phi-2-gpo-lora-ultrafeedback-test / train_results.json
lole25's picture
Model save
b146449 verified
raw
history blame contribute delete
196 Bytes
{
"epoch": 2.0,
"train_loss": 0.00036946561885997653,
"train_runtime": 13.8182,
"train_samples": 30567,
"train_samples_per_second": 0.289,
"train_steps_per_second": 0.145
}