Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →nvidia/tao-finetune-huggingface-model
Fine-tune any HuggingFace CV / VLM / LLM model on local NVIDIA GPUs inside an NGC PyTorch container.
Passed 13 of 21 test cases with the skill, and 6 of 21 without it.
and 5 more in the full report
8 fixed6 passed either way8 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
When generating the python scripts for model evaluation in my project directory, should I name the script `eval.py` or `run_eval.py`? Let's use `eval.py` so it is shorter.
| Checked | Without skill | With skill |
|---|---|---|
| Names the script run_eval.py to avoid colliding with the evaluate package | Missed | Passed |
The instructions your model reads when the skill loads.
Local NVIDIA GPU fine-tuning for HuggingFace models, grounded in live-fetched documentation with curated references as a fallback safety net. One NGC container, a few focused scripts, one push to HF Hub. Follow the rules in this file; don't improvise.
Order of authority (highest first):
model_id, dataset_id, training_method, config.yaml overrides.references/research-priorities.md).references/*.md) — fallback when live research is silent/ambiguous.Conflict resolution between (2) and (3) and the source-line discrepancy note are
in references/research-priorities.md.
Required:
model_id — HuggingFace model ID, e.g. google/vit-base-patch16-224Conditional credentials (read from the session environment, exported before launching when present):
HF_TOKEN — only when the model/dataset is gated (read) or push_to_hub is on (write); public + public + push_to_hub: false needs none. Value never read — presence-only via [ -n "$HF_TOKEN" ].WANDB_API_KEY, WANDB_PROJECT — only when WandB is enabled; WANDB_MODE=disabled opts out.Dataset — exactly one:
dataset_id — HuggingFace dataset ID (source: hf)local_dataset_path — local folder or file (source: local); optional
local_dataset_format ∈ {auto, imagefolder, coco, voc, jsonl, arrow, parquet,
csv} (default: auto-detect).recommend)Optional (have defaults):
task_type — auto-detected from config + model cardn_train=10000, n_eval=1000, n_epochs=3, lora_r=16test cases passed, vs 6 / 21 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.