Spaces:
Running
on
Zero
Running
on
Zero
Fixes version...
Browse files- README.md +5 -4
- labelizer/__init__.py +2 -2
README.md
CHANGED
|
@@ -8,8 +8,7 @@ sdk_version: 6.0.2
|
|
| 8 |
app_file: app.py
|
| 9 |
pinned: false
|
| 10 |
license: mit
|
| 11 |
-
python_version: "3.
|
| 12 |
-
duplicated_from: metal3d/labelm2
|
| 13 |
---
|
| 14 |
|
| 15 |
# ๐ผ๏ธ Labelizer - AI Image Labeling Tool
|
|
@@ -27,7 +26,7 @@ An intelligent image labeling tool that uses Florence-2 vision-language model to
|
|
| 27 |
## ๐ How to Use
|
| 28 |
|
| 29 |
1. **Upload Images**: Click "๐ Upload images" to select multiple image files
|
| 30 |
-
2. **Generate Labels**:
|
| 31 |
- Click "โจ Generate label" below individual images
|
| 32 |
- Or click "๐ท๏ธ Labelize all images" for batch processing
|
| 33 |
3. **Review & Edit**: Modify any generated labels as needed
|
|
@@ -43,6 +42,7 @@ An intelligent image labeling tool that uses Florence-2 vision-language model to
|
|
| 43 |
## ๐ Supported Tasks
|
| 44 |
|
| 45 |
The tool supports various captioning tasks:
|
|
|
|
| 46 |
- `<MORE_DETAILED_CAPTION>`: Comprehensive image descriptions
|
| 47 |
- `<DETAILED_CAPTION>`: Detailed but concise descriptions
|
| 48 |
- `<CAPTION>`: Basic image captions
|
|
@@ -62,4 +62,5 @@ The tool supports various captioning tasks:
|
|
| 62 |
|
| 63 |
---
|
| 64 |
|
| 65 |
-
Built with โค๏ธ using [Gradio](https://gradio.app/) and [Florence-2](https://huggingface.co/microsoft/Florence-2-large)
|
|
|
|
|
|
| 8 |
app_file: app.py
|
| 9 |
pinned: false
|
| 10 |
license: mit
|
| 11 |
+
python_version: "3.9"
|
|
|
|
| 12 |
---
|
| 13 |
|
| 14 |
# ๐ผ๏ธ Labelizer - AI Image Labeling Tool
|
|
|
|
| 26 |
## ๐ How to Use
|
| 27 |
|
| 28 |
1. **Upload Images**: Click "๐ Upload images" to select multiple image files
|
| 29 |
+
2. **Generate Labels**:
|
| 30 |
- Click "โจ Generate label" below individual images
|
| 31 |
- Or click "๐ท๏ธ Labelize all images" for batch processing
|
| 32 |
3. **Review & Edit**: Modify any generated labels as needed
|
|
|
|
| 42 |
## ๐ Supported Tasks
|
| 43 |
|
| 44 |
The tool supports various captioning tasks:
|
| 45 |
+
|
| 46 |
- `<MORE_DETAILED_CAPTION>`: Comprehensive image descriptions
|
| 47 |
- `<DETAILED_CAPTION>`: Detailed but concise descriptions
|
| 48 |
- `<CAPTION>`: Basic image captions
|
|
|
|
| 62 |
|
| 63 |
---
|
| 64 |
|
| 65 |
+
Built with โค๏ธ using [Gradio](https://gradio.app/) and [Florence-2](https://huggingface.co/microsoft/Florence-2-large)
|
| 66 |
+
|
labelizer/__init__.py
CHANGED
|
@@ -21,10 +21,10 @@ def _load_model():
|
|
| 21 |
|
| 22 |
print(f"Loading model {MODEL_ID} on {_device} with dtype {_torch_dtype}...")
|
| 23 |
_model = Florence2ForConditionalGeneration.from_pretrained(
|
| 24 |
-
MODEL_ID,
|
| 25 |
).to(_device) # type: ignore
|
| 26 |
_processor = Florence2Processor.from_pretrained(
|
| 27 |
-
MODEL_ID, trust_remote_code=True
|
| 28 |
)
|
| 29 |
print("Model loaded successfully!")
|
| 30 |
|
|
|
|
| 21 |
|
| 22 |
print(f"Loading model {MODEL_ID} on {_device} with dtype {_torch_dtype}...")
|
| 23 |
_model = Florence2ForConditionalGeneration.from_pretrained(
|
| 24 |
+
MODEL_ID, dtype=_torch_dtype, trust_remote_code=True
|
| 25 |
).to(_device) # type: ignore
|
| 26 |
_processor = Florence2Processor.from_pretrained(
|
| 27 |
+
MODEL_ID, trust_remote_code=True, use_fast=True
|
| 28 |
)
|
| 29 |
print("Model loaded successfully!")
|
| 30 |
|