healed1 / train_results.json
novcor's picture
Duplicate from Ffohturk/mistral-7b_ablated_model_layer_1_gate_only_healed
de3d899
Raw
History Blame Contribute Delete
196 Bytes
{
"epoch": 1.0,
"train_loss": 0.9989171136335196,
"train_runtime": 47811.9362,
"train_samples": 207864,
"train_samples_per_second": 2.916,
"train_steps_per_second": 0.182
}