|
--- |
|
license: mit |
|
tags: |
|
- donut |
|
- image-to-text |
|
- vision |
|
- endpoints-template |
|
--- |
|
|
|
# Fork of [naver-clova-ix/donut-base-finetuned-cord-v2](https://huggingface.co/naver-clova-ix/donut-base-finetuned-cord-v2) |
|
|
|
> This is fork of [naver-clova-ix/donut-base-finetuned-cord-v2](https://huggingface.co/naver-clova-ix/donut-base-finetuned-cord-v2) implementing a custom `handler.py` as an example for how to use `donut` models with [inference-endpoints](https://hf.co/inference-endpoints) |
|
|
|
--- |
|
|
|
# Donut (base-sized model, fine-tuned on CORD) |
|
|
|
Donut model fine-tuned on CORD. It was introduced in the paper [OCR-free Document Understanding Transformer](https://arxiv.org/abs/2111.15664) by Geewok et al. and first released in [this repository](https://github.com/clovaai/donut). |
|
|
|
Donut consists of a vision encoder (Swin Transformer) and a text decoder (BART). Given an image, the encoder first encodes the image into a tensor of embeddings (of shape batch_size, seq_len, hidden_size), after which the decoder autoregressively generates text, conditioned on the encoding of the encoder. |
|
|
|
# Use with Inference Endpoints |
|
|
|
Hugging Face Inference endpoints can directly work with binary data, this means that we can directly send our image from our document to the endpoint. We are going to use requests to send our requests. (make your you have it installed `pip install requests`) |
|
|
|
 |
|
|
|
## Send requests with Pyton |
|
|
|
load sample image |
|
|
|
```bash |
|
wget https://huggingface.co/philschmid/donut-base-finetuned-cord-v2/resolve/main/sample.png |
|
``` |
|
|
|
send request to endpoint |
|
|
|
```python |
|
import json |
|
import requests as r |
|
import mimetypes |
|
|
|
ENDPOINT_URL="" # url of your endpoint |
|
HF_TOKEN="" # organization token where you deployed your endpoint |
|
|
|
def predict(path_to_image:str=None): |
|
with open(path_to_image, "rb") as i: |
|
b = i.read() |
|
headers= { |
|
"Authorization": f"Bearer {HF_TOKEN}", |
|
"Content-Type": mimetypes.guess_type(path_to_image)[0] |
|
} |
|
response = r.post(ENDPOINT_URL, headers=headers, data=b) |
|
return response.json() |
|
|
|
prediction = predict(path_to_image="sample.png") |
|
|
|
print(prediction) |
|
# {'menu': [{'nm': '0571-1854 BLUS WANITA', |
|
# 'unitprice': '@120.000', |
|
# 'cnt': '1', |
|
# 'price': '120,000'}, |
|
# {'nm': '1002-0060 SHOPPING BAG', 'cnt': '1', 'price': '0'}], |
|
# 'total': {'total_price': '120,000', |
|
# 'changeprice': '0', |
|
# 'creditcardprice': '120,000', |
|
# 'menuqty_cnt': '1'}} |
|
``` |
|
|
|
|
|
|
|
**curl example** |
|
|
|
```bash |
|
curl https://ak7gduay2ypyr9vp.us-east-1.aws.endpoints.huggingface.cloud \ |
|
-X POST \ |
|
--data-binary 'sample.png' \ |
|
-H "Authorization: Bearer XXX" \ |
|
-H "Content-Type: null" |
|
``` |
