phi_3-offline-dpo-noise-0.0-42 / train_results.json
Wenboz's picture
Model save
49256fd verified
{
"epoch": 0.9230769230769231,
"total_flos": 0.0,
"train_loss": 0.6940749088923136,
"train_runtime": 53.4359,
"train_samples": 200,
"train_samples_per_second": 3.743,
"train_steps_per_second": 0.056
}