| { | |
| "best_metric": 0.47900134325027466, | |
| "best_model_checkpoint": "Action_model/checkpoint-800", | |
| "epoch": 10.0, | |
| "eval_steps": 100, | |
| "global_step": 1340, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 1.667541265487671, | |
| "learning_rate": 9.850746268656717e-05, | |
| "loss": 1.0751, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 4.206006050109863, | |
| "learning_rate": 9.701492537313434e-05, | |
| "loss": 0.8775, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 2.1778831481933594, | |
| "learning_rate": 9.552238805970149e-05, | |
| "loss": 0.7614, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 2.380641222000122, | |
| "learning_rate": 9.402985074626867e-05, | |
| "loss": 0.701, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 3.6440467834472656, | |
| "learning_rate": 9.253731343283582e-05, | |
| "loss": 0.7766, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "eval_accuracy": 0.8224956063268892, | |
| "eval_loss": 0.6780304312705994, | |
| "eval_runtime": 8.1524, | |
| "eval_samples_per_second": 69.796, | |
| "eval_steps_per_second": 8.832, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 3.5913617610931396, | |
| "learning_rate": 9.104477611940299e-05, | |
| "loss": 0.8058, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "grad_norm": 3.761014938354492, | |
| "learning_rate": 8.955223880597016e-05, | |
| "loss": 0.7853, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "grad_norm": 2.7116057872772217, | |
| "learning_rate": 8.813432835820896e-05, | |
| "loss": 0.6584, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "grad_norm": 4.109683990478516, | |
| "learning_rate": 8.664179104477612e-05, | |
| "loss": 0.6056, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "grad_norm": 3.7420246601104736, | |
| "learning_rate": 8.514925373134329e-05, | |
| "loss": 0.61, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "eval_accuracy": 0.8242530755711776, | |
| "eval_loss": 0.6279409527778625, | |
| "eval_runtime": 7.7566, | |
| "eval_samples_per_second": 73.357, | |
| "eval_steps_per_second": 9.282, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "grad_norm": 4.6456379890441895, | |
| "learning_rate": 8.365671641791046e-05, | |
| "loss": 0.6016, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "grad_norm": 3.174193859100342, | |
| "learning_rate": 8.216417910447761e-05, | |
| "loss": 0.5734, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "grad_norm": 4.6657304763793945, | |
| "learning_rate": 8.067164179104479e-05, | |
| "loss": 0.5684, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "grad_norm": 4.963318347930908, | |
| "learning_rate": 7.917910447761194e-05, | |
| "loss": 0.5957, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "grad_norm": 5.692253589630127, | |
| "learning_rate": 7.776119402985074e-05, | |
| "loss": 0.4734, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.24, | |
| "eval_accuracy": 0.827768014059754, | |
| "eval_loss": 0.5593364834785461, | |
| "eval_runtime": 7.7448, | |
| "eval_samples_per_second": 73.468, | |
| "eval_steps_per_second": 9.297, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "grad_norm": 3.6985843181610107, | |
| "learning_rate": 7.626865671641792e-05, | |
| "loss": 0.4646, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.54, | |
| "grad_norm": 3.7862443923950195, | |
| "learning_rate": 7.477611940298508e-05, | |
| "loss": 0.4568, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "grad_norm": 4.695005893707275, | |
| "learning_rate": 7.328358208955224e-05, | |
| "loss": 0.4876, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.84, | |
| "grad_norm": 3.9707813262939453, | |
| "learning_rate": 7.179104477611941e-05, | |
| "loss": 0.4723, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "grad_norm": 3.133072853088379, | |
| "learning_rate": 7.029850746268657e-05, | |
| "loss": 0.5275, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "eval_accuracy": 0.8418277680140598, | |
| "eval_loss": 0.5148488879203796, | |
| "eval_runtime": 7.8181, | |
| "eval_samples_per_second": 72.78, | |
| "eval_steps_per_second": 9.209, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.13, | |
| "grad_norm": 3.1703858375549316, | |
| "learning_rate": 6.880597014925374e-05, | |
| "loss": 0.353, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 3.28, | |
| "grad_norm": 3.551959276199341, | |
| "learning_rate": 6.73134328358209e-05, | |
| "loss": 0.3559, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "grad_norm": 3.8375322818756104, | |
| "learning_rate": 6.582089552238806e-05, | |
| "loss": 0.376, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.58, | |
| "grad_norm": 4.613718032836914, | |
| "learning_rate": 6.432835820895523e-05, | |
| "loss": 0.4183, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 3.73, | |
| "grad_norm": 9.122322082519531, | |
| "learning_rate": 6.283582089552239e-05, | |
| "loss": 0.3767, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.73, | |
| "eval_accuracy": 0.843585237258348, | |
| "eval_loss": 0.5129419565200806, | |
| "eval_runtime": 7.7011, | |
| "eval_samples_per_second": 73.886, | |
| "eval_steps_per_second": 9.349, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.88, | |
| "grad_norm": 1.9917536973953247, | |
| "learning_rate": 6.134328358208955e-05, | |
| "loss": 0.3943, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 4.03, | |
| "grad_norm": 3.007828712463379, | |
| "learning_rate": 5.985074626865672e-05, | |
| "loss": 0.3885, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 4.18, | |
| "grad_norm": 6.075244426727295, | |
| "learning_rate": 5.835820895522388e-05, | |
| "loss": 0.3312, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "grad_norm": 4.8264641761779785, | |
| "learning_rate": 5.686567164179105e-05, | |
| "loss": 0.3408, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 4.48, | |
| "grad_norm": 2.8609578609466553, | |
| "learning_rate": 5.537313432835821e-05, | |
| "loss": 0.3207, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 4.48, | |
| "eval_accuracy": 0.8558875219683656, | |
| "eval_loss": 0.4966126084327698, | |
| "eval_runtime": 7.6408, | |
| "eval_samples_per_second": 74.468, | |
| "eval_steps_per_second": 9.423, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 4.63, | |
| "grad_norm": 5.234914302825928, | |
| "learning_rate": 5.388059701492537e-05, | |
| "loss": 0.3306, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 4.78, | |
| "grad_norm": 4.566553592681885, | |
| "learning_rate": 5.238805970149254e-05, | |
| "loss": 0.3532, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 4.93, | |
| "grad_norm": 4.077399253845215, | |
| "learning_rate": 5.08955223880597e-05, | |
| "loss": 0.3501, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 5.07, | |
| "grad_norm": 3.5527923107147217, | |
| "learning_rate": 4.940298507462687e-05, | |
| "loss": 0.3147, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 5.22, | |
| "grad_norm": 4.944146633148193, | |
| "learning_rate": 4.7910447761194035e-05, | |
| "loss": 0.3155, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 5.22, | |
| "eval_accuracy": 0.8453427065026362, | |
| "eval_loss": 0.5251042246818542, | |
| "eval_runtime": 7.8327, | |
| "eval_samples_per_second": 72.644, | |
| "eval_steps_per_second": 9.192, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 5.37, | |
| "grad_norm": 2.9990365505218506, | |
| "learning_rate": 4.6417910447761195e-05, | |
| "loss": 0.3121, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 5.52, | |
| "grad_norm": 1.7021130323410034, | |
| "learning_rate": 4.492537313432836e-05, | |
| "loss": 0.3563, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 5.67, | |
| "grad_norm": 4.41218376159668, | |
| "learning_rate": 4.343283582089552e-05, | |
| "loss": 0.3447, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 5.82, | |
| "grad_norm": 2.955658197402954, | |
| "learning_rate": 4.194029850746269e-05, | |
| "loss": 0.2839, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 5.97, | |
| "grad_norm": 3.0852389335632324, | |
| "learning_rate": 4.044776119402985e-05, | |
| "loss": 0.2565, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 5.97, | |
| "eval_accuracy": 0.8629173989455184, | |
| "eval_loss": 0.47900134325027466, | |
| "eval_runtime": 7.78, | |
| "eval_samples_per_second": 73.136, | |
| "eval_steps_per_second": 9.255, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 6.12, | |
| "grad_norm": 2.099174976348877, | |
| "learning_rate": 3.895522388059702e-05, | |
| "loss": 0.2818, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 6.27, | |
| "grad_norm": 3.712127685546875, | |
| "learning_rate": 3.746268656716418e-05, | |
| "loss": 0.2444, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 6.42, | |
| "grad_norm": 2.1818361282348633, | |
| "learning_rate": 3.5970149253731346e-05, | |
| "loss": 0.2418, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 6.57, | |
| "grad_norm": 2.211638927459717, | |
| "learning_rate": 3.447761194029851e-05, | |
| "loss": 0.2684, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 6.72, | |
| "grad_norm": 6.349803924560547, | |
| "learning_rate": 3.298507462686568e-05, | |
| "loss": 0.2791, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 6.72, | |
| "eval_accuracy": 0.8523725834797891, | |
| "eval_loss": 0.5110830664634705, | |
| "eval_runtime": 7.7612, | |
| "eval_samples_per_second": 73.313, | |
| "eval_steps_per_second": 9.277, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 6.87, | |
| "grad_norm": 4.6033759117126465, | |
| "learning_rate": 3.149253731343284e-05, | |
| "loss": 0.2444, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 7.01, | |
| "grad_norm": 6.079771995544434, | |
| "learning_rate": 3e-05, | |
| "loss": 0.2812, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 7.16, | |
| "grad_norm": 3.743011474609375, | |
| "learning_rate": 2.8507462686567167e-05, | |
| "loss": 0.183, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 7.31, | |
| "grad_norm": 4.840090751647949, | |
| "learning_rate": 2.701492537313433e-05, | |
| "loss": 0.2689, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 7.46, | |
| "grad_norm": 3.213412284851074, | |
| "learning_rate": 2.5522388059701496e-05, | |
| "loss": 0.1987, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 7.46, | |
| "eval_accuracy": 0.8453427065026362, | |
| "eval_loss": 0.5002422332763672, | |
| "eval_runtime": 7.6836, | |
| "eval_samples_per_second": 74.054, | |
| "eval_steps_per_second": 9.371, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 7.61, | |
| "grad_norm": 2.2559454441070557, | |
| "learning_rate": 2.402985074626866e-05, | |
| "loss": 0.2254, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 7.76, | |
| "grad_norm": 4.895073413848877, | |
| "learning_rate": 2.2537313432835822e-05, | |
| "loss": 0.283, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 7.91, | |
| "grad_norm": 3.8669703006744385, | |
| "learning_rate": 2.1044776119402985e-05, | |
| "loss": 0.281, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 8.06, | |
| "grad_norm": 2.0825252532958984, | |
| "learning_rate": 1.9552238805970148e-05, | |
| "loss": 0.1955, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 8.21, | |
| "grad_norm": 1.101592779159546, | |
| "learning_rate": 1.8059701492537314e-05, | |
| "loss": 0.2083, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 8.21, | |
| "eval_accuracy": 0.8629173989455184, | |
| "eval_loss": 0.5034471154212952, | |
| "eval_runtime": 7.7486, | |
| "eval_samples_per_second": 73.432, | |
| "eval_steps_per_second": 9.292, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 8.36, | |
| "grad_norm": 2.6236236095428467, | |
| "learning_rate": 1.6567164179104477e-05, | |
| "loss": 0.1409, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 8.51, | |
| "grad_norm": 1.433937668800354, | |
| "learning_rate": 1.5074626865671642e-05, | |
| "loss": 0.2434, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 8.66, | |
| "grad_norm": 2.8774006366729736, | |
| "learning_rate": 1.3582089552238805e-05, | |
| "loss": 0.2044, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 8.81, | |
| "grad_norm": 4.404654026031494, | |
| "learning_rate": 1.2089552238805971e-05, | |
| "loss": 0.2153, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 8.96, | |
| "grad_norm": 4.886945724487305, | |
| "learning_rate": 1.0597014925373134e-05, | |
| "loss": 0.2567, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 8.96, | |
| "eval_accuracy": 0.8576449912126538, | |
| "eval_loss": 0.4995073080062866, | |
| "eval_runtime": 7.7157, | |
| "eval_samples_per_second": 73.745, | |
| "eval_steps_per_second": 9.332, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 9.1, | |
| "grad_norm": 0.4216682016849518, | |
| "learning_rate": 9.104477611940299e-06, | |
| "loss": 0.203, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 9.25, | |
| "grad_norm": 4.639057159423828, | |
| "learning_rate": 7.611940298507463e-06, | |
| "loss": 0.1934, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 9.4, | |
| "grad_norm": 4.426870822906494, | |
| "learning_rate": 6.119402985074627e-06, | |
| "loss": 0.2067, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 9.55, | |
| "grad_norm": 2.948902130126953, | |
| "learning_rate": 4.626865671641791e-06, | |
| "loss": 0.2065, | |
| "step": 1280 | |
| }, | |
| { | |
| "epoch": 9.7, | |
| "grad_norm": 2.3631768226623535, | |
| "learning_rate": 3.134328358208955e-06, | |
| "loss": 0.2127, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 9.7, | |
| "eval_accuracy": 0.8488576449912126, | |
| "eval_loss": 0.5034462809562683, | |
| "eval_runtime": 7.6386, | |
| "eval_samples_per_second": 74.49, | |
| "eval_steps_per_second": 9.426, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 9.85, | |
| "grad_norm": 2.909392833709717, | |
| "learning_rate": 1.6417910447761194e-06, | |
| "loss": 0.1547, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 2.5194036960601807, | |
| "learning_rate": 1.4925373134328358e-07, | |
| "loss": 0.2161, | |
| "step": 1340 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 1340, | |
| "total_flos": 3.3230947683690086e+18, | |
| "train_loss": 0.3864157530798841, | |
| "train_runtime": 1135.6162, | |
| "train_samples_per_second": 37.759, | |
| "train_steps_per_second": 1.18 | |
| } | |
| ], | |
| "logging_steps": 20, | |
| "max_steps": 1340, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 100, | |
| "total_flos": 3.3230947683690086e+18, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |