356 lines
9.3 KiB
JSON
356 lines
9.3 KiB
JSON
|
{
|
||
|
|
"best_global_step": null,
|
||
|
|
"best_metric": null,
|
||
|
|
"best_model_checkpoint": null,
|
||
|
|
"epoch": 7.0,
|
||
|
|
"eval_steps": 500,
|
||
|
|
"global_step": 140,
|
||
|
|
"is_hyper_param_search": false,
|
||
|
|
"is_local_process_zero": true,
|
||
|
|
"is_world_process_zero": true,
|
||
|
|
"log_history": [
|
||
|
|
{
|
||
|
|
"epoch": 0.25,
|
||
|
|
"grad_norm": 6.958695468962432,
|
||
|
|
"learning_rate": 1.1428571428571429e-05,
|
||
|
|
"loss": 0.8727,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.47663891315460205,
|
||
|
|
"step": 5,
|
||
|
|
"valid_targets_mean": 2538.0,
|
||
|
|
"valid_targets_min": 960
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.5,
|
||
|
|
"grad_norm": 1.8207863318741868,
|
||
|
|
"learning_rate": 2.5714285714285718e-05,
|
||
|
|
"loss": 0.7493,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.3401631712913513,
|
||
|
|
"step": 10,
|
||
|
|
"valid_targets_mean": 3047.4,
|
||
|
|
"valid_targets_min": 881
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.75,
|
||
|
|
"grad_norm": 1.2285257783776453,
|
||
|
|
"learning_rate": 4e-05,
|
||
|
|
"loss": 0.6659,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.4193618893623352,
|
||
|
|
"step": 15,
|
||
|
|
"valid_targets_mean": 3820.1,
|
||
|
|
"valid_targets_min": 1000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.0,
|
||
|
|
"grad_norm": 1.081996271850223,
|
||
|
|
"learning_rate": 3.984478413200345e-05,
|
||
|
|
"loss": 0.6289,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.2763247489929199,
|
||
|
|
"step": 20,
|
||
|
|
"valid_targets_mean": 2313.0,
|
||
|
|
"valid_targets_min": 1082
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.25,
|
||
|
|
"grad_norm": 0.9507456507977273,
|
||
|
|
"learning_rate": 3.938154572458156e-05,
|
||
|
|
"loss": 0.5711,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.29192113876342773,
|
||
|
|
"step": 25,
|
||
|
|
"valid_targets_mean": 2284.5,
|
||
|
|
"valid_targets_min": 1012
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.5,
|
||
|
|
"grad_norm": 0.7306775600487055,
|
||
|
|
"learning_rate": 3.861747497288409e-05,
|
||
|
|
"loss": 0.5306,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.21971933543682098,
|
||
|
|
"step": 30,
|
||
|
|
"valid_targets_mean": 3120.9,
|
||
|
|
"valid_targets_min": 1082
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.75,
|
||
|
|
"grad_norm": 0.8229696406571964,
|
||
|
|
"learning_rate": 3.756443146740457e-05,
|
||
|
|
"loss": 0.5133,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.26939940452575684,
|
||
|
|
"step": 35,
|
||
|
|
"valid_targets_mean": 2751.1,
|
||
|
|
"valid_targets_min": 906
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"grad_norm": 1.1425727207706642,
|
||
|
|
"learning_rate": 3.623876011431714e-05,
|
||
|
|
"loss": 0.5376,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.22945210337638855,
|
||
|
|
"step": 40,
|
||
|
|
"valid_targets_mean": 2098.4,
|
||
|
|
"valid_targets_min": 888
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.25,
|
||
|
|
"grad_norm": 0.8793107537574675,
|
||
|
|
"learning_rate": 3.4661037436596526e-05,
|
||
|
|
"loss": 0.4984,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.2042250633239746,
|
||
|
|
"step": 45,
|
||
|
|
"valid_targets_mean": 1877.8,
|
||
|
|
"valid_targets_min": 281
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.5,
|
||
|
|
"grad_norm": 0.9277264133426655,
|
||
|
|
"learning_rate": 3.285575219373079e-05,
|
||
|
|
"loss": 0.4675,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.24294951558113098,
|
||
|
|
"step": 50,
|
||
|
|
"valid_targets_mean": 1935.4,
|
||
|
|
"valid_targets_min": 1180
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.75,
|
||
|
|
"grad_norm": 0.8356405001375695,
|
||
|
|
"learning_rate": 3.0850925277315193e-05,
|
||
|
|
"loss": 0.4494,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.2609573304653168,
|
||
|
|
"step": 55,
|
||
|
|
"valid_targets_mean": 3075.0,
|
||
|
|
"valid_targets_min": 1023
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.0,
|
||
|
|
"grad_norm": 0.8994016375669758,
|
||
|
|
"learning_rate": 2.8677674782351164e-05,
|
||
|
|
"loss": 0.4392,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.22629202902317047,
|
||
|
|
"step": 60,
|
||
|
|
"valid_targets_mean": 2096.5,
|
||
|
|
"valid_targets_min": 906
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.25,
|
||
|
|
"grad_norm": 0.8719408186029495,
|
||
|
|
"learning_rate": 2.6369733005033693e-05,
|
||
|
|
"loss": 0.4353,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.21587803959846497,
|
||
|
|
"step": 65,
|
||
|
|
"valid_targets_mean": 2585.2,
|
||
|
|
"valid_targets_min": 1046
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.5,
|
||
|
|
"grad_norm": 0.8754944835267683,
|
||
|
|
"learning_rate": 2.3962922863987956e-05,
|
||
|
|
"loss": 0.4026,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.19089141488075256,
|
||
|
|
"step": 70,
|
||
|
|
"valid_targets_mean": 1806.1,
|
||
|
|
"valid_targets_min": 1012
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 3.75,
|
||
|
|
"grad_norm": 1.0148994845965762,
|
||
|
|
"learning_rate": 2.149460187172849e-05,
|
||
|
|
"loss": 0.4024,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.21601556241512299,
|
||
|
|
"step": 75,
|
||
|
|
"valid_targets_mean": 2151.1,
|
||
|
|
"valid_targets_min": 1249
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.0,
|
||
|
|
"grad_norm": 0.6801895926025394,
|
||
|
|
"learning_rate": 1.9003082286786056e-05,
|
||
|
|
"loss": 0.4011,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.13709492981433868,
|
||
|
|
"step": 80,
|
||
|
|
"valid_targets_mean": 2535.5,
|
||
|
|
"valid_targets_min": 1048
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.25,
|
||
|
|
"grad_norm": 0.8588772114134782,
|
||
|
|
"learning_rate": 1.6527036446661396e-05,
|
||
|
|
"loss": 0.4039,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.21471670269966125,
|
||
|
|
"step": 85,
|
||
|
|
"valid_targets_mean": 2557.5,
|
||
|
|
"valid_targets_min": 693
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.5,
|
||
|
|
"grad_norm": 0.9032798666713535,
|
||
|
|
"learning_rate": 1.4104896511781916e-05,
|
||
|
|
"loss": 0.3779,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.20103931427001953,
|
||
|
|
"step": 90,
|
||
|
|
"valid_targets_mean": 2660.1,
|
||
|
|
"valid_targets_min": 1229
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 4.75,
|
||
|
|
"grad_norm": 0.7615772109612865,
|
||
|
|
"learning_rate": 1.1774257937387774e-05,
|
||
|
|
"loss": 0.3646,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.1381789743900299,
|
||
|
|
"step": 95,
|
||
|
|
"valid_targets_mean": 2183.5,
|
||
|
|
"valid_targets_min": 1147
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.0,
|
||
|
|
"grad_norm": 0.9699370062930872,
|
||
|
|
"learning_rate": 9.57129593241004e-06,
|
||
|
|
"loss": 0.3588,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.18376871943473816,
|
||
|
|
"step": 100,
|
||
|
|
"valid_targets_mean": 2001.8,
|
||
|
|
"valid_targets_min": 903
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.25,
|
||
|
|
"grad_norm": 0.9079217183176301,
|
||
|
|
"learning_rate": 7.530203962825331e-06,
|
||
|
|
"loss": 0.3493,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.19518150389194489,
|
||
|
|
"step": 105,
|
||
|
|
"valid_targets_mean": 3182.5,
|
||
|
|
"valid_targets_min": 281
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.5,
|
||
|
|
"grad_norm": 1.0126584456406313,
|
||
|
|
"learning_rate": 5.682663014805631e-06,
|
||
|
|
"loss": 0.3352,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.17489181458950043,
|
||
|
|
"step": 110,
|
||
|
|
"valid_targets_mean": 2201.6,
|
||
|
|
"valid_targets_min": 1242
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 5.75,
|
||
|
|
"grad_norm": 1.0687509117646672,
|
||
|
|
"learning_rate": 4.057349855541557e-06,
|
||
|
|
"loss": 0.3566,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.18527688086032867,
|
||
|
|
"step": 115,
|
||
|
|
"valid_targets_mean": 1757.4,
|
||
|
|
"valid_targets_min": 896
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 6.0,
|
||
|
|
"grad_norm": 0.8424843756035632,
|
||
|
|
"learning_rate": 2.679491924311226e-06,
|
||
|
|
"loss": 0.3757,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.17017652094364166,
|
||
|
|
"step": 120,
|
||
|
|
"valid_targets_mean": 2468.2,
|
||
|
|
"valid_targets_min": 1479
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 6.25,
|
||
|
|
"grad_norm": 0.8190515884890711,
|
||
|
|
"learning_rate": 1.5704757625918454e-06,
|
||
|
|
"loss": 0.3467,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.19655027985572815,
|
||
|
|
"step": 125,
|
||
|
|
"valid_targets_mean": 2659.4,
|
||
|
|
"valid_targets_min": 1249
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 6.5,
|
||
|
|
"grad_norm": 0.7847715606973761,
|
||
|
|
"learning_rate": 7.475150609997595e-07,
|
||
|
|
"loss": 0.3642,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.24304279685020447,
|
||
|
|
"step": 130,
|
||
|
|
"valid_targets_mean": 3288.0,
|
||
|
|
"valid_targets_min": 888
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 6.75,
|
||
|
|
"grad_norm": 0.7443132050788007,
|
||
|
|
"learning_rate": 2.2338347549742956e-07,
|
||
|
|
"loss": 0.3527,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.18210794031620026,
|
||
|
|
"step": 135,
|
||
|
|
"valid_targets_mean": 2576.2,
|
||
|
|
"valid_targets_min": 693
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 7.0,
|
||
|
|
"grad_norm": 0.8807563056194639,
|
||
|
|
"learning_rate": 6.216359983675091e-09,
|
||
|
|
"loss": 0.3217,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.16928507387638092,
|
||
|
|
"step": 140,
|
||
|
|
"valid_targets_mean": 2156.8,
|
||
|
|
"valid_targets_min": 1031
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 7.0,
|
||
|
|
"loss_nan_ranks": 0,
|
||
|
|
"loss_rank_avg": 0.16928507387638092,
|
||
|
|
"step": 140,
|
||
|
|
"total_flos": 2.2347755431133184e+16,
|
||
|
|
"train_loss": 0.4597333984715598,
|
||
|
|
"train_runtime": 702.992,
|
||
|
|
"train_samples_per_second": 3.147,
|
||
|
|
"train_steps_per_second": 0.199,
|
||
|
|
"valid_targets_mean": 2156.8,
|
||
|
|
"valid_targets_min": 1031
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"logging_steps": 5,
|
||
|
|
"max_steps": 140,
|
||
|
|
"num_input_tokens_seen": 0,
|
||
|
|
"num_train_epochs": 7,
|
||
|
|
"save_steps": 1500,
|
||
|
|
"stateful_callbacks": {
|
||
|
|
"TrainerControl": {
|
||
|
|
"args": {
|
||
|
|
"should_epoch_stop": false,
|
||
|
|
"should_evaluate": false,
|
||
|
|
"should_log": false,
|
||
|
|
"should_save": true,
|
||
|
|
"should_training_stop": true
|
||
|
|
},
|
||
|
|
"attributes": {}
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"total_flos": 2.2347755431133184e+16,
|
||
|
|
"train_batch_size": 1,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": null
|
||
|
|
}
|