Gemma-2B-finetuned-SQL / checkpoint-50 /trainer_state.json
Xonotic's picture
Subiendo modelo Gemma-2B fine-tuned con LoRA
31a364e verified
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 2.2758620689655173,
"eval_steps": 25,
"global_step": 50,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.22988505747126436,
"grad_norm": 1.4753433465957642,
"learning_rate": 0.00011428571428571428,
"loss": 4.5501,
"step": 5
},
{
"epoch": 0.45977011494252873,
"grad_norm": 3.072629451751709,
"learning_rate": 0.00019322033898305085,
"loss": 3.6174,
"step": 10
},
{
"epoch": 0.6896551724137931,
"grad_norm": 4.133386611938477,
"learning_rate": 0.00017627118644067798,
"loss": 2.4791,
"step": 15
},
{
"epoch": 0.9195402298850575,
"grad_norm": 2.3766379356384277,
"learning_rate": 0.00015932203389830508,
"loss": 1.7968,
"step": 20
},
{
"epoch": 1.1379310344827587,
"grad_norm": 3.2854702472686768,
"learning_rate": 0.0001423728813559322,
"loss": 1.7796,
"step": 25
},
{
"epoch": 1.367816091954023,
"grad_norm": 2.0210888385772705,
"learning_rate": 0.00012542372881355933,
"loss": 1.4906,
"step": 30
},
{
"epoch": 1.5977011494252875,
"grad_norm": 2.1205132007598877,
"learning_rate": 0.00010847457627118644,
"loss": 1.4216,
"step": 35
},
{
"epoch": 1.8275862068965516,
"grad_norm": 1.489486575126648,
"learning_rate": 9.152542372881357e-05,
"loss": 1.1526,
"step": 40
},
{
"epoch": 2.045977011494253,
"grad_norm": 1.803501844406128,
"learning_rate": 7.457627118644068e-05,
"loss": 1.3507,
"step": 45
},
{
"epoch": 2.2758620689655173,
"grad_norm": 1.8296611309051514,
"learning_rate": 5.76271186440678e-05,
"loss": 1.1522,
"step": 50
}
],
"logging_steps": 5,
"max_steps": 66,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 50,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 302864384655360.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}