zephyr-7b-dpo-lora / train_results.json
TanJing's picture
Model save
df67eb4
raw
history blame contribute delete
195 Bytes
{
"epoch": 3.0,
"train_loss": 0.5642068754707158,
"train_runtime": 89225.6094,
"train_samples": 61966,
"train_samples_per_second": 2.083,
"train_steps_per_second": 0.033
}