step": 1360
},
{
"epoch": 0.71,
"learning_rate": 8.42335202576225e-06,
"loss": 0.3741,
"step": 1370
},
{
"epoch": 0.72,
"learning_rate": 8.388954303300651e-06,
"loss": 0.3636,
"step": 1380
},
{
"epoch": 0.72,
"learning_rate": 8.354257292697927e-06,
"loss": 0.3815,
"step": 1390
},
{
"epoch": 0.73,
"learning_rate": 8.319264058145227e-06,
"loss": 0.3596,
"step": 1400
},
{
"epoch": 0.73,
"learning_rate": 8.283977689994085e-06,
"loss": 0.3826,
"step": 1410
},
{
"epoch": 0.74,
"learning_rate": 8.248401304483492e-06,
"loss": 0.41,
"step": 1420
},
{
"epoch": 0.74,
"learning_rate": 8.21253804346469e-06,
"loss": 0.3648,
"step": 1430
},
{
"epoch": 0.75,
"learning_rate": 8.176391074123705e-06,
"loss": 0.3439,
"step": 1440
},
{
"epoch": 0.75,
"learning_rate": 8.139963588701656e-06,
"loss": 0.3635,
"step": 1450
},
{
"epoch": 0.76,
"learning_rate": 8.103258804212824e-06,
"loss": 0.3633,
"step": 1460
},
{
"epoch": 0.77,
"learning_rate": 8.066279962160554e-06,
"loss": 0.3989,
"step": 1470
},
{
"epoch": 0.77,
"learning_rate": 8.029030328250997e-06,
"loss": 0.3504,
"step": 1480
},
{
"epoch": 0.78,
"learning_rate": 7.991513192104694e-06,
"loss": 0.3665,
"step": 1490
},
{
"epoch": 0.78,
"learning_rate": 7.953731866966063e-06,
"loss": 0.3874,
"step": 1500
},
{
"epoch": 0.78,
"eval_webgpt_accuracy": 0.6113381001021451,
"eval_webgpt_loss": 0.6596858501434326,
"eval_webgpt_runtime": 179.0732,
"eval_webgpt_samples_per_second": 21.868,
"eval_webgpt_steps_per_second": 4.378,
"step": 1500
},
{
"epoch": 0.78,
"eval_hfsummary_accuracy": 0.7207327025964997,
"eval_hfsummary_loss": 0.5737717747688293,
"eval_hfsummary_runtime": 3606.3862,
"eval_hfsummary_samples_per_second": 9.173,
"eval_hfsummary_steps_per_second": 1.835,
"step": 1500
},
{
"epoch": 0.78,
"eval_gptsynthetic_accuracy": 0.9993966817496229,
"eval_gptsynthetic_loss": 0.002385695930570364,
"eval_gptsynthetic_runtime": 109.239,
"eval_gptsynthetic_samples_per_second": 30.346,
"eval_gptsynthetic_steps_per_second": 6.069,
"step": 1500
},
{
"epoch": 0.79,
"learning_rate": 7.915689689410806e-06,
"loss": 0.3532,
"step": 1510
},
{
"epoch": 0.79,
"learning_rate": 7.877390019051232e-06,
"loss": 0.403,
"step": 1520
},
{
"epoch": 0.8,
"learning_rate": 7.838836238239574e-06,
"loss": 0.3735,
"step": 1530
},
{
"epoch": 0.8,
"learning_rate": 7.800031751769274e-06,
"loss": 0.3773,
"step": 1540
},
{
"epoch": 0.81,
"learning_rate": 7.760979986574304e-06,
"loss": 0.3644,
"step": 1550
},
{
"epoch": 0.81,
"learning_rate": 7.721684391426516e-06,
"loss": 0.3692,
"step": 1560
},
{
"epoch": 0.82,
"learning_rate": 7.682148436631075e-06,
"loss": 0.3804,
"step": 1570
},
{
"epoch": 0.82,
"learning_rate": 7.642375613719988e-06,
"loss": 0.397,
"step": 1580
},
{
"epoch": 0.83,
"learning_rate": 7.602369435143755e-06,
"loss": 0.3842,
"step": 1590
},
{
"epoch": 0.83,
"learning_rate": 7.562133433961176e-06,
"loss": 0.3571,
"step": 1600
},
{
"epoch": 0.84,
"learning_rate": 7.5216711635273305e-06,
"loss": 0.3784,
"step": 1610
},
{
"epoch": 0.84,
"learning_rate": 7.480986197179785e-06,
"loss": 0.3947,
"step": 1620
},
{
"epoch": 0.85,
"learning_rate": 7.440082127923002e-06,
"loss": 0.3583,
"step": 1630
},
{
"epoch": 0.85,
"learning_rate": 7.39896256811105e-06,
"loss": 0.3671,
"step": 1640
},
{
"epoch": 0.86,
"learning_rate": 7.357631149128568e-06,
"loss": 0.3691,
"step": 1650
},
{
"epoch": 0.86,
"learning_rate": 7.31609152107009e-06,
"loss": 0.4069,
"step": 1660
},
{
"epoch": 0.87,
"learning_rate": 7.2743473524176734e-06,
"loss": 0.3515,
"step": 1670
},
{
"epoch": 0.87,
"learning_rate": 7.232402329716938e-06,
"loss": 0.3783,
"step": 1680
},
{
"epoch": 0.88,
"learning_rate": 7.19026015725149e-06,
"loss": 0.362,
"step": 1690
},
{
"epoch": 0.88,
"learning_rate": 7.147924556715793e-06,
"loss": 0.3807,
"step": 1700
},
{
"epoch": 0.89,
"learning_rate": 7.1053992668864855e-06,
"loss": 0.3632,
"step": 1710
},
{
"epoch": 0.9,
"learning_rate": 7.0626880432922095e-06,
"loss": 0.3619,
"step": 1720
},
{
"epoch": 0.9,
"learning_rate": 7.019794657881942e-06,
"loss": 0.3883,
"step": 1730
},
{
"epoch": 0.91,
"learning_rate": 6.976722898691881e-06,
"loss": 0.351,
"step": 1740
},
{
"epoch": 0.91,
"learning_rate": 6.933476569510924e-06,
"loss": 0.3613,
"step": 1750
},
{
"epoch": 0.91,
"eval_webgpt_accuracy": 0.609805924412666,
"eval_webgpt_loss": 0.6571772694587708,
"eval_webgpt_runtime": 167.3559,
"eval_webgpt_samples_per_second": 23.399,
"eval_webgpt_steps_per_second": 4.685,
"step": 1750
},
{
"epoch": 0.91,
"eval_hfsummary_accuracy": 0.7195538494090621,
"eval_hfsummary_loss": 0.5861465334892273,
"eval_hfsummary_runtime": 3235.9038,
"eval_hfsummary_samples_per_second": 10.224,
"eval_hfsummary_steps_per_second": 2.045,
"step": 1750
},
{
"epoch": 0.91,
"eval_gptsynthetic_accuracy": 0.9996983408748115,
"eval_gptsynthetic_loss": 0.0027254121378064156,
"eval_gptsynthetic_runtime": 101.6222,
"eval_gptsynthetic_samples_per_second": 32.621,
"eval_gptsynthetic_steps_per_second": 6.524,
"step": 1750
},
{
"epoch": 0.92,
"learning_rate": 6.890059489544735e-06,
"loss": 0.3651,
"step": 1760
},
{
"epoch": 0.92,
"learning_rate": 6.846475493078462e-06,
"loss": 0.3617,
"step": 1770
},
{
"epoch": 0.93,
"learning_rate": 6.802728429138122e-06,
"loss": 0.3174,
"step": 1780
},
{
"epoch": 0.93,
"learning_rate": 6.758822161150687e-06,
"loss": 0.3399,
"step": 1790
},
{
"epoch": 0.94,
"learning_rate": 6.714760566602876e-06,
"loss": 0.3664,
"step": 1800
},
{
"epoch": 0.94,
"learning_rate": 6.67054753669875e-06,
"loss": 0.3432,
"step": 1810
},
{
"epoch": 0.95,
"learning_rate": 6.626186976016049e-06,
"loss": 0.3673,
"step": 1820
},
{
"epoch": 0.95,
"learning_rate": 6.581682802161371e-06,
"loss": 0.3562,
"step": 1830
},
{
"epoch": 0.96,
"learning_rate": 6.537038945424201e-06,
"loss": 0.338,
"step": 1840
},
{
"epoch": 0.96,
"learning_rate": 6.492259348429818e-06,
"loss": 0.3387,
"step": 1850
},
{
"epoch": 0.97,
"learning_rate": 6.4473479657910995e-06,
"loss": 0.3504,
"step": 1860
},
{
"epoch": 0.97,
"learning_rate": 6.402308763759296e-06,
"loss": 0.3276,
"step": 1870
},
{
"epoch": 0.98,
"learning_rate": 6.357145719873742e-06,
"loss": 0.3243,
"step": 1880
},
{
"epoch": 0.98,
"learning_rate": 6.3118628226105995e-06,
"loss": 0.3548,
"step": 1890
},
{
"epoch": 0.99,
"learning_rate": 6.266464071030621e-06,
"loss": 0.3099,
"step": 1900
},
{
"epoch": 0.99,
"learning_rate": 6.220953474425984e-06,
"loss": 0.3387,
"step": 1910
},
{
"epoch": 1.0,
"learning_rate": 6.175335051966211e-06,
"loss": 0.3403,
"step": 1920
},
{
"epoch": 1.0,
"learning_rate": 6.129612832343233e-06,
"loss": 0.3385,
"step": 1930
},
{
"epoch": 1.01,
"learning_rate": 6.0837908534156076e-06,
"loss": 0.3305,
"step": 1940
},
{
"epoch": 1.02,
"learning_rate": 6.037873161851911e-06,
"loss": 0.3421,
"step": 1950
},
{
"epoch": 1.02,
"learning_rate": 5.9918638127733764e-06,
"loss": 0.3381,
"step": 1960
},
{
"epoch": 1.03,
"learning_rate": 5.945766869395772e-06,
"loss": 0.3062,
"step": 1970
},
{
"epoch": 1.03,
"learning_rate": 5.899586402670566e-06,
"loss": 0.3459,
"step": 1980
},
{
"epoch": 1.04,
"learning_rate": 5.853326490925407e-06,
"loss": 0.3175,
"step": 1990
},
{
"epoch": 1.04,
"learning_rate": 5.806991219503963e-06,
"loss": 0.3117,
"step": 2000
},
{
"epoch": 1.04,
"eval_webgpt_accuracy": 0.6046986721144024,
"eval_webgpt_loss": 0.6649239659309387,
"eval_webgpt_runtime": 166.8662,
"eval_webgpt_samples_per_second": 23.468,
"eval_webgpt_steps_per_second": 4.698,
"step": 2000
},
{
"epoch": 1.04,
"eval_hfsummary_accuracy": 0.7152616147265968,
"eval_hfsummary_loss": 0.6051351428031921,
"eval_hfsummary_runtime": 3227.7808,
"eval_hfsummary_samples_per_second": 10.249,
"eval_hfsummary_steps_per_second": 2.05,
"step": 2000
},
{
"epoch": 1.04,
"eval_gptsynthetic_accuracy": 1.0,
"eval_gptsynthetic_loss": 0.002014968078583479,
"eval_gptsynthetic_runtime": 101.558,
"eval_gptsynthetic_samples_per_second": 32.641,
"eval_gptsynthetic_steps_per_second": 6.528,
"step": 2000
},
{
"epoch": 1.05,
"learning_rate": 5.765228424606831e-06,
"loss": 0.3206,
"step": 2010
},
{
"epoch": 1.05,
"learning_rate": 5.718761248466732e-06,
"loss": 0.3073,
"step": 2020
},
{
"epoch": 1.06,
"learning_rate": 5.672230596487563e-06,
"loss": 0.3113,
"step": 2030
},
{
"epoch": 1.06,
"learning_rate": 5.625640577922828e-06,
"loss": 0.3213,
"step": 2040
},
{
"epoch": 1.07,
"learning_rate": 5.578995307268855e-06,
"loss": 0.3353,
"step": 2050
},
{
"epoch": 1.07,
"learning_rate": 5.532298903901448e-06,
"loss": 0.3293,
"step": 2060
},
{
"epoch": 1.08,
"learning_rate": 5.485555491712082e-06,
"loss": 0.3158,
"step": 2070
},
{
"epoch": 1.08,
"learning_rate": 5.438769198743713e-06,
"loss": 0.2968,
"step": 2080
},
{
"epoch": 1.09,
"learning_rate": 5.391944156826221e-06,
"loss": 0.2877,
"step": 2090
},
{
"epoch": 1.09,
"learning_rate": 5.345084501211517e-06,
"loss": 0.3361,
"step": 2100
},
{
"epoch": 1.1,
"learning_rate": 5.298194370208341e-06,
"loss": 0.3051,
"step": 2110
},
{
"epoch": 1.1,
"learning_rate": 5.251277904816809e-06,
"loss": 0.3398,
"step": 2120
},
{
"epoch": 1.11,
"learning_rate": 5.2043392483626906e-06,
"loss": 0.3155,
"step": 2130
},
{
"epoch": 1.11,
"learning_rate": 5.157382546131519e-06,
"loss": 0.315,
"step": 2140
},
{
"epoch": 1.12,
"learning_rate": 5.110411945002497e-06,
"loss": 0.304,
"step": 2150
},
{
"epoch": 1.12,
"learning_rate": 5.063431593082283e-06,
"loss": 0.3458,
"step": 2160
},
{
"epoch": 1.13,
"learning_rate": 5.016445639338647e-06,
"loss": 0.3192,
"step": 2170
},
{
"epoch": 1.13,
"learning_rate": 4.969458233234083e-06,
"loss": 0.3188,
"step": 2180
},
{
"epoch": 1.14,
"learning_rate": 4.922473524359335e-06,
"loss": 0.308,
"step": 2190
},
{
"epoch": 1.15,
"learning_rate": 4.875495662066958e-06,
"loss": 0.3052,
"step": 2200
},
{
"epoch": 1.15,
"learning_rate": 4.82852879510486e-06,
"loss": 0.2962,
"step": 2210
},
{
"epoch": 1.16,
"learning_rate": 4.781577071249918e-06,
"loss": 0.3067,
"step": 2220
},
{
"epoch": 1.16,
"learning_rate": 4.734644636941687e-06,
"loss": 0.2886,
"step": 2230
},
{
"epoch": 1.17,
"learning_rate": 4.6877356369161975e-06,
"loss": 0.2832,
"step": 2240
},
{
"epoch": 1.17,
"learning_rate": 4.64085421383994e-06,
"loss": 0.3177,
"step": 2250
},
{
"epoch": 1.17,
"eval_webgpt_accuracy": 0.6052093973442288,
"eval_webgpt_loss": 0.67301344871521,
"eval_webgpt_runtime": 167.0393,
"eval_webgpt_samples_per_second": 23.444,
"eval_webgpt_steps_per_second": 4.694,
"step": 2250
},
{
"epoch": 1.17,
"eval_hfsummary_accuracy": 0.7188586282985219,
"eval_hfsummary_loss": 0.665126621723175,
"eval_hfsummary_runtime": 3228.0014,
"eval_hfsummary_samples_per_second": 10.249,
"eval_hfsummary_steps_per_second": 2.05,
"step": 2250
},
{
"epoch": 1.17,
"eval_gptsynthetic_accuracy": 1.0,
"eval_gptsynthetic_loss": 0.0015043159946799278,
"eval_gptsynthetic_runtime": 101.5554,
"eval_gptsynthetic_samples_per_second": 32.642,
"eval_gptsynthetic_steps_per_second": 6.528,
"step": 2250
},
{
"epoch": 1.18,
"learning_rate": 4.594004507944001e-06,
"loss": 0.2988,
"step": 2260
},
{
"epoch": 1.18,
"learning_rate": 4.547190656658441e-06,
"loss": 0.3037,
"step": 2270
},
{
"epoch": 1.19,
"learning_rate": 4.500416794246885e-06,
"loss": 0.3024,
"step": 2280
},
{
"epoch": 1.19,
"learning_rate": 4.453687051441443e-06,
"loss": 0.3008,
"step": 2290
},
{
"epoch": 1.2,
"learning_rate": 4.407005555077885e-06,
"loss": 0.3059,
"step": 2300
},
{
"epoch": 1.2,
"learning_rate": 4.36037642773121e-06,
"loss": 0.3191,
"step": 2310
},
{
"epoch": 1.21,
"learning_rate": 4.313803787351555e-06,
"loss": 0.3228,
"step": 2320
},
{
"epoch": 1.21,
"learning_rate": 4.267291746900538e-06,
"loss": 0.3054,
"step": 2330
},
{
"epoch": 1.22,
"learning_rate": 4.220844413988022e-06,
"loss": 0.3161,
"step": 2340
},
{
"epoch": 1.22,
"learning_rate": 4.174465890509365e-06,
"loss": 0.296,
"step": 2350
},
{
"epoch": 1.23,
"learning_rate": 4.128160272283172e-06,
"loss": 0.2718,
"step": 2360
},
{
"epoch": 1.23,
"learning_rate": 4.081931648689577e-06,
"loss": 0.3204,
"step": 2370
},
{
"epoch": 1.24,
"learning_rate": 4.035784102309097e-06,
"loss": 0.3418,
"step": 2380
},
{
"epoch": 1.24,
"learning_rate": 3.989721708562098e-06,
"loss": 0.3246,
"step": 2390
},
{
"epoch": 1.25,
"learning_rate": 3.943748535348872e-06,
"loss": 0.2989,
"step": 2400
},
{
"epoch": 1.25,
"learning_rate": 3.8978686426903885e-06,
"loss": 0.3318,
"step": 2410
},
{
"epoch": 1.26,
"learning_rate": 3.852086082369758e-06,
"loss": 0.2856,
"step": 2420
},
{
"epoch": 1.26,
"learning_rate": 3.806404897574388e-06,
"loss": 0.2934,
"step": 2430
},
{
"epoch": 1.27,
"learning_rate": 3.760829122538931e-06,
"loss": 0.3017,
"step": 2440
},
{
"epoch": 1.28,
"learning_rate": 3.7153627821890026e-06,
"loss": 0.318,
"step": 2450
},
{
"epoch": 1.28,
"learning_rate": 3.6700098917857387e-06,
"loss": 0.2964,
"step": 2460
},
{
"epoch": 1.29,
"learning_rate": 3.624774456571176e-06,
"loss": 0.2909,
"step": 2470
},
{
"epoch": 1.29,
"learning_rate": 3.579660471414565e-06,
"loss": 0.3043,
"step": 2480
},
{
"epoch": 1.3,
"learning_rate": 3.5346719204595497e-06,
"loss": 0.305,
"step": 2490
},
{
"epoch": 1.3,
"learning_rate": 3.48981277677233e-06,
"loss": 0.2756,
"step": 2500
},
{
"epoch": 1.3,
"eval_webgpt_accuracy": 0.6131256384065373,
"eval_webgpt_loss": 0.6625245809555054,
"eval_webgpt_runtime": 166.7899,
"eval_webgpt_samples_per_second": 23.479,
"eval_webgpt_steps_per_second": 4.701,
"step": 2500
},
{
"epoch": 1.3,
"eval_hfsummary_accuracy": 0.7223347338512227,
"eval_hfsummary_loss": 0.6251139044761658,
"eval_hfsummary_runtime": 3221.6572,
"eval_hfsummary_samples_per_second": 10.269,
"eval_hfsummary_steps_per_second": 2.054,
"step": 2500
},
{
"epoch": 1.3,
"eval_gptsynthetic_accuracy": 0.9993966817496229,
"eval_gptsynthetic_loss": 0.00219121971167624,
"eval_gptsynthetic_runtime": 101.3035,
"eval_gptsynthetic_samples_per_second": 32.723,
"eval_gptsynthetic_steps_per_second": 6.545,
"step": 2500
}
],
"max_steps": 3842,
"num_train_epochs": 2,
"total_flos": 4.2100836618337517e+17,
"trial_name": null,
"trial_params": null
}