qlora-llama2-13b / checkpoint-2000 /trainer_state.json
ZeroUniqueness's picture
Training in progress, step 3000
fa9838e
raw
history blame
3.23 kB
{
"best_metric": 0.7452704310417175,
"best_model_checkpoint": "./qlora-out/checkpoint-2000",
"epoch": 0.07456843518138771,
"global_step": 2000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 0.00019999938245325715,
"loss": 0.9023,
"step": 100
},
{
"epoch": 0.01,
"learning_rate": 0.00019999724773356797,
"loss": 0.8027,
"step": 200
},
{
"epoch": 0.01,
"learning_rate": 0.0001999935882494411,
"loss": 0.8041,
"step": 300
},
{
"epoch": 0.01,
"learning_rate": 0.00019998840405667672,
"loss": 0.7944,
"step": 400
},
{
"epoch": 0.02,
"learning_rate": 0.00019998169523432365,
"loss": 0.81,
"step": 500
},
{
"epoch": 0.02,
"learning_rate": 0.0001999734618846785,
"loss": 0.7855,
"step": 600
},
{
"epoch": 0.03,
"learning_rate": 0.00019996370413328385,
"loss": 0.7849,
"step": 700
},
{
"epoch": 0.03,
"learning_rate": 0.00019995242212892653,
"loss": 0.7564,
"step": 800
},
{
"epoch": 0.03,
"learning_rate": 0.00019993961604363532,
"loss": 0.7724,
"step": 900
},
{
"epoch": 0.04,
"learning_rate": 0.00019992528607267815,
"loss": 0.7308,
"step": 1000
},
{
"epoch": 0.04,
"eval_loss": 0.7677998542785645,
"eval_runtime": 1774.3517,
"eval_samples_per_second": 0.305,
"eval_steps_per_second": 0.305,
"step": 1000
},
{
"epoch": 0.04,
"learning_rate": 0.0001999094324345594,
"loss": 0.7844,
"step": 1100
},
{
"epoch": 0.04,
"learning_rate": 0.00019989205537101633,
"loss": 0.7668,
"step": 1200
},
{
"epoch": 0.05,
"learning_rate": 0.00019987315514701553,
"loss": 0.7727,
"step": 1300
},
{
"epoch": 0.05,
"learning_rate": 0.00019985273205074878,
"loss": 0.7467,
"step": 1400
},
{
"epoch": 0.06,
"learning_rate": 0.00019983078639362883,
"loss": 0.7516,
"step": 1500
},
{
"epoch": 0.06,
"learning_rate": 0.00019980731851028445,
"loss": 0.7267,
"step": 1600
},
{
"epoch": 0.06,
"learning_rate": 0.0001997823287585554,
"loss": 0.7632,
"step": 1700
},
{
"epoch": 0.07,
"learning_rate": 0.000199755817519487,
"loss": 0.7392,
"step": 1800
},
{
"epoch": 0.07,
"learning_rate": 0.00019972778519732436,
"loss": 0.7528,
"step": 1900
},
{
"epoch": 0.07,
"learning_rate": 0.0001996982322195061,
"loss": 0.725,
"step": 2000
},
{
"epoch": 0.07,
"eval_loss": 0.7452704310417175,
"eval_runtime": 1787.7554,
"eval_samples_per_second": 0.303,
"eval_steps_per_second": 0.303,
"step": 2000
}
],
"max_steps": 80463,
"num_train_epochs": 3,
"total_flos": 5.618921541299405e+17,
"trial_name": null,
"trial_params": null
}