Upcycled-Qwen1.5-MoE2.7B-LoRA / all_results.json
gabrielmbmb's picture
gabrielmbmb HF staff
Upload folder using huggingface_hub
3999164 verified
raw
history blame contribute delete
165 Bytes
{
"epoch": 3.0,
"train_loss": 4.515984590848287,
"train_runtime": 5513.8168,
"train_samples_per_second": 0.696,
"train_steps_per_second": 0.087
}