mistral-7B-v0.3-Base-CPT_SFT_DPO / train_results.json
mjbuehler's picture
Model save
492faec verified
raw
history blame contribute delete
218 Bytes
{
"epoch": 1.0,
"total_flos": 0.0,
"train_loss": 0.22562925693344074,
"train_runtime": 6258.663,
"train_samples": 70516,
"train_samples_per_second": 11.267,
"train_steps_per_second": 0.176
}