| { | |
| "best_metric": NaN, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-25", | |
| "epoch": 0.6552006552006552, | |
| "eval_steps": 25, | |
| "global_step": 50, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.013104013104013105, | |
| "grad_norm": 3.4486238956451416, | |
| "learning_rate": 5e-05, | |
| "loss": 1.7472, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.013104013104013105, | |
| "eval_loss": NaN, | |
| "eval_runtime": 6.7355, | |
| "eval_samples_per_second": 76.312, | |
| "eval_steps_per_second": 9.65, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02620802620802621, | |
| "grad_norm": 2.0970981121063232, | |
| "learning_rate": 0.0001, | |
| "loss": 2.1588, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.03931203931203931, | |
| "grad_norm": 2.1514782905578613, | |
| "learning_rate": 9.989294616193017e-05, | |
| "loss": 2.5184, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.05241605241605242, | |
| "grad_norm": 1.7106603384017944, | |
| "learning_rate": 9.957224306869053e-05, | |
| "loss": 2.0056, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.06552006552006552, | |
| "grad_norm": 1.7927194833755493, | |
| "learning_rate": 9.903926402016153e-05, | |
| "loss": 2.1024, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.07862407862407862, | |
| "grad_norm": 2.5896522998809814, | |
| "learning_rate": 9.829629131445342e-05, | |
| "loss": 2.0688, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.09172809172809172, | |
| "grad_norm": 2.71931791305542, | |
| "learning_rate": 9.73465064747553e-05, | |
| "loss": 2.3227, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.10483210483210484, | |
| "grad_norm": 5.414276123046875, | |
| "learning_rate": 9.619397662556435e-05, | |
| "loss": 2.3482, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.11793611793611794, | |
| "grad_norm": 3.869675636291504, | |
| "learning_rate": 9.484363707663442e-05, | |
| "loss": 2.1806, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.13104013104013104, | |
| "grad_norm": 2.9133687019348145, | |
| "learning_rate": 9.330127018922194e-05, | |
| "loss": 2.0149, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.14414414414414414, | |
| "grad_norm": 3.4659268856048584, | |
| "learning_rate": 9.157348061512727e-05, | |
| "loss": 2.2198, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.15724815724815724, | |
| "grad_norm": 4.818014621734619, | |
| "learning_rate": 8.966766701456177e-05, | |
| "loss": 2.0188, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.17035217035217035, | |
| "grad_norm": 2.0361897945404053, | |
| "learning_rate": 8.759199037394887e-05, | |
| "loss": 1.3849, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.18345618345618345, | |
| "grad_norm": 2.208857774734497, | |
| "learning_rate": 8.535533905932738e-05, | |
| "loss": 1.4099, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.19656019656019655, | |
| "grad_norm": 1.992568850517273, | |
| "learning_rate": 8.296729075500344e-05, | |
| "loss": 1.8764, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.20966420966420968, | |
| "grad_norm": 1.5991078615188599, | |
| "learning_rate": 8.043807145043604e-05, | |
| "loss": 1.5679, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.22276822276822278, | |
| "grad_norm": 1.3598281145095825, | |
| "learning_rate": 7.777851165098012e-05, | |
| "loss": 1.5211, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.23587223587223588, | |
| "grad_norm": 1.720762014389038, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 1.5706, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.24897624897624898, | |
| "grad_norm": 1.6641697883605957, | |
| "learning_rate": 7.211443451095007e-05, | |
| "loss": 1.4383, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.2620802620802621, | |
| "grad_norm": 1.7044271230697632, | |
| "learning_rate": 6.91341716182545e-05, | |
| "loss": 1.6984, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2751842751842752, | |
| "grad_norm": 2.450592279434204, | |
| "learning_rate": 6.607197326515808e-05, | |
| "loss": 1.2786, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.2882882882882883, | |
| "grad_norm": 2.3138625621795654, | |
| "learning_rate": 6.294095225512603e-05, | |
| "loss": 1.3847, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.3013923013923014, | |
| "grad_norm": 2.143385410308838, | |
| "learning_rate": 5.9754516100806423e-05, | |
| "loss": 1.6139, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.3144963144963145, | |
| "grad_norm": 2.384326219558716, | |
| "learning_rate": 5.6526309611002594e-05, | |
| "loss": 1.6064, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.3276003276003276, | |
| "grad_norm": 3.7034049034118652, | |
| "learning_rate": 5.327015646150716e-05, | |
| "loss": 1.9488, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.3276003276003276, | |
| "eval_loss": NaN, | |
| "eval_runtime": 6.7516, | |
| "eval_samples_per_second": 76.13, | |
| "eval_steps_per_second": 9.627, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.3407043407043407, | |
| "grad_norm": 1.821609616279602, | |
| "learning_rate": 5e-05, | |
| "loss": 0.9843, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.3538083538083538, | |
| "grad_norm": 1.8066356182098389, | |
| "learning_rate": 4.6729843538492847e-05, | |
| "loss": 1.3468, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.3669123669123669, | |
| "grad_norm": 1.5410196781158447, | |
| "learning_rate": 4.347369038899744e-05, | |
| "loss": 1.4991, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.38001638001638, | |
| "grad_norm": 1.6339073181152344, | |
| "learning_rate": 4.0245483899193595e-05, | |
| "loss": 1.1596, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.3931203931203931, | |
| "grad_norm": 1.6657391786575317, | |
| "learning_rate": 3.705904774487396e-05, | |
| "loss": 1.363, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.4062244062244062, | |
| "grad_norm": 1.8848774433135986, | |
| "learning_rate": 3.392802673484193e-05, | |
| "loss": 1.012, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.41932841932841936, | |
| "grad_norm": 1.8691872358322144, | |
| "learning_rate": 3.086582838174551e-05, | |
| "loss": 1.4458, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.43243243243243246, | |
| "grad_norm": 2.03696346282959, | |
| "learning_rate": 2.7885565489049946e-05, | |
| "loss": 1.1339, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.44553644553644556, | |
| "grad_norm": 2.19081711769104, | |
| "learning_rate": 2.500000000000001e-05, | |
| "loss": 1.0733, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.45864045864045866, | |
| "grad_norm": 2.4450438022613525, | |
| "learning_rate": 2.2221488349019903e-05, | |
| "loss": 1.2061, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.47174447174447176, | |
| "grad_norm": 2.2913472652435303, | |
| "learning_rate": 1.9561928549563968e-05, | |
| "loss": 1.1912, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.48484848484848486, | |
| "grad_norm": 2.4383180141448975, | |
| "learning_rate": 1.703270924499656e-05, | |
| "loss": 1.1965, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.49795249795249796, | |
| "grad_norm": 1.6899724006652832, | |
| "learning_rate": 1.4644660940672627e-05, | |
| "loss": 1.0133, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.5110565110565111, | |
| "grad_norm": 1.4650294780731201, | |
| "learning_rate": 1.2408009626051137e-05, | |
| "loss": 0.9544, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.5241605241605242, | |
| "grad_norm": 1.6712995767593384, | |
| "learning_rate": 1.0332332985438248e-05, | |
| "loss": 1.4383, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.5372645372645373, | |
| "grad_norm": 1.7614799737930298, | |
| "learning_rate": 8.426519384872733e-06, | |
| "loss": 1.234, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.5503685503685504, | |
| "grad_norm": 1.5106946229934692, | |
| "learning_rate": 6.698729810778065e-06, | |
| "loss": 1.3066, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.5634725634725635, | |
| "grad_norm": 1.6505160331726074, | |
| "learning_rate": 5.156362923365588e-06, | |
| "loss": 1.1251, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.5765765765765766, | |
| "grad_norm": 1.6521756649017334, | |
| "learning_rate": 3.8060233744356633e-06, | |
| "loss": 1.3106, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.5896805896805897, | |
| "grad_norm": 1.6643149852752686, | |
| "learning_rate": 2.653493525244721e-06, | |
| "loss": 1.0814, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.6027846027846028, | |
| "grad_norm": 1.5773600339889526, | |
| "learning_rate": 1.70370868554659e-06, | |
| "loss": 0.9147, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.6158886158886159, | |
| "grad_norm": 1.7343523502349854, | |
| "learning_rate": 9.607359798384785e-07, | |
| "loss": 1.0469, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.628992628992629, | |
| "grad_norm": 1.9467123746871948, | |
| "learning_rate": 4.277569313094809e-07, | |
| "loss": 1.0538, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.6420966420966421, | |
| "grad_norm": 1.9722298383712769, | |
| "learning_rate": 1.0705383806982606e-07, | |
| "loss": 1.0307, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.6552006552006552, | |
| "grad_norm": 2.531688690185547, | |
| "learning_rate": 0.0, | |
| "loss": 1.524, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.6552006552006552, | |
| "eval_loss": NaN, | |
| "eval_runtime": 6.7523, | |
| "eval_samples_per_second": 76.122, | |
| "eval_steps_per_second": 9.626, | |
| "step": 50 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 50, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 25, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 1, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 1 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.589463780163584e+17, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |