zephyr-7b-dpo-qlora / train_results.json
meixiang123's picture
Training in progress, step 1
4876cdf verified
{
"epoch": 1.0,
"total_flos": 0.0,
"train_loss": 0.6187007086617606,
"train_runtime": 1438.7543,
"train_samples": 100,
"train_samples_per_second": 0.07,
"train_steps_per_second": 0.005
}