Upload setup_and_train_300m.sh with huggingface_hub

setup_and_train_300m.sh

@@ -63,6 +63,27 @@ else
     exit 1
 fi
 
+python -c "
+from pathlib import Path
+import json
+data_dir = Path('$DATA_DIR')
+index_path = data_dir / 'index.json'
+if not index_path.exists():
+    candidates = sorted(data_dir.glob('**/index.json'))
+    if candidates:
+        print('ERROR: index.json still not at dataset root after fetch.')
+        print('Found candidates:')
+        for candidate in candidates[:10]:
+            print(f'  - {candidate}')
+    else:
+        print(f'ERROR: No index.json found under {data_dir}')
+    raise SystemExit(1)
+with open(index_path, encoding='utf-8') as f:
+    idx = json.load(f)
+total = sum(c.get('dim', 0) for c in idx.get('chunks', []))
+print(f'  Dataset ready: {len(idx.get("chunks", []))} chunks, {total:,} tokens at {index_path}')
+"
+
 # ── 3. System probe ──────────────────────────────────────────────────────────
 echo ""
 echo "[3/5] System probe..."