datasety 0.1.0__tar.gz → 0.2.0__tar.gz

datasety-0.2.0/.gitignore

@@ -0,0 +1,94 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+build/
+dist/
+*.egg-info/
+*.egg
+wheels/
+pip-wheel-metadata/
+MANIFEST
+
+# Virtual environments
+venv/
+.venv/
+env/
+.env/
+ENV/
+
+# IDE / Editors
+.idea/
+.vscode/
+*.swp
+*.swo
+*~
+.project
+.pydevproject
+.settings/
+*.sublime-project
+*.sublime-workspace
+
+# Testing / Coverage
+.pytest_cache/
+.tox/
+.nox/
+.coverage
+.coverage.*
+htmlcov/
+.hypothesis/
+*.cover
+coverage.xml
+nosetests.xml
+
+# Type checking
+.mypy_cache/
+.dmypy.json
+dmypy.json
+.pytype/
+
+# Linting
+.ruff_cache/
+
+# Jupyter
+.ipynb_checkpoints/
+*.ipynb
+
+# OS
+.DS_Store
+.DS_Store?
+._*
+.Spotlight-V100
+.Trashes
+Thumbs.db
+ehthumbs.db
+Desktop.ini
+
+# Logs
+*.log
+logs/
+
+# ML / AI specific
+*.pt
+*.pth
+*.ckpt
+*.safetensors
+*.bin
+.cache/
+huggingface/
+~/.cache/huggingface/
+
+# Project specific - test data directories
+test_data/
+tests/fixtures/images/
+/in/
+/out/
+/synthetic/
+/captions/
+/resized/
+/dataset/

datasety-0.2.0/PKG-INFO

@@ -0,0 +1,231 @@
+Metadata-Version: 2.4
+Name: datasety
+Version: 0.2.0
+Summary: CLI tool for dataset preparation: resize, caption, and synthetic image generation
+Project-URL: Homepage, https://github.com/kontextox/datasety
+Project-URL: Repository, https://github.com/kontextox/datasety
+Project-URL: Issues, https://github.com/kontextox/datasety/issues
+Author: kontextox
+License-Expression: MIT
+License-File: LICENSE
+Keywords: captioning,cli,dataset,diffusers,florence-2,image-editing,image-processing,machine-learning,synthetic
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Scientific/Engineering :: Image Processing
+Requires-Python: >=3.10
+Requires-Dist: pillow>=9.0.0
+Provides-Extra: all
+Requires-Dist: accelerate; extra == 'all'
+Requires-Dist: diffusers>=0.32.0; extra == 'all'
+Requires-Dist: einops; extra == 'all'
+Requires-Dist: sentencepiece; extra == 'all'
+Requires-Dist: timm; extra == 'all'
+Requires-Dist: torch>=2.0.0; extra == 'all'
+Requires-Dist: transformers<4.46.0,>=4.38.0; extra == 'all'
+Requires-Dist: transformers>=4.38.0; extra == 'all'
+Provides-Extra: caption
+Requires-Dist: einops; extra == 'caption'
+Requires-Dist: timm; extra == 'caption'
+Requires-Dist: torch>=2.0.0; extra == 'caption'
+Requires-Dist: transformers<4.46.0,>=4.38.0; extra == 'caption'
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0.0; extra == 'dev'
+Requires-Dist: ruff>=0.1.0; extra == 'dev'
+Provides-Extra: synthetic
+Requires-Dist: accelerate; extra == 'synthetic'
+Requires-Dist: diffusers>=0.32.0; extra == 'synthetic'
+Requires-Dist: sentencepiece; extra == 'synthetic'
+Requires-Dist: torch>=2.0.0; extra == 'synthetic'
+Requires-Dist: transformers>=4.38.0; extra == 'synthetic'
+Description-Content-Type: text/markdown
+
+# datasety
+
+CLI tool for dataset preparation: resize, caption, and synthetic image generation.
+
+## Installation
+
+```bash
+pip install datasety
+```
+
+Install with specific features:
+
+```bash
+pip install datasety[caption]     # Florence-2 captioning
+pip install datasety[synthetic]   # Qwen image editing
+pip install datasety[all]         # All features
+```
+
+## Usage
+
+### Resize Images
+
+Resize and crop images to a target resolution:
+
+```bash
+datasety resize --input ./images --output ./resized --resolution 768x1024
+```
+
+**Options:**
+
+| Option                  | Description                                               | Default             |
+| ----------------------- | --------------------------------------------------------- | ------------------- |
+| `--input`, `-i`         | Input directory                                           | (required)          |
+| `--output`, `-o`        | Output directory                                          | (required)          |
+| `--resolution`, `-r`    | Target resolution (WIDTHxHEIGHT)                          | (required)          |
+| `--crop-position`       | Crop position: `top`, `center`, `bottom`, `left`, `right` | `center`            |
+| `--input-format`        | Comma-separated formats                                   | `jpg,jpeg,png,webp` |
+| `--output-format`       | Output format: `jpg`, `png`, `webp`                       | `jpg`               |
+| `--output-name-numbers` | Rename files to 1.jpg, 2.jpg, ...                         | `false`             |
+
+**Example:**
+
+```bash
+datasety resize \
+    --input ./raw_photos \
+    --output ./dataset \
+    --resolution 1024x1024 \
+    --crop-position top \
+    --output-format jpg \
+    --output-name-numbers
+```
+
+**How it works:**
+
+1. Finds all images matching input formats
+2. Skips images where either dimension is smaller than target
+3. Resizes proportionally so the smaller side matches target
+4. Crops from the specified position to exact dimensions
+5. Saves with high quality (95% for jpg/webp)
+
+### Generate Captions
+
+Generate captions for images using Microsoft's Florence-2 model:
+
+```bash
+datasety caption --input ./images --output ./captions --florence-2-large
+```
+
+**Options:**
+
+| Option               | Description                     | Default                   |
+| -------------------- | ------------------------------- | ------------------------- |
+| `--input`, `-i`      | Input directory                 | (required)                |
+| `--output`, `-o`     | Output directory for .txt files | (required)                |
+| `--device`           | `cpu` or `cuda`                 | `cpu`                     |
+| `--trigger-word`     | Text to prepend to captions     | (none)                    |
+| `--prompt`           | Florence-2 task prompt          | `<MORE_DETAILED_CAPTION>` |
+| `--florence-2-base`  | Use base model (0.23B, faster)  |                           |
+| `--florence-2-large` | Use large model (0.77B, better) | (default)                 |
+
+**Available prompts:**
+
+- `<CAPTION>` - Brief caption
+- `<DETAILED_CAPTION>` - Detailed caption
+- `<MORE_DETAILED_CAPTION>` - Most detailed caption (default)
+
+**Example:**
+
+```bash
+datasety caption \
+    --input ./dataset \
+    --output ./dataset \
+    --device cuda \
+    --trigger-word "photo of sks person," \
+    --florence-2-large
+```
+
+This creates a `.txt` file for each image with the generated caption.
+
+### Generate Synthetic Images
+
+Generate synthetic variations of images using Qwen-Image-Edit:
+
+```bash
+datasety synthetic --input ./images --output ./synthetic --prompt "add a winter hat"
+```
+
+**Options:**
+
+| Option              | Description                       | Default                    |
+| ------------------- | --------------------------------- | -------------------------- |
+| `--input`, `-i`     | Input directory                   | (required)                 |
+| `--output`, `-o`    | Output directory                  | (required)                 |
+| `--prompt`, `-p`    | Edit prompt                       | (required)                 |
+| `--model`           | Model to use                      | `Qwen/Qwen-Image-Edit-2511`|
+| `--device`          | `cpu` or `cuda`                   | `cuda`                     |
+| `--steps`           | Number of inference steps         | `40`                       |
+| `--cfg-scale`       | Guidance scale                    | `1.0`                      |
+| `--true-cfg-scale`  | True CFG scale                    | `4.0`                      |
+| `--negative-prompt` | Negative prompt                   | `" "`                      |
+| `--num-images`      | Images to generate per input      | `1`                        |
+| `--seed`            | Random seed for reproducibility   | (random)                   |
+
+**Example:**
+
+```bash
+datasety synthetic \
+    --input ./dataset \
+    --output ./synthetic \
+    --prompt "add sunglasses to the person, keep everything else the same" \
+    --device cuda \
+    --steps 40 \
+    --true-cfg-scale 4.0 \
+    --seed 42
+```
+
+## Common Workflows
+
+### Prepare a LoRA Training Dataset
+
+```bash
+# 1. Resize images to 1024x1024
+datasety resize -i ./raw -o ./dataset -r 1024x1024 --crop-position center
+
+# 2. Generate captions with trigger word
+datasety caption -i ./dataset -o ./dataset --trigger-word "[trigger]" --device cuda
+```
+
+### Augment Dataset with Synthetic Variations
+
+```bash
+# Generate variations with different accessories
+datasety synthetic \
+    -i ./dataset \
+    -o ./synthetic \
+    --prompt "add a red scarf" \
+    --num-images 2 \
+    --device cuda
+```
+
+### Batch Process with Numbered Files
+
+```bash
+datasety resize \
+    -i ./photos \
+    -o ./processed \
+    -r 768x1024 \
+    --output-name-numbers \
+    --crop-position top
+```
+
+## Requirements
+
+- Python 3.10+
+- Pillow (for resize)
+- PyTorch + Transformers (for caption: `pip install datasety[caption]`)
+- PyTorch + Diffusers (for synthetic: `pip install datasety[synthetic]`)
+
+## License
+
+MIT

datasety-0.1.0/PKG-INFO → datasety-0.2.0/README.md

@@ -1,41 +1,6 @@
-Metadata-Version: 2.4
-Name: datasety
-Version: 0.1.0
-Summary: CLI tool for dataset preparation: image resizing and captioning with Florence-2
-Project-URL: Homepage, https://github.com/kontextox/datasety
-Project-URL: Repository, https://github.com/kontextox/datasety
-Project-URL: Issues, https://github.com/kontextox/datasety/issues
-Author: kontextox
-License-Expression: MIT
-License-File: LICENSE
-Keywords: captioning,cli,dataset,florence-2,image-processing,machine-learning
-Classifier: Development Status :: 4 - Beta
-Classifier: Environment :: Console
-Classifier: Intended Audience :: Developers
-Classifier: Intended Audience :: Science/Research
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Operating System :: OS Independent
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.10
-Classifier: Programming Language :: Python :: 3.11
-Classifier: Programming Language :: Python :: 3.12
-Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
-Classifier: Topic :: Scientific/Engineering :: Image Processing
-Requires-Python: >=3.10
-Requires-Dist: pillow>=9.0.0
-Provides-Extra: caption
-Requires-Dist: einops; extra == 'caption'
-Requires-Dist: timm; extra == 'caption'
-Requires-Dist: torch>=2.0.0; extra == 'caption'
-Requires-Dist: transformers<4.46.0,>=4.38.0; extra == 'caption'
-Provides-Extra: dev
-Requires-Dist: pytest>=7.0.0; extra == 'dev'
-Requires-Dist: ruff>=0.1.0; extra == 'dev'
-Description-Content-Type: text/markdown
-
 # datasety
 
-CLI tool for dataset preparation: image resizing and captioning with Florence-2.
+CLI tool for dataset preparation: resize, caption, and synthetic image generation.
 
 ## Installation
 
@@ -43,10 +8,12 @@ CLI tool for dataset preparation: image resizing and captioning with Florence-2.
 pip install datasety
 ```
 
-For captioning support (requires PyTorch and Transformers):
+Install with specific features:
 
 ```bash
-pip install datasety[caption]
+pip install datasety[caption]     # Florence-2 captioning
+pip install datasety[synthetic]   # Qwen image editing
+pip install datasety[all]         # All features
 ```
 
 ## Usage
@@ -88,7 +55,7 @@ datasety resize \
 1. Finds all images matching input formats
 2. Skips images where either dimension is smaller than target
 3. Resizes proportionally so the smaller side matches target
-4. Crops from the specified area to exact dimensions
+4. Crops from the specified position to exact dimensions
 5. Saves with high quality (95% for jpg/webp)
 
 ### Generate Captions
@@ -130,6 +97,43 @@ datasety caption \
 
 This creates a `.txt` file for each image with the generated caption.
 
+### Generate Synthetic Images
+
+Generate synthetic variations of images using Qwen-Image-Edit:
+
+```bash
+datasety synthetic --input ./images --output ./synthetic --prompt "add a winter hat"
+```
+
+**Options:**
+
+| Option              | Description                       | Default                    |
+| ------------------- | --------------------------------- | -------------------------- |
+| `--input`, `-i`     | Input directory                   | (required)                 |
+| `--output`, `-o`    | Output directory                  | (required)                 |
+| `--prompt`, `-p`    | Edit prompt                       | (required)                 |
+| `--model`           | Model to use                      | `Qwen/Qwen-Image-Edit-2511`|
+| `--device`          | `cpu` or `cuda`                   | `cuda`                     |
+| `--steps`           | Number of inference steps         | `40`                       |
+| `--cfg-scale`       | Guidance scale                    | `1.0`                      |
+| `--true-cfg-scale`  | True CFG scale                    | `4.0`                      |
+| `--negative-prompt` | Negative prompt                   | `" "`                      |
+| `--num-images`      | Images to generate per input      | `1`                        |
+| `--seed`            | Random seed for reproducibility   | (random)                   |
+
+**Example:**
+
+```bash
+datasety synthetic \
+    --input ./dataset \
+    --output ./synthetic \
+    --prompt "add sunglasses to the person, keep everything else the same" \
+    --device cuda \
+    --steps 40 \
+    --true-cfg-scale 4.0 \
+    --seed 42
+```
+
 ## Common Workflows
 
 ### Prepare a LoRA Training Dataset
@@ -142,6 +146,18 @@ datasety resize -i ./raw -o ./dataset -r 1024x1024 --crop-position center
 datasety caption -i ./dataset -o ./dataset --trigger-word "[trigger]" --device cuda
 ```
 
+### Augment Dataset with Synthetic Variations
+
+```bash
+# Generate variations with different accessories
+datasety synthetic \
+    -i ./dataset \
+    -o ./synthetic \
+    --prompt "add a red scarf" \
+    --num-images 2 \
+    --device cuda
+```
+
 ### Batch Process with Numbered Files
 
 ```bash
@@ -157,7 +173,8 @@ datasety resize \
 
 - Python 3.10+
 - Pillow (for resize)
-- PyTorch + Transformers (for caption, install with `pip install datasety[caption]`)
+- PyTorch + Transformers (for caption: `pip install datasety[caption]`)
+- PyTorch + Diffusers (for synthetic: `pip install datasety[synthetic]`)
 
 ## License
 

{datasety-0.1.0 → datasety-0.2.0}/pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "hatchling.build"
 [project]
 name = "datasety"
 dynamic = ["version"]
-description = "CLI tool for dataset preparation: image resizing and captioning with Florence-2"
+description = "CLI tool for dataset preparation: resize, caption, and synthetic image generation"
 readme = "README.md"
 license = "MIT"
 requires-python = ">=3.10"
@@ -17,6 +17,9 @@ keywords = [
     "image-processing",
     "captioning",
     "florence-2",
+    "synthetic",
+    "image-editing",
+    "diffusers",
     "machine-learning",
     "cli",
 ]
@@ -45,6 +48,16 @@ caption = [
     "einops",
     "timm",
 ]
+synthetic = [
+    "torch>=2.0.0",
+    "diffusers>=0.32.0",
+    "transformers>=4.38.0",
+    "accelerate",
+    "sentencepiece",
+]
+all = [
+    "datasety[caption,synthetic]",
+]
 dev = [
     "pytest>=7.0.0",
     "ruff>=0.1.0",

{datasety-0.1.0 → datasety-0.2.0}/src/datasety/__init__.py

@@ -1,3 +1,3 @@
 """datasety - CLI tool for dataset preparation: image resizing and captioning."""
 
-__version__ = "0.1.0"
+__version__ = "0.2.0"

{datasety-0.1.0 → datasety-0.2.0}/src/datasety/cli.py

@@ -1,10 +1,11 @@
 #!/usr/bin/env python3
 """
-datasety - CLI tool for dataset preparation: image resizing and captioning.
+datasety - CLI tool for dataset preparation: resize, caption, and synthetic generation.
 
 Usage:
     datasety resize --input ./in --output ./out --resolution 768x1024 --crop-position top
     datasety caption --input ./in --output ./out --trigger-word "[trigger]" --florence-2-large
+    datasety synthetic --input ./in --output ./out --prompt "add a winter hat"
 """
 
 import argparse
@@ -283,6 +284,114 @@ def cmd_caption(args):
     print(f"Done! Processed: {processed} images")
 
 
+def cmd_synthetic(args):
+    """Execute the synthetic image generation command."""
+    # Lazy import for faster CLI startup
+    try:
+        import torch
+    except ImportError:
+        print("Error: PyTorch not installed.")
+        print("Run: pip install 'datasety[synthetic]'")
+        sys.exit(1)
+
+    input_dir = Path(args.input)
+    output_dir = Path(args.output)
+
+    if not input_dir.exists():
+        print(f"Error: Input directory '{input_dir}' does not exist.")
+        sys.exit(1)
+
+    output_dir.mkdir(parents=True, exist_ok=True)
+
+    # Determine device
+    if args.device == "cuda" and not torch.cuda.is_available():
+        print("Warning: CUDA not available, falling back to CPU")
+        device = "cpu"
+    else:
+        device = args.device
+
+    # Import the correct pipeline based on model
+    try:
+        from diffusers import QwenImageEditPlusPipeline
+        pipeline_class = QwenImageEditPlusPipeline
+    except ImportError:
+        print("Error: QwenImageEditPlusPipeline not found.")
+        print("Make sure you have the latest diffusers: pip install -U diffusers")
+        sys.exit(1)
+
+    print(f"Loading model: {args.model}")
+    print(f"Device: {device}")
+
+    torch_dtype = torch.bfloat16 if device == "cuda" else torch.float32
+
+    try:
+        pipeline = pipeline_class.from_pretrained(
+            args.model,
+            torch_dtype=torch_dtype
+        )
+        pipeline.to(device)
+        pipeline.set_progress_bar_config(disable=False)
+    except Exception as e:
+        print(f"Error loading model: {e}")
+        sys.exit(1)
+
+    # Find images
+    formats = ["jpg", "jpeg", "png", "webp", "bmp", "tiff"]
+    image_files = get_image_files(input_dir, formats)
+
+    if not image_files:
+        print(f"No images found in '{input_dir}'")
+        sys.exit(0)
+
+    print(f"Found {len(image_files)} images")
+    print(f"Prompt: {args.prompt}")
+    print(f"Steps: {args.steps}, CFG: {args.cfg_scale}, True CFG: {args.true_cfg_scale}")
+    print("-" * 50)
+
+    processed = 0
+
+    for img_path in image_files:
+        try:
+            image = Image.open(img_path).convert("RGB")
+
+            # Set up generation parameters
+            gen_kwargs = {
+                "image": [image],
+                "prompt": args.prompt,
+                "negative_prompt": args.negative_prompt,
+                "num_inference_steps": args.steps,
+                "guidance_scale": args.cfg_scale,
+                "true_cfg_scale": args.true_cfg_scale,
+                "num_images_per_prompt": args.num_images,
+            }
+
+            # Add seed if specified
+            if args.seed is not None:
+                gen_kwargs["generator"] = torch.manual_seed(args.seed)
+
+            with torch.inference_mode():
+                output = pipeline(**gen_kwargs)
+
+            # Save output image(s)
+            for idx, out_img in enumerate(output.images):
+                if args.num_images > 1:
+                    out_name = f"{img_path.stem}_{idx + 1}.png"
+                else:
+                    out_name = f"{img_path.stem}.png"
+
+                out_path = output_dir / out_name
+                out_img.save(out_path)
+
+            print(f"[OK] {img_path.name} -> {len(output.images)} image(s)")
+            processed += 1
+
+        except Exception as e:
+            print(f"[ERROR] {img_path.name}: {e}")
+
+    print("-" * 50)
+    print(f"Done! Processed: {processed} images")
+
+
 def main():
     parser = argparse.ArgumentParser(
         prog="datasety",
@@ -379,6 +488,74 @@ def main():
     )
     caption_parser.set_defaults(func=cmd_caption)
 
+    # === SYNTHETIC command ===
+    synthetic_parser = subparsers.add_parser(
+        "synthetic",
+        help="Generate synthetic images using image editing models"
+    )
+    synthetic_parser.add_argument(
+        "--input", "-i",
+        required=True,
+        help="Input directory containing images"
+    )
+    synthetic_parser.add_argument(
+        "--output", "-o",
+        required=True,
+        help="Output directory for generated images"
+    )
+    synthetic_parser.add_argument(
+        "--prompt", "-p",
+        required=True,
+        help="Edit prompt (e.g., 'add a winter hat to the person')"
+    )
+    synthetic_parser.add_argument(
+        "--model",
+        default="Qwen/Qwen-Image-Edit-2511",
+        help="Model to use (default: Qwen/Qwen-Image-Edit-2511)"
+    )
+    synthetic_parser.add_argument(
+        "--device",
+        choices=["cpu", "cuda"],
+        default="cuda",
+        help="Device to run model on (default: cuda)"
+    )
+    synthetic_parser.add_argument(
+        "--steps",
+        type=int,
+        default=40,
+        help="Number of inference steps (default: 40)"
+    )
+    synthetic_parser.add_argument(
+        "--cfg-scale",
+        type=float,
+        default=1.0,
+        help="Guidance scale (default: 1.0)"
+    )
+    synthetic_parser.add_argument(
+        "--true-cfg-scale",
+        type=float,
+        default=4.0,
+        help="True CFG scale (default: 4.0)"
+    )
+    synthetic_parser.add_argument(
+        "--negative-prompt",
+        default=" ",
+        help="Negative prompt (default: ' ')"
+    )
+    synthetic_parser.add_argument(
+        "--num-images",
+        type=int,
+        default=1,
+        help="Number of images to generate per input (default: 1)"
+    )
+    synthetic_parser.add_argument(
+        "--seed",
+        type=int,
+        default=None,
+        help="Random seed for reproducibility"
+    )
+    synthetic_parser.set_defaults(func=cmd_synthetic)
+
     # Parse and execute
     args = parser.parse_args()
     args.func(args)

datasety-0.1.0/.gitignore

@@ -1,36 +0,0 @@
-# Byte-compiled / optimized / DLL files
-__pycache__/
-*.py[cod]
-*$py.class
-
-# Distribution / packaging
-build/
-dist/
-*.egg-info/
-*.egg
-
-# Virtual environments
-venv/
-.venv/
-env/
-
-# IDE
-.idea/
-.vscode/
-*.swp
-*.swo
-
-# Testing
-.pytest_cache/
-.coverage
-htmlcov/
-
-# OS
-.DS_Store
-Thumbs.db
-
-# Project specific
-*.jpg
-*.jpeg
-*.png
-*.webp

datasety-0.1.0/README.md

@@ -1,129 +0,0 @@
-# datasety
-
-CLI tool for dataset preparation: image resizing and captioning with Florence-2.
-
-## Installation
-
-```bash
-pip install datasety
-```
-
-For captioning support (requires PyTorch and Transformers):
-
-```bash
-pip install datasety[caption]
-```
-
-## Usage
-
-### Resize Images
-
-Resize and crop images to a target resolution:
-
-```bash
-datasety resize --input ./images --output ./resized --resolution 768x1024
-```
-
-**Options:**
-
-| Option                  | Description                                               | Default             |
-| ----------------------- | --------------------------------------------------------- | ------------------- |
-| `--input`, `-i`         | Input directory                                           | (required)          |
-| `--output`, `-o`        | Output directory                                          | (required)          |
-| `--resolution`, `-r`    | Target resolution (WIDTHxHEIGHT)                          | (required)          |
-| `--crop-position`       | Crop position: `top`, `center`, `bottom`, `left`, `right` | `center`            |
-| `--input-format`        | Comma-separated formats                                   | `jpg,jpeg,png,webp` |
-| `--output-format`       | Output format: `jpg`, `png`, `webp`                       | `jpg`               |
-| `--output-name-numbers` | Rename files to 1.jpg, 2.jpg, ...                         | `false`             |
-
-**Example:**
-
-```bash
-datasety resize \
-    --input ./raw_photos \
-    --output ./dataset \
-    --resolution 1024x1024 \
-    --crop-position top \
-    --output-format jpg \
-    --output-name-numbers
-```
-
-**How it works:**
-
-1. Finds all images matching input formats
-2. Skips images where either dimension is smaller than target
-3. Resizes proportionally so the smaller side matches target
-4. Crops from the specified area to exact dimensions
-5. Saves with high quality (95% for jpg/webp)
-
-### Generate Captions
-
-Generate captions for images using Microsoft's Florence-2 model:
-
-```bash
-datasety caption --input ./images --output ./captions --florence-2-large
-```
-
-**Options:**
-
-| Option               | Description                     | Default                   |
-| -------------------- | ------------------------------- | ------------------------- |
-| `--input`, `-i`      | Input directory                 | (required)                |
-| `--output`, `-o`     | Output directory for .txt files | (required)                |
-| `--device`           | `cpu` or `cuda`                 | `cpu`                     |
-| `--trigger-word`     | Text to prepend to captions     | (none)                    |
-| `--prompt`           | Florence-2 task prompt          | `<MORE_DETAILED_CAPTION>` |
-| `--florence-2-base`  | Use base model (0.23B, faster)  |                           |
-| `--florence-2-large` | Use large model (0.77B, better) | (default)                 |
-
-**Available prompts:**
-
-- `<CAPTION>` - Brief caption
-- `<DETAILED_CAPTION>` - Detailed caption
-- `<MORE_DETAILED_CAPTION>` - Most detailed caption (default)
-
-**Example:**
-
-```bash
-datasety caption \
-    --input ./dataset \
-    --output ./dataset \
-    --device cuda \
-    --trigger-word "photo of sks person," \
-    --florence-2-large
-```
-
-This creates a `.txt` file for each image with the generated caption.
-
-## Common Workflows
-
-### Prepare a LoRA Training Dataset
-
-```bash
-# 1. Resize images to 1024x1024
-datasety resize -i ./raw -o ./dataset -r 1024x1024 --crop-position center
-
-# 2. Generate captions with trigger word
-datasety caption -i ./dataset -o ./dataset --trigger-word "[trigger]" --device cuda
-```
-
-### Batch Process with Numbered Files
-
-```bash
-datasety resize \
-    -i ./photos \
-    -o ./processed \
-    -r 768x1024 \
-    --output-name-numbers \
-    --crop-position top
-```
-
-## Requirements
-
-- Python 3.10+
-- Pillow (for resize)
-- PyTorch + Transformers (for caption, install with `pip install datasety[caption]`)
-
-## License
-
-MIT