Fix dataset target: exp004 trained on diffusion-pretrain-set-ft1 data/deepfashion (has caption_vlm_json), not qwen-deepfashion-fused; shard listing filtered to the config (the repo holds 582 parquet across configs). Caught by CELL 0 refusing before spend — as designed.

colab/anima_closeout.ipynb

@@ -138,11 +138,20 @@
     "print(\"Anima split_files present \u2014 NC weights; derived ckpts inherit NC\")\n",
     "\n",
     "# -- 5. dataset columns, verified BEFORE anything spends ------------------\n",
-    "DATASET = \"AbstractPhil/qwen-deepfashion-fused\"\n",
+    "# exp004's actual source: diffusion-pretrain-set-ft1, deepfashion config \u2014\n",
+    "# the dataset whose shards were the pod's ft1_shards and which carries\n",
+    "# caption_vlm_json. (qwen-deepfashion-fused does NOT have that column; the\n",
+    "# first campaign run caught this exactly here, at zero cost.) The repo\n",
+    "# holds 582 parquet files across many configs \u2014 filter to deepfashion or\n",
+    "# the split cell would download all of them.\n",
+    "DATASET = \"AbstractPhil/diffusion-pretrain-set-ft1\"\n",
+    "DATASET_SUBDIR = \"data/deepfashion/\"\n",
     "from huggingface_hub import HfApi\n",
     "SHARDS = sorted(f for f in HfApi().list_repo_files(\n",
-    "    DATASET, repo_type=\"dataset\") if f.endswith(\".parquet\"))\n",
-    "assert SHARDS, f\"no parquet shards in {DATASET}\"\n",
+    "    DATASET, repo_type=\"dataset\")\n",
+    "    if f.endswith(\".parquet\") and f.startswith(DATASET_SUBDIR))\n",
+    "assert SHARDS, f\"no parquet shards under {DATASET_SUBDIR} in {DATASET}\"\n",
+    "print(f\"deepfashion shards: {len(SHARDS)}\")\n",
     "import pyarrow.parquet as pq\n",
     "_p0 = hf_hub_download(DATASET, SHARDS[0], repo_type=\"dataset\")\n",
     "_cols = set(pq.read_schema(_p0).names)\n",