Meta-Llama-3-8B-Instruct-MI-1e-6 / train_results.json
tengxiao1
TX
1f70f5a
{
"epoch": 0.9982631930527722,
"total_flos": 0.0,
"train_loss": 1.2022870587587866,
"train_runtime": 20947.8363,
"train_samples": 59876,
"train_samples_per_second": 2.858,
"train_steps_per_second": 0.022
}