|
--- |
|
language: ja |
|
thumbnail: https://github.com/rinnakk/japanese-pretrained-models/blob/master/rinna.png |
|
license: apache-2.0 |
|
tags: |
|
- feature-extraction |
|
- ja |
|
- japanese |
|
- clip |
|
- cloob |
|
- vision |
|
--- |
|
|
|
# rinna/japanese-cloob-vit-b-16 |
|
|
|
 |
|
|
|
This repository provides a Japanese [CLOOB (Contrastive Leave One Out Boost)](https://arxiv.org/abs/2110.11316) model. The model was trained by [rinna Co., Ltd.](https://corp.rinna.co.jp/) |
|
|
|
# How to use the model |
|
|
|
|
|
1. Install package |
|
|
|
```shell |
|
$ pip install git+https://github.com/rinnakk/japanese-clip.git |
|
``` |
|
|
|
2. Run |
|
|
|
```python |
|
import io |
|
import requests |
|
from PIL import Image |
|
import torch |
|
import japanese_clip as ja_clip |
|
|
|
device = "cuda" if torch.cuda.is_available() else "cpu" |
|
|
|
|
|
model, preprocess = ja_clip.load("rinna/japanese-cloob-vit-b-16", device=device) |
|
tokenizer = ja_clip.load_tokenizer() |
|
|
|
img = Image.open(io.BytesIO(requests.get('https://images.pexels.com/photos/2253275/pexels-photo-2253275.jpeg?auto=compress&cs=tinysrgb&dpr=3&h=750&w=1260').content)) |
|
image = preprocess(img).unsqueeze(0).to(device) |
|
encodings = ja_clip.tokenize( |
|
texts=["犬", "猫", "象"], |
|
max_seq_len=77, |
|
device=device, |
|
tokenizer=tokenizer, # this is optional. if you don't pass, load tokenizer each time |
|
) |
|
|
|
with torch.no_grad(): |
|
image_features = model.get_image_features(image) |
|
text_features = model.get_text_features(**encodings) |
|
|
|
text_probs = (100.0 * image_features @ text_features.T).softmax(dim=-1) |
|
|
|
print("Label probs:", text_probs) # prints: [[1.0, 0.0, 0.0]] |
|
``` |
|
# Model architecture |
|
The model was trained a ViT-B/16 Transformer architecture as an image encoder and uses a 12-layer RoBERTa as a text encoder. The text encoder was trained upon the pre-trained Japanese RoBERTa model [rinna/japanese-roberta-base](https://huggingface.co/rinna/japanese-roberta-base) with the same sentencepiece tokenizer. |
|
|
|
# Training |
|
The model was trained on [CC12M](https://github.com/google-research-datasets/conceptual-12m) translated the captions to Japanese. |
|
|
|
|
|
# License |
|
|
|
[Apache-2.0 license](https://www.apache.org/licenses/LICENSE-2.0) |
