Llama3.1-Mamba2-8B-dpo / train_results.json
Junxiong Wang
add models
7ae6913
raw
history blame contribute delete
219 Bytes
{
"epoch": 1.0,
"total_flos": 0.0,
"train_loss": 0.4822349088434523,
"train_runtime": 13347.0673,
"train_samples": 133368,
"train_samples_per_second": 9.992,
"train_steps_per_second": 0.312
}