Record visual conversion provenance and exclusions

provenance.json

@@ -3,16 +3,16 @@
   "revision": "934f3928b7311ec70c080206d8199bbf3697b6c7",
   "format_version": 1,
   "splits": {
-    "validation": {
-      "rows": 83,
-      "questions": 83
-    },
     "train": {
       "rows": 94157,
       "questions": 100000
+    },
+    "validation": {
+      "rows": 83,
+      "questions": 83
     }
   },
-  "conversion_sha256": "661d8c8d0e8b96f2e5b0c38c91ffc3ad7f963fcf04d0f0269368be21ae456682",
+  "conversion_sha256": "b7a50f24845f40adad540de01426d7d688e894313641f19a1a11cf580440fcdf",
   "excluded_rows": 0,
   "license": "research use",
   "license_url": "https://github.com/YangTianze009/TRON#license",
@@ -58,8 +58,8 @@
     "tron.py": "59b1fed112224ec1aabff0ad2780b710a13b9010ac3d8252c9c81931e3d1d474"
   },
   "parquet_sha256": {
-    "train.parquet": "cf5f50c29142961ccd86bb759258c56a13368b7ab946f00ee01c42ed3195cfbf",
-    "validation.parquet": "c6d89f1e7451bf1ce8b5d72b8929d092f459854362ba6f29a3731312bc722052"
+    "train.parquet": "590799478644d5d4e5b4bef594a39b948f3b9c3ef7e04550844b27c84407bb0b",
+    "validation.parquet": "92bc63e1e36655a68d69c97e647baec66b9d5766c9e06404e83108ba3e488c39"
   },
   "image_overlap_excluded": 126,
   "checkpoint_provenance": [
@@ -2222,5 +2222,10 @@
     "old_scale_sha256": "d8d7b2dffc00d14dc534b0eb02897a160646f3ef88d151a5c32949ea163953dc",
     "new_scale_sha256": "7829cb5948b61bbc5b7b51e7f6e540116589406037921587c8c49e8b117caae2",
     "reason": "Preserve global deduplication state through Datasets generator iteration; assert final counts. Generator, seeds, native source code and checkpoints are unchanged."
-  }
+  },
+  "base_revision": "7da8ffc23dde85af9d985cabd2fc71c31b8f886d",
+  "question_repairs": {
+    "rounded_measurements": 26670
+  },
+  "repair_script_sha256": "cdbda27c98aa017e8067703113fd1e0daa23df9f7ec519bd9ac8855bf0ff0fe6"
 }