Commit
Β·
400d1bf
1
Parent(s):
89f58ba
Refactor README.md for clarity and quick start instructions for OCR scripts
Browse files
README.md
CHANGED
|
@@ -3,100 +3,68 @@ viewer: false
|
|
| 3 |
tags: [uv-script, ocr, vision-language-model, document-processing]
|
| 4 |
---
|
| 5 |
|
| 6 |
-
# UV Scripts
|
| 7 |
|
| 8 |
-
|
| 9 |
|
| 10 |
-
##
|
| 11 |
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
## Available Scripts
|
| 15 |
-
|
| 16 |
-
### 1. Nanonets OCR (`nanonets-ocr.py`)
|
| 17 |
-
|
| 18 |
-
Converts document images to structured markdown using the Nanonets-OCR-s model.
|
| 19 |
-
|
| 20 |
-
**Features:**
|
| 21 |
-
- LaTeX equation recognition
|
| 22 |
-
- Table extraction and formatting
|
| 23 |
-
- Document structure preservation
|
| 24 |
-
- Batch processing with vLLM
|
| 25 |
-
|
| 26 |
-
**Requirements:**
|
| 27 |
-
- GPU with CUDA support
|
| 28 |
-
- Python 3.11+
|
| 29 |
-
|
| 30 |
-
## Quick Test
|
| 31 |
-
|
| 32 |
-
To test the script with a sample dataset:
|
| 33 |
|
| 34 |
```bash
|
| 35 |
-
|
| 36 |
-
uv
|
| 37 |
-
|
| 38 |
-
my-test-ocr-output \
|
| 39 |
-
--max-samples 5
|
| 40 |
-
|
| 41 |
-
# Or if you have a specific dataset with images
|
| 42 |
-
uv run nanonets-ocr.py \
|
| 43 |
-
your-username/your-image-dataset \
|
| 44 |
-
your-username/test-ocr-results \
|
| 45 |
-
--image-column image \
|
| 46 |
-
--max-samples 10
|
| 47 |
```
|
| 48 |
|
| 49 |
-
|
| 50 |
|
| 51 |
-
|
| 52 |
-
-
|
| 53 |
-
-
|
| 54 |
-
- Mathematical equations in LaTeX
|
| 55 |
-
- Lists and other formatting
|
| 56 |
|
| 57 |
-
##
|
| 58 |
|
| 59 |
-
|
| 60 |
|
| 61 |
-
|
| 62 |
-
uv run nanonets-ocr.py input output \
|
| 63 |
-
--batch-size 4 \
|
| 64 |
-
--gpu-memory-utilization 0.5
|
| 65 |
-
```
|
| 66 |
|
| 67 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 68 |
|
| 69 |
-
|
| 70 |
|
| 71 |
-
###
|
|
|
|
|
|
|
| 72 |
|
| 73 |
```bash
|
| 74 |
-
# Basic
|
| 75 |
hf jobs uv run --flavor l4x1 \
|
| 76 |
https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py \
|
| 77 |
-
input-dataset-
|
| 78 |
|
| 79 |
-
#
|
| 80 |
-
hf jobs uv run
|
|
|
|
|
|
|
|
|
|
| 81 |
https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py \
|
| 82 |
-
|
| 83 |
-
your-username/
|
| 84 |
--image-column image \
|
| 85 |
--max-model-len 16384 \
|
| 86 |
-
--batch-size
|
| 87 |
|
| 88 |
-
#
|
| 89 |
-
hf jobs uv run --flavor
|
|
|
|
| 90 |
https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py \
|
| 91 |
-
input-
|
| 92 |
-
--private
|
| 93 |
-
|
| 94 |
-
# With vLLM Docker image for optimized performance
|
| 95 |
-
hf jobs uv run \
|
| 96 |
-
--flavor l4x1 \
|
| 97 |
-
--image vllm/vllm-openai:latest \
|
| 98 |
-
https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py \
|
| 99 |
-
input-dataset output-dataset \
|
| 100 |
--batch-size 32
|
| 101 |
```
|
| 102 |
|
|
@@ -105,30 +73,37 @@ hf jobs uv run \
|
|
| 105 |
```python
|
| 106 |
from huggingface_hub import run_uv_job
|
| 107 |
|
| 108 |
-
# Run the OCR script
|
| 109 |
job = run_uv_job(
|
| 110 |
"https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py",
|
| 111 |
-
args=[
|
| 112 |
-
|
| 113 |
-
"output-dataset-id",
|
| 114 |
-
"--image-column", "image",
|
| 115 |
-
"--max-model-len", "16384"
|
| 116 |
-
],
|
| 117 |
-
flavor="l4x1",
|
| 118 |
-
secrets={"HF_TOKEN": "your-token"} # if needed
|
| 119 |
)
|
| 120 |
```
|
| 121 |
|
| 122 |
-
###
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 123 |
|
| 124 |
-
|
| 125 |
-
- **`t4-small`** (16GB) - For smaller batches or lower resolution
|
| 126 |
-
- **`a10g-small`** (24GB) - Alternative to L4
|
| 127 |
-
- **`l40sx1`** (48GB) - For very large batches
|
| 128 |
-
- **`a100-large`** (80GB) - Maximum performance
|
| 129 |
|
| 130 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 131 |
|
| 132 |
-
|
| 133 |
-
- Performance benchmarks
|
| 134 |
-
- More examples and use cases
|
|
|
|
| 3 |
tags: [uv-script, ocr, vision-language-model, document-processing]
|
| 4 |
---
|
| 5 |
|
| 6 |
+
# OCR UV Scripts
|
| 7 |
|
| 8 |
+
Ready-to-run OCR scripts that work with `uv run` - no setup required!
|
| 9 |
|
| 10 |
+
## π Quick Start with HuggingFace Jobs
|
| 11 |
|
| 12 |
+
Run OCR on any dataset without a GPU:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
|
| 14 |
```bash
|
| 15 |
+
hf jobs uv run --flavor l4x1 \
|
| 16 |
+
https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py \
|
| 17 |
+
your-input-dataset your-output-dataset
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
```
|
| 19 |
|
| 20 |
+
That's it! The script will:
|
| 21 |
|
| 22 |
+
- β
Process all images in your dataset
|
| 23 |
+
- β
Add OCR results as a new `markdown` column
|
| 24 |
+
- β
Push the results to a new dataset
|
|
|
|
|
|
|
| 25 |
|
| 26 |
+
## π Available Scripts
|
| 27 |
|
| 28 |
+
### Nanonets OCR (`nanonets-ocr.py`)
|
| 29 |
|
| 30 |
+
State-of-the-art document OCR using [nanonets/Nanonets-OCR-s](https://huggingface.co/nanonets/Nanonets-OCR-s) that handles:
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
|
| 32 |
+
- π **LaTeX equations** - Mathematical formulas preserved
|
| 33 |
+
- π **Tables** - Extracted as HTML format
|
| 34 |
+
- π **Document structure** - Headers, lists, formatting maintained
|
| 35 |
+
- πΌοΈ **Images** - Captions and descriptions included
|
| 36 |
+
- βοΈ **Forms** - Checkboxes rendered as β/β
|
| 37 |
|
| 38 |
+
## π» Usage Examples
|
| 39 |
|
| 40 |
+
### Run on HuggingFace Jobs (Recommended)
|
| 41 |
+
|
| 42 |
+
No GPU? No problem! Run on HF infrastructure:
|
| 43 |
|
| 44 |
```bash
|
| 45 |
+
# Basic OCR job
|
| 46 |
hf jobs uv run --flavor l4x1 \
|
| 47 |
https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py \
|
| 48 |
+
your-input-dataset your-output-dataset
|
| 49 |
|
| 50 |
+
# Real example with UFO dataset πΈ
|
| 51 |
+
hf jobs uv run \
|
| 52 |
+
--flavor a10g-large \
|
| 53 |
+
--image vllm/vllm-openai:latest \
|
| 54 |
+
-e HF_TOKEN=$(python3 -c "from huggingface_hub import get_token; print(get_token())") \
|
| 55 |
https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py \
|
| 56 |
+
davanstrien/ufo-ColPali \
|
| 57 |
+
your-username/ufo-ocr \
|
| 58 |
--image-column image \
|
| 59 |
--max-model-len 16384 \
|
| 60 |
+
--batch-size 64
|
| 61 |
|
| 62 |
+
# Private dataset with custom settings
|
| 63 |
+
hf jobs uv run --flavor l40sx1 \
|
| 64 |
+
-e HF_TOKEN=$(python3 -c "from huggingface_hub import get_token; print(get_token())") \
|
| 65 |
https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py \
|
| 66 |
+
private-input private-output \
|
| 67 |
+
--private \
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 68 |
--batch-size 32
|
| 69 |
```
|
| 70 |
|
|
|
|
| 73 |
```python
|
| 74 |
from huggingface_hub import run_uv_job
|
| 75 |
|
|
|
|
| 76 |
job = run_uv_job(
|
| 77 |
"https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py",
|
| 78 |
+
args=["input-dataset", "output-dataset", "--batch-size", "16"],
|
| 79 |
+
flavor="l4x1"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
)
|
| 81 |
```
|
| 82 |
|
| 83 |
+
### Run Locally (Requires GPU)
|
| 84 |
+
|
| 85 |
+
```bash
|
| 86 |
+
# Clone and run
|
| 87 |
+
git clone https://huggingface.co/datasets/uv-scripts/ocr
|
| 88 |
+
cd ocr
|
| 89 |
+
uv run nanonets-ocr.py input-dataset output-dataset
|
| 90 |
+
|
| 91 |
+
# Or run directly from URL
|
| 92 |
+
uv run https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py \
|
| 93 |
+
input-dataset output-dataset
|
| 94 |
+
```
|
| 95 |
|
| 96 |
+
## ποΈ Configuration Options
|
|
|
|
|
|
|
|
|
|
|
|
|
| 97 |
|
| 98 |
+
| Option | Default | Description |
|
| 99 |
+
| -------------------------- | ------- | --------------------------- |
|
| 100 |
+
| `--image-column` | `image` | Column containing images |
|
| 101 |
+
| `--batch-size` | `8` | Images processed together |
|
| 102 |
+
| `--max-model-len` | `8192` | Max context length |
|
| 103 |
+
| `--max-tokens` | `4096` | Max output tokens |
|
| 104 |
+
| `--gpu-memory-utilization` | `0.7` | GPU memory usage |
|
| 105 |
+
| `--split` | `train` | Dataset split to process |
|
| 106 |
+
| `--max-samples` | None | Limit samples (for testing) |
|
| 107 |
+
| `--private` | False | Make output dataset private |
|
| 108 |
|
| 109 |
+
More OCR VLM Scripts coming soon! Stay tuned for updates!
|
|
|
|
|
|