ds3-img-classification / trainer_state.json
jnmrr's picture
Upload model with metrics - 2024-12-12 15:17
d94ca4a verified
{
"best_metric": 0.8528685149601977,
"best_model_checkpoint": "ds3-img-classification/checkpoint-284",
"epoch": 26.41860465116279,
"eval_steps": 142,
"global_step": 1704,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.015503875968992248,
"grad_norm": 14.292332649230957,
"learning_rate": 1.736111111111111e-06,
"loss": 5.8979,
"step": 1
},
{
"epoch": 2.201550387596899,
"grad_norm": 3.2709274291992188,
"learning_rate": 0.0002465277777777778,
"loss": 2.2063,
"step": 142
},
{
"epoch": 2.201550387596899,
"eval_accuracy": 0.8300850947021685,
"eval_loss": 0.6956288814544678,
"eval_macro_f1": 0.8020115300522307,
"eval_macro_precision": 0.8326277180719005,
"eval_macro_recall": 0.8130180643302005,
"eval_micro_f1": 0.8300850947021685,
"eval_micro_precision": 0.8300850947021685,
"eval_micro_recall": 0.8300850947021685,
"eval_runtime": 8.1254,
"eval_samples_per_second": 448.347,
"eval_steps_per_second": 3.569,
"eval_weighted_f1": 0.8217784359221563,
"eval_weighted_precision": 0.8439879212180366,
"eval_weighted_recall": 0.8300850947021685,
"step": 142
},
{
"epoch": 4.403100775193798,
"grad_norm": 3.7781729698181152,
"learning_rate": 0.0004930555555555556,
"loss": 0.4906,
"step": 284
},
{
"epoch": 4.403100775193798,
"eval_accuracy": 0.8528685149601977,
"eval_loss": 0.5704612731933594,
"eval_macro_f1": 0.8447646928874578,
"eval_macro_precision": 0.8434580548053667,
"eval_macro_recall": 0.8590562294756982,
"eval_micro_f1": 0.8528685149601977,
"eval_micro_precision": 0.8528685149601977,
"eval_micro_recall": 0.8528685149601977,
"eval_runtime": 8.3026,
"eval_samples_per_second": 438.779,
"eval_steps_per_second": 3.493,
"eval_weighted_f1": 0.8542422373852568,
"eval_weighted_precision": 0.8661150819486806,
"eval_weighted_recall": 0.8528685149601977,
"step": 284
},
{
"epoch": 6.604651162790698,
"grad_norm": 2.7355144023895264,
"learning_rate": 0.0007395833333333334,
"loss": 0.4125,
"step": 426
},
{
"epoch": 6.604651162790698,
"eval_accuracy": 0.7400494098270656,
"eval_loss": 1.1121388673782349,
"eval_macro_f1": 0.7104399369512154,
"eval_macro_precision": 0.7984066441843661,
"eval_macro_recall": 0.7246907362358344,
"eval_micro_f1": 0.7400494098270657,
"eval_micro_precision": 0.7400494098270656,
"eval_micro_recall": 0.7400494098270656,
"eval_runtime": 8.3145,
"eval_samples_per_second": 438.152,
"eval_steps_per_second": 3.488,
"eval_weighted_f1": 0.7343329270391191,
"eval_weighted_precision": 0.8109139038669717,
"eval_weighted_recall": 0.7400494098270656,
"step": 426
},
{
"epoch": 8.806201550387597,
"grad_norm": 2.7910587787628174,
"learning_rate": 0.0009861111111111112,
"loss": 0.3948,
"step": 568
},
{
"epoch": 8.806201550387597,
"eval_accuracy": 0.769146307987922,
"eval_loss": 0.9053287506103516,
"eval_macro_f1": 0.7380470144030622,
"eval_macro_precision": 0.7929019396341611,
"eval_macro_recall": 0.7416184656713186,
"eval_micro_f1": 0.769146307987922,
"eval_micro_precision": 0.769146307987922,
"eval_micro_recall": 0.769146307987922,
"eval_runtime": 8.4314,
"eval_samples_per_second": 432.076,
"eval_steps_per_second": 3.44,
"eval_weighted_f1": 0.7647284160437106,
"eval_weighted_precision": 0.804728448308173,
"eval_weighted_recall": 0.769146307987922,
"step": 568
},
{
"epoch": 11.007751937984496,
"grad_norm": 6.627900123596191,
"learning_rate": 0.0009983522903302263,
"loss": 0.3706,
"step": 710
},
{
"epoch": 11.007751937984496,
"eval_accuracy": 0.7672248147131485,
"eval_loss": 0.9641202688217163,
"eval_macro_f1": 0.7528371715147637,
"eval_macro_precision": 0.787293122179915,
"eval_macro_recall": 0.7589348598973276,
"eval_micro_f1": 0.7672248147131485,
"eval_micro_precision": 0.7672248147131485,
"eval_micro_recall": 0.7672248147131485,
"eval_runtime": 8.3804,
"eval_samples_per_second": 434.705,
"eval_steps_per_second": 3.46,
"eval_weighted_f1": 0.7623056450515923,
"eval_weighted_precision": 0.7946085579995245,
"eval_weighted_recall": 0.7672248147131485,
"step": 710
},
{
"epoch": 13.209302325581396,
"grad_norm": 3.9615182876586914,
"learning_rate": 0.0009930222532081596,
"loss": 0.3302,
"step": 852
},
{
"epoch": 13.209302325581396,
"eval_accuracy": 0.7669503156738952,
"eval_loss": 0.9978815317153931,
"eval_macro_f1": 0.731377351970227,
"eval_macro_precision": 0.7752862718850362,
"eval_macro_recall": 0.7408605973033737,
"eval_micro_f1": 0.766950315673895,
"eval_micro_precision": 0.7669503156738952,
"eval_micro_recall": 0.7669503156738952,
"eval_runtime": 8.3024,
"eval_samples_per_second": 438.791,
"eval_steps_per_second": 3.493,
"eval_weighted_f1": 0.7591614272562255,
"eval_weighted_precision": 0.7900342900911238,
"eval_weighted_recall": 0.7669503156738952,
"step": 852
},
{
"epoch": 15.410852713178295,
"grad_norm": 2.4717087745666504,
"learning_rate": 0.0009840434590541415,
"loss": 0.2924,
"step": 994
},
{
"epoch": 15.410852713178295,
"eval_accuracy": 0.7233049684326105,
"eval_loss": 1.1638933420181274,
"eval_macro_f1": 0.7113593382146706,
"eval_macro_precision": 0.759489526605669,
"eval_macro_recall": 0.7371465474721075,
"eval_micro_f1": 0.7233049684326105,
"eval_micro_precision": 0.7233049684326105,
"eval_micro_recall": 0.7233049684326105,
"eval_runtime": 8.2739,
"eval_samples_per_second": 440.3,
"eval_steps_per_second": 3.505,
"eval_weighted_f1": 0.7249947526290721,
"eval_weighted_precision": 0.7928301153657854,
"eval_weighted_recall": 0.7233049684326105,
"step": 994
},
{
"epoch": 17.612403100775193,
"grad_norm": 1.886505126953125,
"learning_rate": 0.0009714823580904379,
"loss": 0.271,
"step": 1136
},
{
"epoch": 17.612403100775193,
"eval_accuracy": 0.7631073291243481,
"eval_loss": 0.9723167419433594,
"eval_macro_f1": 0.7630758622320426,
"eval_macro_precision": 0.8404983162678952,
"eval_macro_recall": 0.7684491516518803,
"eval_micro_f1": 0.7631073291243481,
"eval_micro_precision": 0.7631073291243481,
"eval_micro_recall": 0.7631073291243481,
"eval_runtime": 8.2536,
"eval_samples_per_second": 441.381,
"eval_steps_per_second": 3.514,
"eval_weighted_f1": 0.7641468306327522,
"eval_weighted_precision": 0.8384853206385829,
"eval_weighted_recall": 0.7631073291243481,
"step": 1136
},
{
"epoch": 19.813953488372093,
"grad_norm": 1.751386284828186,
"learning_rate": 0.000955431912460588,
"loss": 0.2493,
"step": 1278
},
{
"epoch": 19.813953488372093,
"eval_accuracy": 0.7556958550645073,
"eval_loss": 0.9924145340919495,
"eval_macro_f1": 0.7186954972930405,
"eval_macro_precision": 0.7780354194554568,
"eval_macro_recall": 0.7246815062488177,
"eval_micro_f1": 0.7556958550645072,
"eval_micro_precision": 0.7556958550645073,
"eval_micro_recall": 0.7556958550645073,
"eval_runtime": 8.2332,
"eval_samples_per_second": 442.478,
"eval_steps_per_second": 3.522,
"eval_weighted_f1": 0.7424053023538434,
"eval_weighted_precision": 0.7815999680084916,
"eval_weighted_recall": 0.7556958550645073,
"step": 1278
},
{
"epoch": 22.015503875968992,
"grad_norm": 1.887330174446106,
"learning_rate": 0.0009360109082355581,
"loss": 0.2276,
"step": 1420
},
{
"epoch": 22.015503875968992,
"eval_accuracy": 0.7488333790831732,
"eval_loss": 1.0906082391738892,
"eval_macro_f1": 0.7232412414342934,
"eval_macro_precision": 0.7888326922300277,
"eval_macro_recall": 0.7295020066680354,
"eval_micro_f1": 0.7488333790831732,
"eval_micro_precision": 0.7488333790831732,
"eval_micro_recall": 0.7488333790831732,
"eval_runtime": 8.3573,
"eval_samples_per_second": 435.905,
"eval_steps_per_second": 3.47,
"eval_weighted_f1": 0.7411123038164863,
"eval_weighted_precision": 0.78924044171482,
"eval_weighted_recall": 0.7488333790831732,
"step": 1420
},
{
"epoch": 24.217054263565892,
"grad_norm": 1.4166276454925537,
"learning_rate": 0.0009133630763022861,
"loss": 0.2147,
"step": 1562
},
{
"epoch": 24.217054263565892,
"eval_accuracy": 0.7595388416140544,
"eval_loss": 1.046807050704956,
"eval_macro_f1": 0.7218058945395823,
"eval_macro_precision": 0.7806200306355751,
"eval_macro_recall": 0.7363816501236903,
"eval_micro_f1": 0.7595388416140544,
"eval_micro_precision": 0.7595388416140544,
"eval_micro_recall": 0.7595388416140544,
"eval_runtime": 8.3705,
"eval_samples_per_second": 435.218,
"eval_steps_per_second": 3.465,
"eval_weighted_f1": 0.7500356793933665,
"eval_weighted_precision": 0.7963937500158152,
"eval_weighted_recall": 0.7595388416140544,
"step": 1562
},
{
"epoch": 26.41860465116279,
"grad_norm": 1.3001500368118286,
"learning_rate": 0.0008876560286407329,
"loss": 0.1915,
"step": 1704
},
{
"epoch": 26.41860465116279,
"eval_accuracy": 0.7757342849300027,
"eval_loss": 0.9456538558006287,
"eval_macro_f1": 0.7518200808982276,
"eval_macro_precision": 0.7941242669868954,
"eval_macro_recall": 0.7603347162166142,
"eval_micro_f1": 0.7757342849300026,
"eval_micro_precision": 0.7757342849300027,
"eval_micro_recall": 0.7757342849300027,
"eval_runtime": 8.4106,
"eval_samples_per_second": 433.145,
"eval_steps_per_second": 3.448,
"eval_weighted_f1": 0.7688586665917762,
"eval_weighted_precision": 0.8022618721710044,
"eval_weighted_recall": 0.7757342849300027,
"step": 1704
},
{
"epoch": 26.41860465116279,
"step": 1704,
"total_flos": 2.3734952710727598e+20,
"train_loss": 0.4731367647368024,
"train_runtime": 2312.991,
"train_samples_per_second": 1275.725,
"train_steps_per_second": 2.49
}
],
"logging_steps": 142,
"max_steps": 5760,
"num_input_tokens_seen": 0,
"num_train_epochs": 90,
"save_steps": 142,
"stateful_callbacks": {
"EarlyStoppingCallback": {
"args": {
"early_stopping_patience": 10,
"early_stopping_threshold": 0.01
},
"attributes": {
"early_stopping_patience_counter": 10
}
},
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 2.3734952710727598e+20,
"train_batch_size": 256,
"trial_name": null,
"trial_params": null
}