NYXMed-V18-Model / training_summary.json
vineetdaniels's picture
Upload training_summary.json with huggingface_hub
cdedaec verified
Raw
History Blame Contribute Delete
407 Bytes
{
"base_model": "vineetdaniels/NYXMed-V17-Merged",
"output_model": "vineetdaniels/NYXMed-V18-Model",
"train_examples": 59170,
"val_examples": 500,
"epochs": 2,
"max_length": 2560,
"grad_accum": 8,
"effective_batch": 32,
"learning_rate": 1e-05,
"lora_r": 64,
"lora_alpha": 128,
"final_eval_loss": 0.07097452133893967,
"total_steps": 1700,
"train_runtime_hrs": 10.879446277777777
}