| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.0, | |
| "eval_steps": 50, | |
| "global_step": 2000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 7.943530082702637, | |
| "learning_rate": 1.0000000000000002e-06, | |
| "loss": 1.8114, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "eval_loss": 1.1615628004074097, | |
| "eval_runtime": 3.153, | |
| "eval_samples_per_second": 49.476, | |
| "eval_steps_per_second": 2.537, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 5.800487518310547, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "loss": 0.9446, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "eval_loss": 0.8041068315505981, | |
| "eval_runtime": 3.0734, | |
| "eval_samples_per_second": 50.759, | |
| "eval_steps_per_second": 2.603, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 3.3142590522766113, | |
| "learning_rate": 3e-06, | |
| "loss": 0.829, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "eval_loss": 0.7686946392059326, | |
| "eval_runtime": 3.0993, | |
| "eval_samples_per_second": 50.334, | |
| "eval_steps_per_second": 2.581, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 4.635634422302246, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 0.8045, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_loss": 0.754519522190094, | |
| "eval_runtime": 3.0961, | |
| "eval_samples_per_second": 50.386, | |
| "eval_steps_per_second": 2.584, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 3.970125913619995, | |
| "learning_rate": 5e-06, | |
| "loss": 0.7957, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "eval_loss": 0.7530429363250732, | |
| "eval_runtime": 3.0913, | |
| "eval_samples_per_second": 50.465, | |
| "eval_steps_per_second": 2.588, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 3.469130516052246, | |
| "learning_rate": 6e-06, | |
| "loss": 0.7833, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "eval_loss": 0.7474463582038879, | |
| "eval_runtime": 3.0973, | |
| "eval_samples_per_second": 50.366, | |
| "eval_steps_per_second": 2.583, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 2.752779722213745, | |
| "learning_rate": 7e-06, | |
| "loss": 0.8005, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "eval_loss": 0.7482121586799622, | |
| "eval_runtime": 3.0919, | |
| "eval_samples_per_second": 50.454, | |
| "eval_steps_per_second": 2.587, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 2.9983832836151123, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 0.771, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_loss": 0.7438180446624756, | |
| "eval_runtime": 3.0859, | |
| "eval_samples_per_second": 50.553, | |
| "eval_steps_per_second": 2.592, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 2.574100971221924, | |
| "learning_rate": 9e-06, | |
| "loss": 0.7449, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "eval_loss": 0.7498941421508789, | |
| "eval_runtime": 3.0388, | |
| "eval_samples_per_second": 51.336, | |
| "eval_steps_per_second": 2.633, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 2.4326021671295166, | |
| "learning_rate": 1e-05, | |
| "loss": 0.7533, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "eval_loss": 0.7558138966560364, | |
| "eval_runtime": 3.0822, | |
| "eval_samples_per_second": 50.614, | |
| "eval_steps_per_second": 2.596, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 2.791947364807129, | |
| "learning_rate": 1.1000000000000001e-05, | |
| "loss": 0.8062, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "eval_loss": 0.7553198337554932, | |
| "eval_runtime": 3.0823, | |
| "eval_samples_per_second": 50.611, | |
| "eval_steps_per_second": 2.595, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 2.409931182861328, | |
| "learning_rate": 1.2e-05, | |
| "loss": 0.8354, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_loss": 0.7591729760169983, | |
| "eval_runtime": 3.0925, | |
| "eval_samples_per_second": 50.444, | |
| "eval_steps_per_second": 2.587, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 2.654364585876465, | |
| "learning_rate": 1.3000000000000001e-05, | |
| "loss": 0.8084, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "eval_loss": 0.7605798244476318, | |
| "eval_runtime": 3.0938, | |
| "eval_samples_per_second": 50.423, | |
| "eval_steps_per_second": 2.586, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 2.8715052604675293, | |
| "learning_rate": 1.4e-05, | |
| "loss": 0.7464, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "eval_loss": 0.768295168876648, | |
| "eval_runtime": 3.0904, | |
| "eval_samples_per_second": 50.479, | |
| "eval_steps_per_second": 2.589, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 2.547102451324463, | |
| "learning_rate": 1.5000000000000002e-05, | |
| "loss": 0.7652, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "eval_loss": 0.7770843505859375, | |
| "eval_runtime": 3.0851, | |
| "eval_samples_per_second": 50.566, | |
| "eval_steps_per_second": 2.593, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 2.6396355628967285, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 0.7807, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_loss": 0.7774565815925598, | |
| "eval_runtime": 3.0842, | |
| "eval_samples_per_second": 50.581, | |
| "eval_steps_per_second": 2.594, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 2.243042230606079, | |
| "learning_rate": 1.7e-05, | |
| "loss": 0.8038, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "eval_loss": 0.778019905090332, | |
| "eval_runtime": 3.0739, | |
| "eval_samples_per_second": 50.75, | |
| "eval_steps_per_second": 2.603, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 2.5405020713806152, | |
| "learning_rate": 1.8e-05, | |
| "loss": 0.8292, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "eval_loss": 0.7810414433479309, | |
| "eval_runtime": 3.0805, | |
| "eval_samples_per_second": 50.641, | |
| "eval_steps_per_second": 2.597, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 2.9373252391815186, | |
| "learning_rate": 1.9e-05, | |
| "loss": 0.8057, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "eval_loss": 0.7795991897583008, | |
| "eval_runtime": 3.0813, | |
| "eval_samples_per_second": 50.627, | |
| "eval_steps_per_second": 2.596, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 2.0142874717712402, | |
| "learning_rate": 2e-05, | |
| "loss": 0.8265, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.7865928411483765, | |
| "eval_runtime": 3.0922, | |
| "eval_samples_per_second": 50.449, | |
| "eval_steps_per_second": 2.587, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 2.4737377166748047, | |
| "learning_rate": 1.9998476951563914e-05, | |
| "loss": 0.567, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "eval_loss": 0.8242304921150208, | |
| "eval_runtime": 3.0914, | |
| "eval_samples_per_second": 50.462, | |
| "eval_steps_per_second": 2.588, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "grad_norm": 1.9739576578140259, | |
| "learning_rate": 1.999390827019096e-05, | |
| "loss": 0.5251, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "eval_loss": 0.8305041193962097, | |
| "eval_runtime": 3.0703, | |
| "eval_samples_per_second": 50.81, | |
| "eval_steps_per_second": 2.606, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 3.079801321029663, | |
| "learning_rate": 1.9986295347545738e-05, | |
| "loss": 0.5836, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "eval_loss": 0.8230717778205872, | |
| "eval_runtime": 3.0905, | |
| "eval_samples_per_second": 50.478, | |
| "eval_steps_per_second": 2.589, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 2.380758047103882, | |
| "learning_rate": 1.9975640502598243e-05, | |
| "loss": 0.5404, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "eval_loss": 0.8192301392555237, | |
| "eval_runtime": 3.0886, | |
| "eval_samples_per_second": 50.508, | |
| "eval_steps_per_second": 2.59, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 2.047494649887085, | |
| "learning_rate": 1.9961946980917457e-05, | |
| "loss": 0.5448, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "eval_loss": 0.8297872543334961, | |
| "eval_runtime": 3.0696, | |
| "eval_samples_per_second": 50.821, | |
| "eval_steps_per_second": 2.606, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "grad_norm": 2.4927003383636475, | |
| "learning_rate": 1.9945218953682736e-05, | |
| "loss": 0.5505, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "eval_loss": 0.8312389850616455, | |
| "eval_runtime": 3.0792, | |
| "eval_samples_per_second": 50.662, | |
| "eval_steps_per_second": 2.598, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "grad_norm": 3.409341812133789, | |
| "learning_rate": 1.9925461516413224e-05, | |
| "loss": 0.5726, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "eval_loss": 0.8322041630744934, | |
| "eval_runtime": 3.0883, | |
| "eval_samples_per_second": 50.514, | |
| "eval_steps_per_second": 2.59, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 1.5658016204833984, | |
| "learning_rate": 1.9902680687415704e-05, | |
| "loss": 0.5687, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "eval_loss": 0.8255723118782043, | |
| "eval_runtime": 3.0902, | |
| "eval_samples_per_second": 50.481, | |
| "eval_steps_per_second": 2.589, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 2.2273943424224854, | |
| "learning_rate": 1.9876883405951378e-05, | |
| "loss": 0.5768, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "eval_loss": 0.8274821639060974, | |
| "eval_runtime": 3.0903, | |
| "eval_samples_per_second": 50.48, | |
| "eval_steps_per_second": 2.589, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "grad_norm": 1.8873156309127808, | |
| "learning_rate": 1.9848077530122083e-05, | |
| "loss": 0.5437, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "eval_loss": 0.8350476026535034, | |
| "eval_runtime": 3.0892, | |
| "eval_samples_per_second": 50.498, | |
| "eval_steps_per_second": 2.59, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 1.8373135328292847, | |
| "learning_rate": 1.9816271834476642e-05, | |
| "loss": 0.569, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "eval_loss": 0.8110507726669312, | |
| "eval_runtime": 3.096, | |
| "eval_samples_per_second": 50.388, | |
| "eval_steps_per_second": 2.584, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 2.555246591567993, | |
| "learning_rate": 1.9781476007338058e-05, | |
| "loss": 0.5672, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_loss": 0.8229615688323975, | |
| "eval_runtime": 3.1033, | |
| "eval_samples_per_second": 50.269, | |
| "eval_steps_per_second": 2.578, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "grad_norm": 2.032984972000122, | |
| "learning_rate": 1.9743700647852356e-05, | |
| "loss": 0.5709, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "eval_loss": 0.8179122805595398, | |
| "eval_runtime": 3.0703, | |
| "eval_samples_per_second": 50.81, | |
| "eval_steps_per_second": 2.606, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "grad_norm": 2.0338752269744873, | |
| "learning_rate": 1.9702957262759964e-05, | |
| "loss": 0.5945, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "eval_loss": 0.8166740536689758, | |
| "eval_runtime": 3.0576, | |
| "eval_samples_per_second": 51.021, | |
| "eval_steps_per_second": 2.616, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "grad_norm": 2.7427613735198975, | |
| "learning_rate": 1.9659258262890683e-05, | |
| "loss": 0.5756, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "eval_loss": 0.8103773593902588, | |
| "eval_runtime": 3.0935, | |
| "eval_samples_per_second": 50.429, | |
| "eval_steps_per_second": 2.586, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "grad_norm": 2.876005172729492, | |
| "learning_rate": 1.961261695938319e-05, | |
| "loss": 0.5701, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "eval_loss": 0.8226655125617981, | |
| "eval_runtime": 3.099, | |
| "eval_samples_per_second": 50.338, | |
| "eval_steps_per_second": 2.581, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 2.338442087173462, | |
| "learning_rate": 1.9563047559630356e-05, | |
| "loss": 0.5765, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "eval_loss": 0.8165319561958313, | |
| "eval_runtime": 3.0865, | |
| "eval_samples_per_second": 50.543, | |
| "eval_steps_per_second": 2.592, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 1.9952592849731445, | |
| "learning_rate": 1.9510565162951538e-05, | |
| "loss": 0.6023, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "eval_loss": 0.8059903979301453, | |
| "eval_runtime": 3.057, | |
| "eval_samples_per_second": 51.03, | |
| "eval_steps_per_second": 2.617, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 2.47471284866333, | |
| "learning_rate": 1.945518575599317e-05, | |
| "loss": 0.5544, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "eval_loss": 0.8095174431800842, | |
| "eval_runtime": 3.1028, | |
| "eval_samples_per_second": 50.277, | |
| "eval_steps_per_second": 2.578, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 2.391505002975464, | |
| "learning_rate": 1.9396926207859085e-05, | |
| "loss": 0.5932, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.8160108923912048, | |
| "eval_runtime": 3.0868, | |
| "eval_samples_per_second": 50.537, | |
| "eval_steps_per_second": 2.592, | |
| "step": 2000 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 10000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 400, | |
| "total_flos": 1.140965468316631e+17, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |