| { | |
| "best_metric": 0.02550993673503399, | |
| "best_model_checkpoint": "/data/wheld3/mt5-base-pointer-adv-top_v2/checkpoint-3000", | |
| "epoch": 6.160056430151656, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 0.0009333333333333333, | |
| "loss": 2.2938, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "eval_exact_match": 0.0011655011655011655, | |
| "eval_loss": 0.5532176494598389, | |
| "eval_runtime": 914.7564, | |
| "eval_samples_per_second": 18.759, | |
| "eval_steps_per_second": 2.345, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 0.0008666666666666667, | |
| "loss": 0.671, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "eval_exact_match": 0.16095571095571096, | |
| "eval_loss": 0.16243064403533936, | |
| "eval_runtime": 902.0916, | |
| "eval_samples_per_second": 19.022, | |
| "eval_steps_per_second": 2.378, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "learning_rate": 0.0008, | |
| "loss": 0.5276, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "eval_exact_match": 0.21567599067599066, | |
| "eval_loss": 0.06916385143995285, | |
| "eval_runtime": 882.9446, | |
| "eval_samples_per_second": 19.435, | |
| "eval_steps_per_second": 2.429, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "learning_rate": 0.0007333333333333333, | |
| "loss": 0.4196, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "eval_exact_match": 0.22587412587412586, | |
| "eval_loss": 0.04908030480146408, | |
| "eval_runtime": 882.5419, | |
| "eval_samples_per_second": 19.444, | |
| "eval_steps_per_second": 2.43, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "learning_rate": 0.0006666666666666666, | |
| "loss": 0.3593, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "eval_exact_match": 0.22913752913752913, | |
| "eval_loss": 0.03997402638196945, | |
| "eval_runtime": 888.3883, | |
| "eval_samples_per_second": 19.316, | |
| "eval_steps_per_second": 2.414, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "learning_rate": 0.0006, | |
| "loss": 0.3471, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "eval_exact_match": 0.22966200466200465, | |
| "eval_loss": 0.03349088877439499, | |
| "eval_runtime": 876.9206, | |
| "eval_samples_per_second": 19.568, | |
| "eval_steps_per_second": 2.446, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "learning_rate": 0.0005333333333333334, | |
| "loss": 0.3416, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "eval_exact_match": 0.23175990675990676, | |
| "eval_loss": 0.03065803460776806, | |
| "eval_runtime": 880.241, | |
| "eval_samples_per_second": 19.495, | |
| "eval_steps_per_second": 2.437, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "learning_rate": 0.00046666666666666666, | |
| "loss": 0.3351, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "eval_exact_match": 0.23344988344988346, | |
| "eval_loss": 0.030744880437850952, | |
| "eval_runtime": 884.2271, | |
| "eval_samples_per_second": 19.407, | |
| "eval_steps_per_second": 2.426, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 3.7, | |
| "learning_rate": 0.0004, | |
| "loss": 0.3316, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 3.7, | |
| "eval_exact_match": 0.23432400932400932, | |
| "eval_loss": 0.029669322073459625, | |
| "eval_runtime": 876.7173, | |
| "eval_samples_per_second": 19.573, | |
| "eval_steps_per_second": 2.447, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 4.11, | |
| "learning_rate": 0.0003333333333333333, | |
| "loss": 0.3312, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.11, | |
| "eval_exact_match": 0.23444055944055944, | |
| "eval_loss": 0.02815121039748192, | |
| "eval_runtime": 885.0693, | |
| "eval_samples_per_second": 19.388, | |
| "eval_steps_per_second": 2.424, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.52, | |
| "learning_rate": 0.0002666666666666667, | |
| "loss": 0.3271, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 4.52, | |
| "eval_exact_match": 0.23648018648018648, | |
| "eval_loss": 0.02621879242360592, | |
| "eval_runtime": 890.2789, | |
| "eval_samples_per_second": 19.275, | |
| "eval_steps_per_second": 2.409, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 4.93, | |
| "learning_rate": 0.0002, | |
| "loss": 0.3241, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 4.93, | |
| "eval_exact_match": 0.23653846153846153, | |
| "eval_loss": 0.02629098668694496, | |
| "eval_runtime": 876.3961, | |
| "eval_samples_per_second": 19.58, | |
| "eval_steps_per_second": 2.448, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 5.34, | |
| "learning_rate": 0.00013333333333333334, | |
| "loss": 0.3227, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 5.34, | |
| "eval_exact_match": 0.23677156177156178, | |
| "eval_loss": 0.02586781419813633, | |
| "eval_runtime": 890.1947, | |
| "eval_samples_per_second": 19.277, | |
| "eval_steps_per_second": 2.41, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 5.75, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 0.3201, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 5.75, | |
| "eval_exact_match": 0.23653846153846153, | |
| "eval_loss": 0.025654641911387444, | |
| "eval_runtime": 876.5182, | |
| "eval_samples_per_second": 19.577, | |
| "eval_steps_per_second": 2.447, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 6.16, | |
| "learning_rate": 0.0, | |
| "loss": 0.3227, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 6.16, | |
| "eval_exact_match": 0.23653846153846153, | |
| "eval_loss": 0.02550993673503399, | |
| "eval_runtime": 881.7574, | |
| "eval_samples_per_second": 19.461, | |
| "eval_steps_per_second": 2.433, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 6.16, | |
| "step": 3000, | |
| "total_flos": 8.2913691900475e+16, | |
| "train_loss": 0.5049722137451171, | |
| "train_runtime": 47265.062, | |
| "train_samples_per_second": 32.498, | |
| "train_steps_per_second": 0.063 | |
| } | |
| ], | |
| "max_steps": 3000, | |
| "num_train_epochs": 7, | |
| "total_flos": 8.2913691900475e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |