| { | |
| "best_metric": 0.12020513415336609, | |
| "best_model_checkpoint": "/data/wheld3/mt5-small-pointer-mtop/checkpoint-1200", | |
| "epoch": 99.9795918367347, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 6.65, | |
| "learning_rate": 0.0009333333333333333, | |
| "loss": 2.1451, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 6.65, | |
| "eval_exact_match": 0.013422818791946308, | |
| "eval_loss": 0.5965703725814819, | |
| "eval_runtime": 44.8708, | |
| "eval_samples_per_second": 49.81, | |
| "eval_steps_per_second": 3.12, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 13.33, | |
| "learning_rate": 0.0008666666666666667, | |
| "loss": 0.4695, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 13.33, | |
| "eval_exact_match": 0.29977628635346754, | |
| "eval_loss": 0.22643320262432098, | |
| "eval_runtime": 44.5288, | |
| "eval_samples_per_second": 50.192, | |
| "eval_steps_per_second": 3.144, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 19.98, | |
| "learning_rate": 0.0008, | |
| "loss": 0.2229, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 19.98, | |
| "eval_exact_match": 0.46487695749440716, | |
| "eval_loss": 0.14458051323890686, | |
| "eval_runtime": 47.7042, | |
| "eval_samples_per_second": 46.851, | |
| "eval_steps_per_second": 2.935, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 26.65, | |
| "learning_rate": 0.0007333333333333333, | |
| "loss": 0.1389, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 26.65, | |
| "eval_exact_match": 0.5154362416107383, | |
| "eval_loss": 0.12272302061319351, | |
| "eval_runtime": 44.4647, | |
| "eval_samples_per_second": 50.265, | |
| "eval_steps_per_second": 3.149, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "learning_rate": 0.0006666666666666666, | |
| "loss": 0.097, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 33.33, | |
| "eval_exact_match": 0.5221476510067115, | |
| "eval_loss": 0.12126713246107101, | |
| "eval_runtime": 46.4733, | |
| "eval_samples_per_second": 48.092, | |
| "eval_steps_per_second": 3.012, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 39.98, | |
| "learning_rate": 0.0006, | |
| "loss": 0.0724, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 39.98, | |
| "eval_exact_match": 0.5364653243847874, | |
| "eval_loss": 0.12020513415336609, | |
| "eval_runtime": 44.6036, | |
| "eval_samples_per_second": 50.108, | |
| "eval_steps_per_second": 3.139, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 46.65, | |
| "learning_rate": 0.0005333333333333334, | |
| "loss": 0.0562, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 46.65, | |
| "eval_exact_match": 0.5436241610738255, | |
| "eval_loss": 0.12071848660707474, | |
| "eval_runtime": 44.8086, | |
| "eval_samples_per_second": 49.879, | |
| "eval_steps_per_second": 3.124, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 53.33, | |
| "learning_rate": 0.00046666666666666666, | |
| "loss": 0.0457, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 53.33, | |
| "eval_exact_match": 0.5440715883668904, | |
| "eval_loss": 0.12402455508708954, | |
| "eval_runtime": 44.3425, | |
| "eval_samples_per_second": 50.403, | |
| "eval_steps_per_second": 3.157, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 59.98, | |
| "learning_rate": 0.0004, | |
| "loss": 0.0399, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 59.98, | |
| "eval_exact_match": 0.5440715883668904, | |
| "eval_loss": 0.13486850261688232, | |
| "eval_runtime": 44.9861, | |
| "eval_samples_per_second": 49.682, | |
| "eval_steps_per_second": 3.112, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 66.65, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.0317, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 66.65, | |
| "eval_exact_match": 0.5476510067114094, | |
| "eval_loss": 0.13694295287132263, | |
| "eval_runtime": 45.1401, | |
| "eval_samples_per_second": 49.513, | |
| "eval_steps_per_second": 3.101, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 73.33, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.0271, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 73.33, | |
| "eval_exact_match": 0.548993288590604, | |
| "eval_loss": 0.1409003585577011, | |
| "eval_runtime": 44.2099, | |
| "eval_samples_per_second": 50.554, | |
| "eval_steps_per_second": 3.167, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 79.98, | |
| "learning_rate": 0.0002, | |
| "loss": 0.0237, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 79.98, | |
| "eval_exact_match": 0.5539149888143177, | |
| "eval_loss": 0.14622744917869568, | |
| "eval_runtime": 45.1454, | |
| "eval_samples_per_second": 49.507, | |
| "eval_steps_per_second": 3.101, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 86.65, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.0207, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 86.65, | |
| "eval_exact_match": 0.5516778523489932, | |
| "eval_loss": 0.14700742065906525, | |
| "eval_runtime": 44.6922, | |
| "eval_samples_per_second": 50.009, | |
| "eval_steps_per_second": 3.133, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 93.33, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.0188, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 93.33, | |
| "eval_exact_match": 0.5507829977628635, | |
| "eval_loss": 0.150468111038208, | |
| "eval_runtime": 44.5533, | |
| "eval_samples_per_second": 50.165, | |
| "eval_steps_per_second": 3.142, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 99.98, | |
| "learning_rate": 0.0, | |
| "loss": 0.0174, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 99.98, | |
| "eval_exact_match": 0.5512304250559285, | |
| "eval_loss": 0.15047426521778107, | |
| "eval_runtime": 46.3824, | |
| "eval_samples_per_second": 48.186, | |
| "eval_steps_per_second": 3.018, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 99.98, | |
| "step": 3000, | |
| "total_flos": 3.831174875254272e+16, | |
| "train_loss": 0.22847254665692648, | |
| "train_runtime": 13959.7055, | |
| "train_samples_per_second": 110.031, | |
| "train_steps_per_second": 0.215 | |
| } | |
| ], | |
| "max_steps": 3000, | |
| "num_train_epochs": 100, | |
| "total_flos": 3.831174875254272e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |