Publish the dataset to the Hub and verify the round trip

Aryan006/cone-distance-v1, private: 2.38 GB as three tar shards alongside the
manifest, sixteen calibration files, the derived class priors and a generated
card. Private deliberately -- nuScenes and Argoverse 2 carry non-commercial
research licences that restrict redistributing derived imagery.

Pulled it back down and checked it, since "uploaded" and "trains on Colab" are
not the same claim: 8,446 train and 1,815 val images restored as real files
rather than dangling links, every image paired with a label, data.yaml rewritten
to an absolute path.

Both hub tools now read HF_TOKEN from the environment. A token passed as --token
is visible to anyone who can run ps.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

notebooks/colab_train.ipynb

@@ -63,7 +63,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "REPO_ID = 'YOUR_USER/cone-distance-v1'  # <-- edit me\n",
+    "REPO_ID = 'Aryan006/cone-distance-v1'\n",
     "\n",
     "try:\n",
     "    from google.colab import userdata\n",

src/hub/pull.py

@@ -12,6 +12,7 @@ Usage:
 from __future__ import annotations
 
 import argparse
+import os
 import tarfile
 from pathlib import Path
 
@@ -71,7 +72,11 @@ def main() -> None:
                                      formatter_class=argparse.RawDescriptionHelpFormatter)
     parser.add_argument("--repo-id", required=True)
     parser.add_argument("--dest", type=Path, default=Path("data/hub"))
-    parser.add_argument("--token", default=None)
+    parser.add_argument("--token", default=os.environ.get("HF_TOKEN"),
+                        help="defaults to $HF_TOKEN, then the cached "
+                             "huggingface-cli login. Prefer the environment: a "
+                             "token passed on the command line is visible to "
+                             "anyone who can run ps.")
     parser.add_argument("--keep-shards", action="store_true",
                         help="keep the .tar files after extracting (doubles disk use)")
     args = parser.parse_args()

src/hub/push.py

@@ -27,6 +27,7 @@ Usage:
 from __future__ import annotations
 
 import argparse
+import os
 import shutil
 import tarfile
 import tempfile
@@ -237,8 +238,11 @@ def main() -> None:
     parser.add_argument("--repo-id", required=True, help="e.g. <user>/cone-distance-v1")
     parser.add_argument("--unified-root", type=Path, default=None)
     parser.add_argument("--private", action="store_true")
-    parser.add_argument("--token", default=None,
-                        help="defaults to the cached huggingface-cli login")
+    parser.add_argument("--token", default=os.environ.get("HF_TOKEN"),
+                        help="defaults to $HF_TOKEN, then the cached "
+                             "huggingface-cli login. Prefer the environment: a "
+                             "token passed on the command line is visible to "
+                             "anyone who can run ps.")
     parser.add_argument("--shard-bytes", type=int, default=DEFAULT_SHARD_BYTES)
     parser.add_argument("--staging-dir", type=Path, default=None,
                         help="keep the staging directory here instead of a temp dir")