zephyr-7b-dpo-lora / train_results.json
jikaixuan's picture
Training in progress, epoch 0
64a18a9
raw
history blame contribute delete
194 Bytes
{
"epoch": 3.0,
"train_loss": -1.932115889119454,
"train_runtime": 45081.596,
"train_samples": 61966,
"train_samples_per_second": 4.124,
"train_steps_per_second": 0.064
}