cmv_hard_gemma3-12b-it_full_sft / trainer_state.json
simonycl's picture
Upload folder using huggingface_hub
897c1f4 verified
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.9984162179284131,
"eval_steps": 500,
"global_step": 394,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.025340513145391194,
"grad_norm": 7.014422278343604,
"learning_rate": 4.5e-06,
"loss": 1.1276,
"step": 10
},
{
"epoch": 0.05068102629078239,
"grad_norm": 3.950172585471577,
"learning_rate": 9.5e-06,
"loss": 0.9532,
"step": 20
},
{
"epoch": 0.07602153943617358,
"grad_norm": 2.8852145535313203,
"learning_rate": 9.985718470743916e-06,
"loss": 0.9247,
"step": 30
},
{
"epoch": 0.10136205258156478,
"grad_norm": 4.372398694120704,
"learning_rate": 9.936454954953108e-06,
"loss": 0.9131,
"step": 40
},
{
"epoch": 0.12670256572695598,
"grad_norm": 2.678600773280688,
"learning_rate": 9.852380451890723e-06,
"loss": 0.9038,
"step": 50
},
{
"epoch": 0.15204307887234716,
"grad_norm": 3.1616554933826584,
"learning_rate": 9.734087839742152e-06,
"loss": 0.9032,
"step": 60
},
{
"epoch": 0.17738359201773837,
"grad_norm": 2.711129850639537,
"learning_rate": 9.58241129660755e-06,
"loss": 0.8988,
"step": 70
},
{
"epoch": 0.20272410516312955,
"grad_norm": 2.816882812285983,
"learning_rate": 9.398420418028789e-06,
"loss": 0.9,
"step": 80
},
{
"epoch": 0.22806461830852076,
"grad_norm": 2.4491670100971437,
"learning_rate": 9.183412674395193e-06,
"loss": 0.889,
"step": 90
},
{
"epoch": 0.25340513145391197,
"grad_norm": 2.3671858961630132,
"learning_rate": 8.938904261417088e-06,
"loss": 0.8862,
"step": 100
},
{
"epoch": 0.2787456445993031,
"grad_norm": 25.362434554086317,
"learning_rate": 8.666619408187953e-06,
"loss": 0.8874,
"step": 110
},
{
"epoch": 0.30408615774469433,
"grad_norm": 2.435185075586954,
"learning_rate": 8.368478218232787e-06,
"loss": 0.8809,
"step": 120
},
{
"epoch": 0.32942667089008554,
"grad_norm": 2.3423241237120522,
"learning_rate": 8.046583129285422e-06,
"loss": 0.8759,
"step": 130
},
{
"epoch": 0.35476718403547675,
"grad_norm": 2.295983766372591,
"learning_rate": 7.703204087277989e-06,
"loss": 0.8684,
"step": 140
},
{
"epoch": 0.3801076971808679,
"grad_norm": 2.8544179633447255,
"learning_rate": 7.340762539092858e-06,
"loss": 0.8676,
"step": 150
},
{
"epoch": 0.4054482103262591,
"grad_norm": 2.2374422127818963,
"learning_rate": 6.961814356957308e-06,
"loss": 0.8718,
"step": 160
},
{
"epoch": 0.4307887234716503,
"grad_norm": 2.1472864212507976,
"learning_rate": 6.569031814894962e-06,
"loss": 0.8613,
"step": 170
},
{
"epoch": 0.4561292366170415,
"grad_norm": 2.2467067956901183,
"learning_rate": 6.165184744332824e-06,
"loss": 0.8587,
"step": 180
},
{
"epoch": 0.4814697497624327,
"grad_norm": 2.2528338735220905,
"learning_rate": 5.753121001751161e-06,
"loss": 0.8496,
"step": 190
},
{
"epoch": 0.5068102629078239,
"grad_norm": 2.4179129846566636,
"learning_rate": 5.335746386114814e-06,
"loss": 0.8507,
"step": 200
},
{
"epoch": 0.532150776053215,
"grad_norm": 16.529801921641155,
"learning_rate": 4.9160041477046e-06,
"loss": 0.8465,
"step": 210
},
{
"epoch": 0.5574912891986062,
"grad_norm": 2.1966500837147556,
"learning_rate": 4.4968542328488e-06,
"loss": 0.8543,
"step": 220
},
{
"epoch": 0.5828318023439975,
"grad_norm": 2.2370934180241986,
"learning_rate": 4.081252410917148e-06,
"loss": 0.841,
"step": 230
},
{
"epoch": 0.6081723154893887,
"grad_norm": 2.086887732081978,
"learning_rate": 3.6721294307699786e-06,
"loss": 0.8358,
"step": 240
},
{
"epoch": 0.6335128286347799,
"grad_norm": 2.263852568221778,
"learning_rate": 3.272370353647465e-06,
"loss": 0.8242,
"step": 250
},
{
"epoch": 0.6588533417801711,
"grad_norm": 2.24355503619341,
"learning_rate": 2.8847942082397112e-06,
"loss": 0.83,
"step": 260
},
{
"epoch": 0.6841938549255623,
"grad_norm": 2.2286503006725806,
"learning_rate": 2.512134111406422e-06,
"loss": 0.8256,
"step": 270
},
{
"epoch": 0.7095343680709535,
"grad_norm": 2.1171293835945155,
"learning_rate": 2.1570179947312674e-06,
"loss": 0.8219,
"step": 280
},
{
"epoch": 0.7348748812163446,
"grad_norm": 2.087191484034938,
"learning_rate": 1.8219500728237849e-06,
"loss": 0.8211,
"step": 290
},
{
"epoch": 0.7602153943617358,
"grad_norm": 2.3605396760707866,
"learning_rate": 1.509293184050995e-06,
"loss": 0.8124,
"step": 300
},
{
"epoch": 0.785555907507127,
"grad_norm": 2.1652486052680384,
"learning_rate": 1.2212521282287093e-06,
"loss": 0.8189,
"step": 310
},
{
"epoch": 0.8108964206525182,
"grad_norm": 2.04980952594426,
"learning_rate": 9.59858118772105e-07,
"loss": 0.8151,
"step": 320
},
{
"epoch": 0.8362369337979094,
"grad_norm": 2.0755635014559606,
"learning_rate": 7.269544589461968e-07,
"loss": 0.8189,
"step": 330
},
{
"epoch": 0.8615774469433006,
"grad_norm": 2.0129993073524095,
"learning_rate": 5.241835432246888e-07,
"loss": 0.8107,
"step": 340
},
{
"epoch": 0.8869179600886918,
"grad_norm": 2.0409156226901315,
"learning_rate": 3.5297527542127675e-07,
"loss": 0.8062,
"step": 350
},
{
"epoch": 0.912258473234083,
"grad_norm": 2.3050144167741613,
"learning_rate": 2.1453698526664513e-07,
"loss": 0.8155,
"step": 360
},
{
"epoch": 0.9375989863794741,
"grad_norm": 2.0906040167387188,
"learning_rate": 1.0984491453762402e-07,
"loss": 0.8078,
"step": 370
},
{
"epoch": 0.9629394995248653,
"grad_norm": 1.94641588069489,
"learning_rate": 3.963733277679904e-08,
"loss": 0.802,
"step": 380
},
{
"epoch": 0.9882800126702566,
"grad_norm": 2.333717554003802,
"learning_rate": 4.409331149256013e-09,
"loss": 0.8076,
"step": 390
},
{
"epoch": 0.9984162179284131,
"step": 394,
"total_flos": 318229560426496.0,
"train_loss": 0.8606352945269667,
"train_runtime": 15262.586,
"train_samples_per_second": 1.655,
"train_steps_per_second": 0.026
}
],
"logging_steps": 10,
"max_steps": 394,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 318229560426496.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}