Llama3.1-Mamba-8B-dpo / train_results.json
Junxiong Wang
add models
5009c6e
raw
history blame contribute delete
219 Bytes
{
"epoch": 1.0,
"total_flos": 0.0,
"train_loss": 0.4679888197991303,
"train_runtime": 14179.2142,
"train_samples": 133368,
"train_samples_per_second": 9.406,
"train_steps_per_second": 0.294
}