llama3_mamba_0_5_sft / eval_results.json
Junxiong Wang
add models
a871be3
{
"epoch": 1.0,
"eval_loss": 0.6645376086235046,
"eval_runtime": 15.9472,
"eval_samples": 4096,
"eval_samples_per_second": 44.835,
"eval_steps_per_second": 0.752
}