llama3-sudo-5epochs-tofu_full_sft / train_results.json
Qin Liu
Model save
57fbd7c verified
raw
history blame contribute delete
229 Bytes
{
"epoch": 5.0,
"total_flos": 2442779688960.0,
"train_loss": 0.6922085983276367,
"train_runtime": 19109.4525,
"train_samples": 4000,
"train_samples_per_second": 1.047,
"train_steps_per_second": 0.065
}