Metadata-Version: 2.1
Name: studiolab-ml
Version: 0.1.1
Summary: STUDIOLAB ML inference Package
Home-page: https://github.com/StudioLABdev/ML-inference-package
Author: STUDIOLAB
Author-email: jaeseung.lim@studiolab.ai
License: GPL-3.0
Project-URL: Bug Reports, https://github.com/StudioLABdev/ML-inference-package/issues
Project-URL: Source, https://github.com/StudioLABdev/ML-inference-package
Keywords: STUDIOLAB,MLFT,PCP,PoseCompo,inference,ML,DL
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
Classifier: Programming Language :: Python :: 3.9
Classifier: Topic :: Software Development
Classifier: Topic :: Scientific/Engineering
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Classifier: Topic :: Scientific/Engineering :: Image Recognition
Classifier: Operating System :: POSIX :: Linux
Requires-Python: >=3.9
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: onnx (==1.15.0)
Requires-Dist: onnxsim (==0.4.35)
Requires-Dist: onnxruntime-gpu (==1.16.1)
Requires-Dist: Pillow (==10.0.1)
Requires-Dist: numpy (==1.25.2)
Requires-Dist: openai (==0.28.0)
Requires-Dist: tiktoken (==0.5.1)
Requires-Dist: boto3 (==1.28.75)
Requires-Dist: sentry-sdk

# STUDIOLAB ML inference Package

# Install
- pip install studiolab-ml
# RUN
All input image type is PIL Image 
## MLFT
 ```
from studiolab_ml import MLFT

mlft = MLFT()
out = mlft.predict(img, cat_id)
 ```
- result is same dict type as "get_attributes" in ML-API
## Pose Compo
```
from studiolab_ml import PoseCompo

pcp = PoseCompo()
out = pcp.predict(img)
```
- output examples
 - outfit image - {'cut': 'outfit', 'background': 'blind', 'direction': 'front', 'head': 'head', 'part': 'full', 'pose': 'stand', 'detail': None}
 - product image - {'cut': 'product', 'background': None, 'direction': 'front', 'head': None, 'part': None, 'pose': None, 'detail': None}
 - detail image - {'cut': 'detail', 'background': None, 'direction': None, 'head': None, 'part': None, 'pose': None, 'detail': [shoulder, sleeve, ..]}
 - noise image - {'cut': 'noise', 'background': None, 'direction': None, 'head': None, 'part': None, 'pose': None, 'detail': None}
## FIC
```
from studiolab_ml import PoseCompo

infer = FIC(api_key)
res = infer(attribute_dict, user_inputs_dict)
```
- input and result is same dict type as "get_gpt_content" in ML-API
# TODO
- create model cloud storage
- model download from cloud
- GPU inference
