Instructions to use AlexWortega/tinyvla with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- LeRobot
How to use AlexWortega/tinyvla with LeRobot:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
wds mixture: unitree+nav converters (gnm/vamos), wds source branch in train_fast, AMP autoselect (V100 fp16), embedding expand 16->32, push/fetch/validate tooling, DATA.md
b152c62 verified Download tinyvla_b200/scripts/hf_push_wds.py from AlexWortega/tinyvla: direct link, hf CLI and curl.
- Browser
- Download file 1.91 kB
-
https://huggingface.co/AlexWortega/tinyvla/resolve/main/tinyvla_b200/scripts/hf_push_wds.py
- Command line
-
hf download hf://AlexWortega/tinyvla/tinyvla_b200/scripts/hf_push_wds.py
-
curl -L -o hf_push_wds.py https://huggingface.co/AlexWortega/tinyvla/resolve/main/tinyvla_b200/scripts/hf_push_wds.py
1.91 kB
| #!/usr/bin/env python | |
| """Push a packed wds dataset dir (shard-*.tar + manifest.json + tasks.json) to | |
| the Hub as a private dataset repo, and verify the listing afterwards. | |
| Usage: | |
| python hf_push_wds.py --src /data/wds/tartandrive --repo AlexWortega/wds-tartandrive | |
| """ | |
| from __future__ import annotations | |
| import argparse | |
| import json | |
| from pathlib import Path | |
| def main(): | |
| ap = argparse.ArgumentParser() | |
| ap.add_argument("--src", type=Path, required=True) | |
| ap.add_argument("--repo", required=True) | |
| ap.add_argument("--public", action="store_true") | |
| args = ap.parse_args() | |
| from huggingface_hub import HfApi | |
| # single packed dataset (manifest at top) or a bundle of per-task subdirs | |
| if (args.src / "manifest.json").exists(): | |
| dirs = {"": json.loads((args.src / "manifest.json").read_text())} | |
| else: | |
| dirs = {p.name + "/": json.loads((p / "manifest.json").read_text()) | |
| for p in sorted(args.src.iterdir()) | |
| if p.is_dir() and (p / "manifest.json").exists()} | |
| if not dirs: | |
| raise SystemExit(f"{args.src}: no manifest.json anywhere") | |
| api = HfApi() | |
| api.create_repo(args.repo, repo_type="dataset", private=not args.public, exist_ok=True) | |
| api.upload_large_folder(repo_id=args.repo, repo_type="dataset", | |
| folder_path=str(args.src)) | |
| remote = set(api.list_repo_files(args.repo, repo_type="dataset")) | |
| missing, total = [], 0 | |
| for pfx, mf in dirs.items(): | |
| missing += [pfx + s for s in mf["shards"] if pfx + s not in remote] | |
| missing += [pfx + f for f in ("manifest.json", "tasks.json") if pfx + f not in remote] | |
| total += mf["samples"] | |
| if missing: | |
| raise SystemExit(f"upload incomplete, missing: {missing[:10]}") | |
| print(f"{args.repo}: {len(dirs)} dataset dir(s), {total:,} samples, all shards verified") | |
| if __name__ == "__main__": | |
| main() | |