Hugging Face's logo

context212
/
alhazen-ocr

Image-Text-to-Text
PEFT
Safetensors
arabic
ocr
qwen3-vl
lora
unsloth
document-ai
conversational
Model card Files
xet
 Community
younes0x53 commited on
Commit
2455fca
·
verified ·
1 Parent(s): 2000081

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -41,7 +41,7 @@ benchmark (100 samples each, greedy decoding):
41
  | Model | Eval CER ↓ | Eval WER ↓ | KHATT-bench CER ↓ | KHATT-bench WER ↓ |
42
  |---|---|---|---|---|
43
  | Qwen3-VL-2B-Instruct (base) | 0.653 | 0.951 | 1.498 | 1.594 |
44
- | **Alhazen-OCR (1 epoch)** | **0.368** | **0.567** | 2.761 | 1.707 |
45
 
46
  **Reading the numbers:** fine-tuning cuts character errors nearly in half on
47
  in-distribution documents (0.65 → 0.37 CER) and word errors from 0.95 → 0.57.
 
41
  | Model | Eval CER ↓ | Eval WER ↓ | KHATT-bench CER ↓ | KHATT-bench WER ↓ |
42
  |---|---|---|---|---|
43
  | Qwen3-VL-2B-Instruct (base) | 0.653 | 0.951 | 1.498 | 1.594 |
44
+ | **Alhazen-OCR** | **0.368** | **0.567** | 2.761 | 1.707 |
45
 
46
  **Reading the numbers:** fine-tuning cuts character errors nearly in half on
47
  in-distribution documents (0.65 → 0.37 CER) and word errors from 0.95 → 0.57.