special12b / trainer_state.json
tsor13's picture
Initial upload of fine‑tuned Gemma + custom tokenizer
9cbdcd0 verified
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.5877457310948486,
"eval_steps": 32,
"global_step": 8,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.07346821638685608,
"grad_norm": 52224.03643740955,
"learning_rate": 3e-06,
"loss": 1232.5903,
"mean_token_accuracy": 0.5698598268500064,
"num_tokens": 2439700.0,
"step": 1
},
{
"epoch": 0.14693643277371216,
"grad_norm": 24564.28249818258,
"learning_rate": 2.7692307692307693e-06,
"loss": 1012.342,
"mean_token_accuracy": 0.5935460063046776,
"num_tokens": 4818537.0,
"step": 2
},
{
"epoch": 0.22040464916056823,
"grad_norm": 10030.685256076777,
"learning_rate": 2.5384615384615385e-06,
"loss": 771.7506,
"mean_token_accuracy": 0.6206552012590691,
"num_tokens": 7222999.0,
"step": 3
},
{
"epoch": 0.2938728655474243,
"grad_norm": 9797.578497343955,
"learning_rate": 2.307692307692308e-06,
"loss": 673.8949,
"mean_token_accuracy": 0.6229124862584285,
"num_tokens": 9603627.0,
"step": 4
},
{
"epoch": 0.3673410819342804,
"grad_norm": 6906.536059477202,
"learning_rate": 2.076923076923077e-06,
"loss": 647.7118,
"mean_token_accuracy": 0.6252844646223821,
"num_tokens": 12031078.0,
"step": 5
},
{
"epoch": 0.44080929832113647,
"grad_norm": 4121.501080449312,
"learning_rate": 1.8461538461538462e-06,
"loss": 616.8495,
"mean_token_accuracy": 0.6341593922115862,
"num_tokens": 14403431.0,
"step": 6
},
{
"epoch": 0.5142775147079925,
"grad_norm": 2026.657249791317,
"learning_rate": 1.6153846153846154e-06,
"loss": 601.5251,
"mean_token_accuracy": 0.6352307246997952,
"num_tokens": 16811979.0,
"step": 7
},
{
"epoch": 0.5877457310948486,
"grad_norm": 1892.3110141436696,
"learning_rate": 1.3846153846153846e-06,
"loss": 577.3281,
"mean_token_accuracy": 0.6342741788248532,
"num_tokens": 19217660.0,
"step": 8
}
],
"logging_steps": 1,
"max_steps": 13,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 8,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 128626689310720.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}