| { | |
| "best_metric": 3.696150779724121, | |
| "best_model_checkpoint": "models/GPT2_bigram_function_42/checkpoint-64200", | |
| "epoch": 10.0, | |
| "global_step": 64200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 1e-05, | |
| "loss": 7.7984, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 2e-05, | |
| "loss": 6.1007, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3e-05, | |
| "loss": 5.6345, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 4e-05, | |
| "loss": 5.3208, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 5e-05, | |
| "loss": 5.0834, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 6e-05, | |
| "loss": 4.8903, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.2576252395931425, | |
| "eval_loss": 4.681421279907227, | |
| "eval_runtime": 2.0164, | |
| "eval_samples_per_second": 584.716, | |
| "eval_steps_per_second": 4.959, | |
| "step": 6420 | |
| }, | |
| { | |
| "epoch": 1.09, | |
| "learning_rate": 7e-05, | |
| "loss": 4.7259, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 8e-05, | |
| "loss": 4.5966, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 9e-05, | |
| "loss": 4.5007, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 0.0001, | |
| "loss": 4.4213, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 9.81568265682657e-05, | |
| "loss": 4.3527, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "learning_rate": 9.631180811808118e-05, | |
| "loss": 4.2915, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.29833102923203303, | |
| "eval_loss": 4.1694016456604, | |
| "eval_runtime": 2.037, | |
| "eval_samples_per_second": 578.794, | |
| "eval_steps_per_second": 4.909, | |
| "step": 12840 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "learning_rate": 9.446863468634687e-05, | |
| "loss": 4.2334, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 2.18, | |
| "learning_rate": 9.262361623616237e-05, | |
| "loss": 4.1641, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 9.078044280442805e-05, | |
| "loss": 4.1351, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 2.49, | |
| "learning_rate": 8.893542435424355e-05, | |
| "loss": 4.1097, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "learning_rate": 8.709225092250924e-05, | |
| "loss": 4.0827, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "learning_rate": 8.524723247232472e-05, | |
| "loss": 4.0578, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 8.340405904059041e-05, | |
| "loss": 4.0376, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.32061068702290074, | |
| "eval_loss": 3.976128578186035, | |
| "eval_runtime": 2.0169, | |
| "eval_samples_per_second": 584.567, | |
| "eval_steps_per_second": 4.958, | |
| "step": 19260 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "learning_rate": 8.155904059040591e-05, | |
| "loss": 3.9746, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 3.27, | |
| "learning_rate": 7.971586715867159e-05, | |
| "loss": 3.953, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "learning_rate": 7.787084870848709e-05, | |
| "loss": 3.943, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 3.58, | |
| "learning_rate": 7.602767527675278e-05, | |
| "loss": 3.9332, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 3.74, | |
| "learning_rate": 7.418265682656826e-05, | |
| "loss": 3.9213, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 3.89, | |
| "learning_rate": 7.233948339483395e-05, | |
| "loss": 3.9123, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.3328257631918148, | |
| "eval_loss": 3.876999616622925, | |
| "eval_runtime": 2.0205, | |
| "eval_samples_per_second": 583.525, | |
| "eval_steps_per_second": 4.949, | |
| "step": 25680 | |
| }, | |
| { | |
| "epoch": 4.05, | |
| "learning_rate": 7.049446494464945e-05, | |
| "loss": 3.8816, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 4.21, | |
| "learning_rate": 6.865129151291513e-05, | |
| "loss": 3.8333, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 4.36, | |
| "learning_rate": 6.680627306273063e-05, | |
| "loss": 3.8317, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 4.52, | |
| "learning_rate": 6.496309963099632e-05, | |
| "loss": 3.8297, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 4.67, | |
| "learning_rate": 6.311808118081182e-05, | |
| "loss": 3.8245, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "learning_rate": 6.12749077490775e-05, | |
| "loss": 3.8195, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 4.98, | |
| "learning_rate": 5.942988929889299e-05, | |
| "loss": 3.8146, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.33965792443883447, | |
| "eval_loss": 3.81557297706604, | |
| "eval_runtime": 2.0122, | |
| "eval_samples_per_second": 585.935, | |
| "eval_steps_per_second": 4.97, | |
| "step": 32100 | |
| }, | |
| { | |
| "epoch": 5.14, | |
| "learning_rate": 5.758671586715867e-05, | |
| "loss": 3.7496, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 5.3, | |
| "learning_rate": 5.574169741697417e-05, | |
| "loss": 3.7491, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 5.45, | |
| "learning_rate": 5.389852398523986e-05, | |
| "loss": 3.7495, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 5.61, | |
| "learning_rate": 5.205350553505536e-05, | |
| "loss": 3.7491, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 5.76, | |
| "learning_rate": 5.0210332103321035e-05, | |
| "loss": 3.749, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 5.92, | |
| "learning_rate": 4.836531365313653e-05, | |
| "loss": 3.7441, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.3458355873454749, | |
| "eval_loss": 3.7725794315338135, | |
| "eval_runtime": 2.0068, | |
| "eval_samples_per_second": 587.504, | |
| "eval_steps_per_second": 4.983, | |
| "step": 38520 | |
| }, | |
| { | |
| "epoch": 6.07, | |
| "learning_rate": 4.652214022140222e-05, | |
| "loss": 3.7094, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 6.23, | |
| "learning_rate": 4.4677121771217715e-05, | |
| "loss": 3.6803, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 6.39, | |
| "learning_rate": 4.28339483394834e-05, | |
| "loss": 3.6849, | |
| "step": 41000 | |
| }, | |
| { | |
| "epoch": 6.54, | |
| "learning_rate": 4.098892988929889e-05, | |
| "loss": 3.6849, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 6.7, | |
| "learning_rate": 3.9145756457564576e-05, | |
| "loss": 3.6875, | |
| "step": 43000 | |
| }, | |
| { | |
| "epoch": 6.85, | |
| "learning_rate": 3.7300738007380073e-05, | |
| "loss": 3.6835, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.3496824347338262, | |
| "eval_loss": 3.7411692142486572, | |
| "eval_runtime": 2.0063, | |
| "eval_samples_per_second": 587.653, | |
| "eval_steps_per_second": 4.984, | |
| "step": 44940 | |
| }, | |
| { | |
| "epoch": 7.01, | |
| "learning_rate": 3.545756457564576e-05, | |
| "loss": 3.6812, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 7.17, | |
| "learning_rate": 3.3612546125461256e-05, | |
| "loss": 3.6234, | |
| "step": 46000 | |
| }, | |
| { | |
| "epoch": 7.32, | |
| "learning_rate": 3.1767527675276754e-05, | |
| "loss": 3.63, | |
| "step": 47000 | |
| }, | |
| { | |
| "epoch": 7.48, | |
| "learning_rate": 2.9924354243542435e-05, | |
| "loss": 3.6306, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 7.63, | |
| "learning_rate": 2.8079335793357936e-05, | |
| "loss": 3.6335, | |
| "step": 49000 | |
| }, | |
| { | |
| "epoch": 7.79, | |
| "learning_rate": 2.6236162361623618e-05, | |
| "loss": 3.6333, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 7.94, | |
| "learning_rate": 2.4391143911439115e-05, | |
| "loss": 3.6319, | |
| "step": 51000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.35245403484869736, | |
| "eval_loss": 3.7189927101135254, | |
| "eval_runtime": 1.9995, | |
| "eval_samples_per_second": 589.657, | |
| "eval_steps_per_second": 5.001, | |
| "step": 51360 | |
| }, | |
| { | |
| "epoch": 8.1, | |
| "learning_rate": 2.2547970479704797e-05, | |
| "loss": 3.5962, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 8.26, | |
| "learning_rate": 2.0702952029520294e-05, | |
| "loss": 3.5825, | |
| "step": 53000 | |
| }, | |
| { | |
| "epoch": 8.41, | |
| "learning_rate": 1.885977859778598e-05, | |
| "loss": 3.5851, | |
| "step": 54000 | |
| }, | |
| { | |
| "epoch": 8.57, | |
| "learning_rate": 1.7014760147601477e-05, | |
| "loss": 3.5844, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 8.72, | |
| "learning_rate": 1.5169741697416975e-05, | |
| "loss": 3.5862, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 8.88, | |
| "learning_rate": 1.3326568265682656e-05, | |
| "loss": 3.5859, | |
| "step": 57000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.35551949136128974, | |
| "eval_loss": 3.7039053440093994, | |
| "eval_runtime": 2.0101, | |
| "eval_samples_per_second": 586.552, | |
| "eval_steps_per_second": 4.975, | |
| "step": 57780 | |
| }, | |
| { | |
| "epoch": 9.03, | |
| "learning_rate": 1.1481549815498155e-05, | |
| "loss": 3.5764, | |
| "step": 58000 | |
| }, | |
| { | |
| "epoch": 9.19, | |
| "learning_rate": 9.638376383763839e-06, | |
| "loss": 3.545, | |
| "step": 59000 | |
| }, | |
| { | |
| "epoch": 9.35, | |
| "learning_rate": 7.793357933579336e-06, | |
| "loss": 3.547, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 9.5, | |
| "learning_rate": 5.948339483394835e-06, | |
| "loss": 3.546, | |
| "step": 61000 | |
| }, | |
| { | |
| "epoch": 9.66, | |
| "learning_rate": 4.105166051660517e-06, | |
| "loss": 3.5476, | |
| "step": 62000 | |
| }, | |
| { | |
| "epoch": 9.81, | |
| "learning_rate": 2.2601476014760147e-06, | |
| "loss": 3.5461, | |
| "step": 63000 | |
| }, | |
| { | |
| "epoch": 9.97, | |
| "learning_rate": 4.1697416974169745e-07, | |
| "loss": 3.5463, | |
| "step": 64000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.3568218094875545, | |
| "eval_loss": 3.696150779724121, | |
| "eval_runtime": 2.0131, | |
| "eval_samples_per_second": 585.653, | |
| "eval_steps_per_second": 4.967, | |
| "step": 64200 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 64200, | |
| "total_flos": 5.367983505408e+17, | |
| "train_loss": 4.021457500517182, | |
| "train_runtime": 29518.126, | |
| "train_samples_per_second": 278.392, | |
| "train_steps_per_second": 2.175 | |
| } | |
| ], | |
| "max_steps": 64200, | |
| "num_train_epochs": 10, | |
| "total_flos": 5.367983505408e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |