Upload OCR artifacts
Browse files- .gitattributes +1 -0
- README.md +27 -3
- artifact_metadata.jsonl +1 -0
- ocr.jsonl +3 -0
.gitattributes
CHANGED
|
@@ -58,3 +58,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 58 |
# Video files - compressed
|
| 59 |
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 60 |
*.webm filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 58 |
# Video files - compressed
|
| 59 |
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 60 |
*.webm filter=lfs diff=lfs merge=lfs -text
|
| 61 |
+
ocr.jsonl filter=lfs diff=lfs merge=lfs -text
|
README.md
CHANGED
|
@@ -1,3 +1,27 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# OCR
|
| 2 |
+
|
| 3 |
+
The primary artifact is `ocr.jsonl` còn `artifact_metadata.jsonl` records the input, model, prompt version, subset, and Git lineage
|
| 4 |
+
|
| 5 |
+
## `ocr.jsonl`
|
| 6 |
+
|
| 7 |
+
Each line is one `OCRFragment`, a visible text region in one keyframe.
|
| 8 |
+
|
| 9 |
+
Common fields:
|
| 10 |
+
|
| 11 |
+
- `fragment_id`: OCR observation identifier
|
| 12 |
+
- `video_id`: source video identifier
|
| 13 |
+
- `time_span`: `[timestamp_ms, timestamp_ms]` in source video time
|
| 14 |
+
- `text`: recognized visible text
|
| 15 |
+
- `bbox`: normalized text region with `x_min`, `y_min`, `x_max`, and `y_max`
|
| 16 |
+
|
| 17 |
+
There is no placeholder for frames without text and no fabricated confidence score. Join OCR with captions or detections on `video_id` and timestamp; compare normalized bounding boxes when spatial context is needed.
|
| 18 |
+
|
| 19 |
+
## Generation flow
|
| 20 |
+
|
| 21 |
+
HunyuanOCR reads the canonical keyframe images, emits visible text regions, and the job validates each result as an `OCRFragment`. It publishes one flat JSONL file and removes provider work files and the retired bundled OCR representation after a successful run.
|
| 22 |
+
|
| 23 |
+
Run OCR for the original keyframe set:
|
| 24 |
+
|
| 25 |
+
```bash
|
| 26 |
+
make data-ocr INPUT=/data/aic/shared/artifacts/keyframes/keyframes.jsonl OUTPUT=/data/aic/shared/artifacts/ocr SUBSET=700vid
|
| 27 |
+
```
|
artifact_metadata.jsonl
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"artifact":"ocr.jsonl","input":"../keyframes/keyframes.jsonl","subset":"5vid","stage":"ocr","model":"tencent/HunyuanOCR@a8d9d9c5a30bba088029bc8cbab098c321a71e47","prompt_version":"spotting_json","git_commit":"3133e14a82dd827eb273484314e7a99be85725d3"}
|
ocr.jsonl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4b6aca69c0deb921179c8fd3f8ebfce3ea8570e7cb53027d76875fa83c8692f3
|
| 3 |
+
size 21341544
|