Phase 1 checkpoint — CER=0.0348

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+phase2_files.csv filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,40 @@
+---
+language: bn
+license: mit
+tags: [ocr, bengali, crnn, easyocr, ctc]
+datasets: [mnsm92/bengali-ocr-dataset-1m]
+metrics: [cer, wer]
+---
+
+# Bengali CRNN OCR — Custom EasyOCR Recognition Model
+
+**DocReader BD — CSC4233 NLP Final Project, AIUB**
+
+## Phase 1 Results (500K samples, 10 epochs)
+
+| Model | CER ↓ | WER ↓ |
+|---|---|---|
+| Tesseract baseline | ~0.45 | ~0.60 |
+| EasyOCR default | ~0.25 | ~0.40 |
+| **BengaliCRNN Phase 1** | **0.0348** | **0.1020** |
+
+Phase 2 checkpoint (full 1M) will update this card.
+
+## Files
+- `phase1_best.pth` — Phase 1 weights (load for Phase 2 training)
+- `phase2_best.pth` — Phase 2 weights (final model for EasyOCR)
+- `vocab.json` — Bengali character vocabulary
+- `bengali_crnn.py` — EasyOCR network definition (added in Phase 2)
+
+## Usage (after Phase 2)
+```python
+import easyocr
+reader = easyocr.Reader(
+    lang_list=["bn"],
+    recog_network="bengali_crnn",
+    model_storage_directory="./model",
+    user_network_directory="./model",
+    gpu=True
+)
+results = reader.readtext("bengali_doc.jpg")
+```

phase1_best.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc6ec57fe4615f6c273d39f3efe62e85d5cd2122f36148f61a1b3f4421911cf3
+size 125944385

phase2_files.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:defc227d746be6262504bbd04b0b2068d061a1cf3e1b2b7f7ef354031f70ce4d
+size 53046834

vocab.json

@@ -0,0 +1 @@
+{"charset": "অআইঈউঊঋঌএঐওঔকখগঘঙচছজঝঞটঠডঢণতথদধনপফবভমযরলশষসহড়ঢ়য়ৎািীুূৃেৈোৌ্ংঃঁ০১২৩৪৫৬৭৮৯ ।,.?!()-–:;%/\\ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789", "num_classes": 153}