| { | |
| "best_metric": 0.7120087742805481, | |
| "best_model_checkpoint": "Action_model/checkpoint-600", | |
| "epoch": 2.0, | |
| "eval_steps": 100, | |
| "global_step": 1256, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.03, | |
| "grad_norm": 2.222489356994629, | |
| "learning_rate": 9.840764331210192e-05, | |
| "loss": 2.2257, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 2.293585777282715, | |
| "learning_rate": 9.681528662420382e-05, | |
| "loss": 2.0112, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 2.167264223098755, | |
| "learning_rate": 9.522292993630574e-05, | |
| "loss": 1.8158, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 3.259340286254883, | |
| "learning_rate": 9.363057324840766e-05, | |
| "loss": 1.5504, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 2.540658473968506, | |
| "learning_rate": 9.203821656050956e-05, | |
| "loss": 1.3489, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "eval_accuracy": 0.7, | |
| "eval_loss": 1.2611500024795532, | |
| "eval_runtime": 22.4862, | |
| "eval_samples_per_second": 46.695, | |
| "eval_steps_per_second": 5.87, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 2.064728021621704, | |
| "learning_rate": 9.044585987261147e-05, | |
| "loss": 1.2181, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 4.9456000328063965, | |
| "learning_rate": 8.885350318471338e-05, | |
| "loss": 1.1517, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 3.7164435386657715, | |
| "learning_rate": 8.73407643312102e-05, | |
| "loss": 1.0429, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "grad_norm": 3.3535468578338623, | |
| "learning_rate": 8.57484076433121e-05, | |
| "loss": 0.9935, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 5.574573040008545, | |
| "learning_rate": 8.415605095541401e-05, | |
| "loss": 1.0112, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "eval_accuracy": 0.7590476190476191, | |
| "eval_loss": 0.9050103425979614, | |
| "eval_runtime": 14.9955, | |
| "eval_samples_per_second": 70.021, | |
| "eval_steps_per_second": 8.803, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 2.9566049575805664, | |
| "learning_rate": 8.256369426751593e-05, | |
| "loss": 0.9295, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 2.9411683082580566, | |
| "learning_rate": 8.097133757961783e-05, | |
| "loss": 0.895, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 4.049664497375488, | |
| "learning_rate": 7.937898089171975e-05, | |
| "loss": 0.8278, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 4.934290409088135, | |
| "learning_rate": 7.778662420382165e-05, | |
| "loss": 0.7667, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 5.120416164398193, | |
| "learning_rate": 7.619426751592357e-05, | |
| "loss": 0.7962, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "eval_accuracy": 0.7504761904761905, | |
| "eval_loss": 0.852245032787323, | |
| "eval_runtime": 14.85, | |
| "eval_samples_per_second": 70.707, | |
| "eval_steps_per_second": 8.889, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 2.324152708053589, | |
| "learning_rate": 7.460191082802548e-05, | |
| "loss": 0.8446, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 6.521075248718262, | |
| "learning_rate": 7.300955414012739e-05, | |
| "loss": 0.6878, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 4.726436614990234, | |
| "learning_rate": 7.14171974522293e-05, | |
| "loss": 0.7465, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 3.800800085067749, | |
| "learning_rate": 6.982484076433122e-05, | |
| "loss": 0.706, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 5.507264614105225, | |
| "learning_rate": 6.823248407643312e-05, | |
| "loss": 0.6383, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_accuracy": 0.7219047619047619, | |
| "eval_loss": 0.8676416277885437, | |
| "eval_runtime": 14.6076, | |
| "eval_samples_per_second": 71.881, | |
| "eval_steps_per_second": 9.036, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 6.359522342681885, | |
| "learning_rate": 6.664012738853504e-05, | |
| "loss": 0.6658, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 3.9945261478424072, | |
| "learning_rate": 6.504777070063695e-05, | |
| "loss": 0.6106, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 2.555899143218994, | |
| "learning_rate": 6.345541401273885e-05, | |
| "loss": 0.7034, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 2.68978214263916, | |
| "learning_rate": 6.186305732484077e-05, | |
| "loss": 0.4986, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 4.49608039855957, | |
| "learning_rate": 6.027070063694268e-05, | |
| "loss": 0.6485, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_accuracy": 0.7323809523809524, | |
| "eval_loss": 0.8052415251731873, | |
| "eval_runtime": 14.5568, | |
| "eval_samples_per_second": 72.131, | |
| "eval_steps_per_second": 9.068, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 5.239855766296387, | |
| "learning_rate": 5.867834394904459e-05, | |
| "loss": 0.6176, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 2.8663668632507324, | |
| "learning_rate": 5.70859872611465e-05, | |
| "loss": 0.5519, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 2.615525245666504, | |
| "learning_rate": 5.5493630573248414e-05, | |
| "loss": 0.6374, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 3.312385082244873, | |
| "learning_rate": 5.3901273885350324e-05, | |
| "loss": 0.5816, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 4.399689197540283, | |
| "learning_rate": 5.230891719745223e-05, | |
| "loss": 0.5452, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "eval_accuracy": 0.7847619047619048, | |
| "eval_loss": 0.7120087742805481, | |
| "eval_runtime": 14.577, | |
| "eval_samples_per_second": 72.031, | |
| "eval_steps_per_second": 9.055, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 3.4874184131622314, | |
| "learning_rate": 5.071656050955414e-05, | |
| "loss": 0.5328, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 1.02, | |
| "grad_norm": 5.2181396484375, | |
| "learning_rate": 4.912420382165605e-05, | |
| "loss": 0.5078, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 2.219102621078491, | |
| "learning_rate": 4.753184713375796e-05, | |
| "loss": 0.4969, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 4.785001754760742, | |
| "learning_rate": 4.593949044585987e-05, | |
| "loss": 0.5407, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "grad_norm": 7.441385269165039, | |
| "learning_rate": 4.4347133757961786e-05, | |
| "loss": 0.4882, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "eval_accuracy": 0.7714285714285715, | |
| "eval_loss": 0.7478358745574951, | |
| "eval_runtime": 14.7271, | |
| "eval_samples_per_second": 71.297, | |
| "eval_steps_per_second": 8.963, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 3.0530927181243896, | |
| "learning_rate": 4.2754777070063695e-05, | |
| "loss": 0.423, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "grad_norm": 3.1082653999328613, | |
| "learning_rate": 4.1162420382165605e-05, | |
| "loss": 0.505, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "grad_norm": 5.5019683837890625, | |
| "learning_rate": 3.957006369426752e-05, | |
| "loss": 0.4445, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 1.24, | |
| "grad_norm": 3.35685658454895, | |
| "learning_rate": 3.797770700636943e-05, | |
| "loss": 0.4795, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "grad_norm": 0.8577423691749573, | |
| "learning_rate": 3.638535031847134e-05, | |
| "loss": 0.3409, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.27, | |
| "eval_accuracy": 0.7742857142857142, | |
| "eval_loss": 0.7310556769371033, | |
| "eval_runtime": 14.6273, | |
| "eval_samples_per_second": 71.784, | |
| "eval_steps_per_second": 9.024, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "grad_norm": 2.747500419616699, | |
| "learning_rate": 3.479299363057325e-05, | |
| "loss": 0.3633, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "grad_norm": 2.4795773029327393, | |
| "learning_rate": 3.3200636942675165e-05, | |
| "loss": 0.4641, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "grad_norm": 5.826427936553955, | |
| "learning_rate": 3.1608280254777074e-05, | |
| "loss": 0.4289, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 4.507148742675781, | |
| "learning_rate": 3.0015923566878983e-05, | |
| "loss": 0.4525, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "grad_norm": 2.810245990753174, | |
| "learning_rate": 2.8423566878980896e-05, | |
| "loss": 0.4105, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "eval_accuracy": 0.780952380952381, | |
| "eval_loss": 0.735313892364502, | |
| "eval_runtime": 14.7897, | |
| "eval_samples_per_second": 70.995, | |
| "eval_steps_per_second": 8.925, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "grad_norm": 3.5758163928985596, | |
| "learning_rate": 2.6831210191082805e-05, | |
| "loss": 0.3657, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "grad_norm": 2.174391031265259, | |
| "learning_rate": 2.5238853503184718e-05, | |
| "loss": 0.3409, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 1.53, | |
| "grad_norm": 3.542391300201416, | |
| "learning_rate": 2.372611464968153e-05, | |
| "loss": 0.3414, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "grad_norm": 4.226655006408691, | |
| "learning_rate": 2.2133757961783442e-05, | |
| "loss": 0.383, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "grad_norm": 5.462564945220947, | |
| "learning_rate": 2.054140127388535e-05, | |
| "loss": 0.4011, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "eval_accuracy": 0.7457142857142857, | |
| "eval_loss": 0.8153719305992126, | |
| "eval_runtime": 14.4617, | |
| "eval_samples_per_second": 72.605, | |
| "eval_steps_per_second": 9.128, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "grad_norm": 6.1501569747924805, | |
| "learning_rate": 1.8949044585987264e-05, | |
| "loss": 0.3402, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "grad_norm": 2.9438650608062744, | |
| "learning_rate": 1.7356687898089173e-05, | |
| "loss": 0.2997, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 1.3817728757858276, | |
| "learning_rate": 1.5764331210191083e-05, | |
| "loss": 0.3485, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 1.72, | |
| "grad_norm": 0.5100256204605103, | |
| "learning_rate": 1.4171974522292993e-05, | |
| "loss": 0.3804, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "grad_norm": 7.605688095092773, | |
| "learning_rate": 1.2579617834394904e-05, | |
| "loss": 0.3493, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "eval_accuracy": 0.7752380952380953, | |
| "eval_loss": 0.7397615313529968, | |
| "eval_runtime": 14.8106, | |
| "eval_samples_per_second": 70.895, | |
| "eval_steps_per_second": 8.913, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "grad_norm": 10.322568893432617, | |
| "learning_rate": 1.0987261146496815e-05, | |
| "loss": 0.4022, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "grad_norm": 5.649250030517578, | |
| "learning_rate": 9.394904458598726e-06, | |
| "loss": 0.2426, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 7.395249366760254, | |
| "learning_rate": 7.802547770700637e-06, | |
| "loss": 0.2628, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 1.7934772968292236, | |
| "learning_rate": 6.210191082802548e-06, | |
| "loss": 0.3818, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "grad_norm": 6.324862480163574, | |
| "learning_rate": 4.6178343949044585e-06, | |
| "loss": 0.3389, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "eval_accuracy": 0.7676190476190476, | |
| "eval_loss": 0.7365464568138123, | |
| "eval_runtime": 14.7187, | |
| "eval_samples_per_second": 71.338, | |
| "eval_steps_per_second": 8.968, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 4.285161018371582, | |
| "learning_rate": 3.0254777070063695e-06, | |
| "loss": 0.3351, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "grad_norm": 4.406313896179199, | |
| "learning_rate": 1.4331210191082802e-06, | |
| "loss": 0.2856, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1256, | |
| "total_flos": 1.555375746295849e+18, | |
| "train_loss": 0.6562361546382782, | |
| "train_runtime": 775.5335, | |
| "train_samples_per_second": 25.879, | |
| "train_steps_per_second": 1.62 | |
| } | |
| ], | |
| "logging_steps": 20, | |
| "max_steps": 1256, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 100, | |
| "total_flos": 1.555375746295849e+18, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |