research-distil-e92feb57 / minitron_hf_fallback.json
models4world's picture
Upload folder using huggingface_hub
b9a735f verified
Raw
History Blame Contribute Delete
259 Bytes
{
"teacher": "/home/aadi/medminitron/models/v2/teacher-ft-ckpt2000",
"target_hidden": 1800,
"target_intermediate": 4000,
"calib_samples": 512,
"calib_dataset": "MedRAG/pubmed",
"pruned_params": 3654723272,
"pruned_params_billions": 3.654723272
}