Update README.md
Browse files
README.md
CHANGED
|
@@ -101,14 +101,3 @@ print(tokenizer.decode(out[0][inputs["input_ids"].shape[1]:], skip_special_token
|
|
| 101 |
|
| 102 |
Eval benchmark: `ahmedheakl/arocrbench_khatt` (MIT).
|
| 103 |
|
| 104 |
-
## Limitations
|
| 105 |
-
|
| 106 |
-
- First smoke run (20k samples, 1 epoch) — handwriting and out-of-distribution
|
| 107 |
-
documents remain weak (see KHATT-bench column)
|
| 108 |
-
- Trained on line/paragraph-level crops; full-page layout parsing untested
|
| 109 |
-
- Arabic-first; other scripts not evaluated
|
| 110 |
-
|
| 111 |
-
## Code
|
| 112 |
-
|
| 113 |
-
Training, data pipeline and evaluation:
|
| 114 |
-
[`github.com/context212/atlas-ocr`](https://github.com/context212/atlas-ocr)
|
|
|
|
| 101 |
|
| 102 |
Eval benchmark: `ahmedheakl/arocrbench_khatt` (MIT).
|
| 103 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|