Fork of openai/clip-vit-base-patch32 for a zero-shot-image-classification Inference endpoint.

This repository implements a custom task for zero-shot-image-classification for ? Inference Endpoints. The code for the customized pipeline is in the pipeline.py .

To use deploy this model a an Inference Endpoint you have to select Custom as task to use the pipeline.py file. -> double check if it is selected

expected Request payload

{
  "image": "/9j/4AAQSkZJRgABAQEBLAEsAAD/2wBDAAMCAgICAgMC....", // base64 image as bytes
  "candiates":["sea","palace","car","ship"]
}

below is an example on how to run a request using Python and requests .

Run Request

!wget https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg

import json
from typing import List
import requests as r
import base64

ENDPOINT_URL = ""
HF_TOKEN = ""


def predict(path_to_image: str = None, candiates: List[str] = None):
    with open(path_to_image, "rb") as i:
        b64 = base64.b64encode(i.read())

    payload = {"inputs": {"image": b64.decode("utf-8"), "candiates": candiates}}
    response = r.post(
        ENDPOINT_URL, headers={"Authorization": f"Bearer {HF_TOKEN}"}, json=payload
    )
    return response.json()


prediction = predict(
    path_to_image="palace.jpg", candiates=["sea", "palace", "car", "ship"]
)

expected output

[{'label': 'palace', 'score': 0.9996134638786316},
 {'label': 'car', 'score': 0.0002602009626571089},
 {'label': 'ship', 'score': 0.00011758189066313207},
 {'label': 'sea', 'score': 8.666840585647151e-06}]

openai/clip-vit-base-patch32 的零射击图像分类推理端点的分支。

该仓库实现了?推理端点的零射击图像分类任务的自定义任务。定制的 pipeline 代码位于 pipeline.py 中。

要使用部署此模型作为推理端点，您必须选择 Custom 作为任务，并使用 pipeline.py 文件。-> 请确认是否已选择。

期望的请求载荷

{
  "image": "/9j/4AAQSkZJRgABAQEBLAEsAAD/2wBDAAMCAgICAgMC....", // base64 image as bytes
  "candiates":["sea","palace","car","ship"]
}

以下是使用 Python 和 requests 运行请求的示例。

运行请求

!wget https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg

import json
from typing import List
import requests as r
import base64

ENDPOINT_URL = ""
HF_TOKEN = ""


def predict(path_to_image: str = None, candiates: List[str] = None):
    with open(path_to_image, "rb") as i:
        b64 = base64.b64encode(i.read())

    payload = {"inputs": {"image": b64.decode("utf-8"), "candiates": candiates}}
    response = r.post(
        ENDPOINT_URL, headers={"Authorization": f"Bearer {HF_TOKEN}"}, json=payload
    )
    return response.json()


prediction = predict(
    path_to_image="palace.jpg", candiates=["sea", "palace", "car", "ship"]
)

预期的输出

[{'label': 'palace', 'score': 0.9996134638786316},
 {'label': 'car', 'score': 0.0002602009626571089},
 {'label': 'ship', 'score': 0.00011758189066313207},
 {'label': 'sea', 'score': 8.666840585647151e-06}]