Model Card for tmp-pl-image-classification
์ด ์ ์ฅ์๋ ๐ค Transformers์ pipeline() ๋์์ ์ดํดํ๊ณ ์ฐ์ตํ๊ธฐ ์ํ ํ์ต์ฉ(pipeline practice) ๋ชจ๋ธ repo ์
๋๋ค.
๋ชจ๋ธ ๊ฐ์ค์น๋ ์๋ณธ ๋ชจ๋ธ google/vit-base-patch16-224 ์ ๊ทธ๋๋ก ์ฌ์ฉํ๋ฉฐ, ์ถ๊ฐ์ ์ธ fine-tuning์ ์ํํ์ง ์์์ต๋๋ค.
Model Details
Model Description
๋ณธ ๋ชจ๋ธ์ Vision Transformer(ViT) ๊ธฐ๋ฐ ์ด๋ฏธ์ง ๋ถ๋ฅ ๋ชจ๋ธ์ pipeline("image-classification") ํํ๋ก
Hub์ ์
๋ก๋ํ๊ณ ๋ค์ ๋ถ๋ฌ์ค๋ ์ ์ฒด ํ๋ฆ์ ์ค์ตํ๊ธฐ ์ํด ๊ตฌ์ฑ๋จ.
- Developed by: Google Research (์๋ณธ๋ชจ๋ธ)
- Shared by [optional]: dsaint31
- Model type: Image Classification (Vision Transformer)
- Language(s) (NLP): ํด๋น ์์ (์ด๋ฏธ์ง ์ ๋ ฅ)
- License: Apache-2.0
- Finetuned from model [optional]: google/vit-base-patch16-224 (๊ฐ์ค์น ๋ณ๊ฒฝ ์์. fine-tuning ๋ฏธ์ํ)
Model Sources [optional]
- Baee model Repository: https://huggingface.co/google/vit-base-patch16-224
- Paper [optional]: Dosovitskiy et al., An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale, arXiv:2010.11929
- Demo [optional]: None
Uses
Direct Use
pipeline("image-classification", model=...)์ฌ์ฉ๋ฒ ์ค์ต- Hugging Face Hub์ pipeline ํํ๋ก ๋ชจ๋ธ์ ์ ๋ก๋ / ๋ค์ด๋ก๋ํ๋ ํ๋ฆ ์ดํด
- Vision ๋ชจ๋ธ๊ณผ pipeline์ ๊ด๊ณ ํ์ต
Downstream Use [optional]
- ๋ณธ repo ์์ฒด๋ downstream task๋ฅผ ์ํ fine-tuning์ ๋ชฉ์ ์ผ๋ก ํ์ง ์์ต๋๋ค.
- ํ์ต ๋๋ ์ฑ๋ฅ ๋น๊ต ๋ชฉ์ ์ด๋ผ๋ฉด ์๋ณธ ๋ชจ๋ธ repo๋ฅผ ์ง์ ์ฌ์ฉํ๋ ๊ฒ์ด ์ ์ ํฉ๋๋ค.
Out-of-Scope Use
- ๋ชจ๋ธ ์ฑ๋ฅ ํ๊ฐ ๋๋ ๋ฒค์น๋งํฌ
- ์ค์ ์๋น์ค ํ๊ฒฝ์์์ ๋ชจ๋ธ ๋ฐฐํฌ
- ํน์ ๋๋ฉ์ธ(์๋ฃ, ์ฐ์ ์์ ๋ฑ)์ ๋ํ ์ ๋ขฐ์ฑ ์๋ ์ถ๋ก
Bias, Risks, and Limitations
- ๋ณธ ๋ชจ๋ธ์ ImageNet ๊ธฐ๋ฐ ๋ฐ์ดํฐ๋ก ํ์ต๋ ์ผ๋ฐ ๋ชฉ์ ์ด๋ฏธ์ง ๋ถ๋ฅ ๋ชจ๋ธ์ ํน์ฑ์ ๊ทธ๋๋ก ๊ฐ์ง๋๋ค.
- ํน์ ๊ฐ์ฒด, ๋ฌธํ์ ๋งฅ๋ฝ, ์ ๋ฌธ ๋๋ฉ์ธ์ ๋ํ ๋ถ๋ฅ ์ฑ๋ฅ์ ๋ณด์ฅ๋์ง ์์ต๋๋ค.
- ๋ณธ repo๋ ์ฐ์ต์ฉ pipeline ์ ์ฅ์์ด๋ฏ๋ก ๋ชจ๋ธ์ ์ฌํ์ ์ํฅ์ด๋ ํธํฅ ๋ถ์์ ๋ชฉ์ ์ผ๋ก ํ์ง ์์ต๋๋ค.
Recommendations
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
- ์ค์ ์ฌ์ฉ ๋ชฉ์ ์ด ์๋ ๊ฒฝ์ฐ, ์๋ณธ ๋ชจ๋ธ ์นด๋(
google/vit-base-patch16-224)์ ์ ํ ์ฌํญ์ ๋ฐ๋์ ์ฐธ๊ณ ํ์ญ์์ค. - ์ด repo๋ ํ์ต ๋ฐ ์ค์ต ๋ชฉ์ ์ ํํด ์ฌ์ฉํ๊ธฐ๋ฅผ ๊ถ์ฅํฉ๋๋ค.
How to Get Started with the Model
Use the code below to get started with the model.
์๋ ์์ ๋ Hugging Face pipeline์ ์ด์ฉํด ๋ณธ ๋ชจ๋ธ์ ๋ก๋ํ๊ณ ์ด๋ฏธ์ง๋ฅผ ๋ถ๋ฅํ๋ ์ต์ ์์ ์
๋๋ค.
from transformers import pipeline
from PIL import Image
import requests
img_url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/cats.png"
image = Image.open(requests.get(img_url, stream=True).raw)
clf = pipeline(
task="image-classification",
model="dsaint31/tmp-pl-image-classification",
)
print(clf(image))
[More Information Needed]
Training Details
Training Data
- ๋ณธ repo์์๋ ์ถ๊ฐ ํ์ต์ ์ํํ์ง ์์์ต๋๋ค.
- ์๋ณธ ๋ชจ๋ธ์ ImageNet-21k๋ก ์ฌ์ ํ์ต(pretraining) ํ ImageNet-1k๋ก fine-tuning๋ ๋ชจ๋ธ์ ๋๋ค.
Training Procedure
Preprocessing [optional]
- ์๋ณธ ViT ๋ชจ๋ธ์ ๊ธฐ๋ณธ ์ด๋ฏธ์ง ์ ์ฒ๋ฆฌ(Image Processor)๋ฅผ ๊ทธ๋๋ก ์ฌ์ฉํฉ๋๋ค.
[More Information Needed]
Training Hyperparameters
- Training regime: ํด๋น ์์(ํ์ต ๋ฏธ์ํ) [More Information Needed]
Speeds, Sizes, Times [optional]
Evaluation
Testing Data, Factors & Metrics
Testing Data
- ๋ณธ repo์์๋ ๋ณ๋์ ํ๊ฐ๋ฅผ ์ํํ์ง ์์์ต๋๋ค.
- ์ฑ๋ฅ ์งํ๋ ์๋ณธ ๋ชจ๋ธ ์นด๋์ ํ๊ฐ ๊ฒฐ๊ณผ๋ฅผ ์ฐธ๊ณ ํ์ญ์์ค.
[More Information Needed]
Factors
[More Information Needed]
Metrics
[More Information Needed]
Results
[More Information Needed]
Summary
Model Examination [optional]
[More Information Needed]
Environmental Impact
๋ณธ repo์์๋ ํ์ต์ ์ํํ์ง ์์์ผ๋ฏ๋ก ์ถ๊ฐ์ ์ธ ํ๊ฒฝ์ ์ํฅ์ ์์ต๋๋ค.
- ์๋ณธ ๋ชจ๋ธ ํ์ต์ ๋ํ ํ๊ฒฝ ์ํฅ์ base model ๋ฌธ์๋ฅผ ์ฐธ๊ณ ํ์ญ์์ค.
Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).
- Hardware Type: [More Information Needed]
- Hours used: [More Information Needed]
- Cloud Provider: [More Information Needed]
- Compute Region: [More Information Needed]
- Carbon Emitted: [More Information Needed]
Technical Specifications [optional]
Model Architecture and Objective
- Vision Transformer (ViT-Base, patch size 16, input resolution 224x224)
- Objective: Image classification
Compute Infrastructure
[More Information Needed]
Hardware
- ํด๋น์์ (ํ์ต ๋ฏธ์ํ)
Software
- Transformers
- Pillow
- PyTorch
[More Information Needed]
Citation [optional]
์๋ณธ ๋ชจ๋ธ ์ธ์ฉ ์ ์๋ ๋ ผ๋ฌธ์ ์ฐธ๊ณ ํ์ญ์์ค.
BibTeX:
@article{dosovitskiy2020image,
title={An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale},
author={Dosovitskiy, Alexey and others},
journal={arXiv preprint arXiv:2010.11929},
year={2020}
}
APA:
[More Information Needed]
Glossary [optional]
[More Information Needed]
More Information [optional]
[More Information Needed]
Model Card Authors [optional]
- dsaint31 (pipeline practice repository)
Model Card Contact
- Hugging Face profile: https://huggingface.co/dsaint31
- Downloads last month
- 28
Model tree for dsaint31/tmp-pl-image-classification
Base model
google/vit-base-patch16-224