zephyr-7b-dpo-qlora / train_results.json
taicheng's picture
Model save
d1f6790 verified
raw
history blame
232 Bytes
{
"epoch": 0.9997382884061764,
"total_flos": 0.0,
"train_loss": 0.2358125359600127,
"train_runtime": 19082.7985,
"train_samples": 61134,
"train_samples_per_second": 3.204,
"train_steps_per_second": 0.05
}