Llama3.2-Mamba-3B-dpo / train_results.json
Jingyu6
add models
0da1b35
raw
history blame
219 Bytes
{
"epoch": 1.0,
"total_flos": 0.0,
"train_loss": 0.5327433187535995,
"train_runtime": 4228.4167,
"train_samples": 133368,
"train_samples_per_second": 31.541,
"train_steps_per_second": 0.986
}