| { | |
| "best_metric": 0.030559765174984932, | |
| "best_model_checkpoint": "/data/wheld3/mt5-small-pointer-top_v2/checkpoint-3000", | |
| "epoch": 12.345168539325842, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 0.0009333333333333333, | |
| "loss": 1.9316, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "eval_exact_match": 0.008391608391608392, | |
| "eval_loss": 0.45658260583877563, | |
| "eval_runtime": 661.3564, | |
| "eval_samples_per_second": 25.947, | |
| "eval_steps_per_second": 3.243, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 0.0008666666666666667, | |
| "loss": 0.3713, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "eval_exact_match": 0.12301864801864802, | |
| "eval_loss": 0.14726108312606812, | |
| "eval_runtime": 655.0457, | |
| "eval_samples_per_second": 26.197, | |
| "eval_steps_per_second": 3.275, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 0.0008, | |
| "loss": 0.1747, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "eval_exact_match": 0.19842657342657344, | |
| "eval_loss": 0.07877045124769211, | |
| "eval_runtime": 646.2259, | |
| "eval_samples_per_second": 26.554, | |
| "eval_steps_per_second": 3.319, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "learning_rate": 0.0007333333333333333, | |
| "loss": 0.1104, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "eval_exact_match": 0.21486013986013985, | |
| "eval_loss": 0.056847672909498215, | |
| "eval_runtime": 648.473, | |
| "eval_samples_per_second": 26.462, | |
| "eval_steps_per_second": 3.308, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "learning_rate": 0.0006666666666666666, | |
| "loss": 0.0842, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.12, | |
| "eval_exact_match": 0.22173659673659674, | |
| "eval_loss": 0.04728136211633682, | |
| "eval_runtime": 656.2209, | |
| "eval_samples_per_second": 26.15, | |
| "eval_steps_per_second": 3.269, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.94, | |
| "learning_rate": 0.0006, | |
| "loss": 0.0694, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 4.94, | |
| "eval_exact_match": 0.225990675990676, | |
| "eval_loss": 0.0425742082297802, | |
| "eval_runtime": 650.7447, | |
| "eval_samples_per_second": 26.37, | |
| "eval_steps_per_second": 3.296, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 5.76, | |
| "learning_rate": 0.0005333333333333334, | |
| "loss": 0.0603, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 5.76, | |
| "eval_exact_match": 0.2279137529137529, | |
| "eval_loss": 0.03834143653512001, | |
| "eval_runtime": 652.2239, | |
| "eval_samples_per_second": 26.31, | |
| "eval_steps_per_second": 3.289, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 6.58, | |
| "learning_rate": 0.00046666666666666666, | |
| "loss": 0.0534, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 6.58, | |
| "eval_exact_match": 0.22808857808857808, | |
| "eval_loss": 0.03673423081636429, | |
| "eval_runtime": 652.7536, | |
| "eval_samples_per_second": 26.289, | |
| "eval_steps_per_second": 3.286, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 7.41, | |
| "learning_rate": 0.0004, | |
| "loss": 0.0477, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 7.41, | |
| "eval_exact_match": 0.23006993006993007, | |
| "eval_loss": 0.03471648693084717, | |
| "eval_runtime": 664.2219, | |
| "eval_samples_per_second": 25.835, | |
| "eval_steps_per_second": 3.229, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 8.23, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.0441, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 8.23, | |
| "eval_exact_match": 0.23135198135198135, | |
| "eval_loss": 0.03336101025342941, | |
| "eval_runtime": 672.8476, | |
| "eval_samples_per_second": 25.504, | |
| "eval_steps_per_second": 3.188, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.0413, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 9.05, | |
| "eval_exact_match": 0.23146853146853147, | |
| "eval_loss": 0.03233984857797623, | |
| "eval_runtime": 660.8292, | |
| "eval_samples_per_second": 25.967, | |
| "eval_steps_per_second": 3.246, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 9.88, | |
| "learning_rate": 0.0002, | |
| "loss": 0.0387, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 9.88, | |
| "eval_exact_match": 0.2315850815850816, | |
| "eval_loss": 0.03159063309431076, | |
| "eval_runtime": 652.4785, | |
| "eval_samples_per_second": 26.3, | |
| "eval_steps_per_second": 3.287, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 10.7, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.0366, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 10.7, | |
| "eval_exact_match": 0.2324009324009324, | |
| "eval_loss": 0.03113115206360817, | |
| "eval_runtime": 657.1693, | |
| "eval_samples_per_second": 26.112, | |
| "eval_steps_per_second": 3.264, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 11.52, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.0358, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 11.52, | |
| "eval_exact_match": 0.2324009324009324, | |
| "eval_loss": 0.03069169819355011, | |
| "eval_runtime": 652.5597, | |
| "eval_samples_per_second": 26.296, | |
| "eval_steps_per_second": 3.287, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 12.35, | |
| "learning_rate": 0.0, | |
| "loss": 0.0343, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 12.35, | |
| "eval_exact_match": 0.2326923076923077, | |
| "eval_loss": 0.030559765174984932, | |
| "eval_runtime": 648.4066, | |
| "eval_samples_per_second": 26.465, | |
| "eval_steps_per_second": 3.308, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 12.35, | |
| "step": 3000, | |
| "total_flos": 3.654446752585728e+16, | |
| "train_loss": 0.2089100898106893, | |
| "train_runtime": 30994.3996, | |
| "train_samples_per_second": 49.557, | |
| "train_steps_per_second": 0.097 | |
| } | |
| ], | |
| "max_steps": 3000, | |
| "num_train_epochs": 13, | |
| "total_flos": 3.654446752585728e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |