Feature extraction with Inference API in Javascript

Hi everyone, I need to use a model for feature extraction through the Inference API in Javascript.

Using the examples described in Parameters, by using this code:

async function query(data) {
    const response = await fetch(
        "https://api-inference.huggingface.co/models/sentence-transformers/all-mpnet-base-v2",
        {
            method: "POST",
            headers: {
                Authorization: `Bearer hf_XXXXXXXXXXXXXXXXXX`,
                "Content-Type": "application/json"
            },
            body: JSON.stringify(data),
        }
    );
    const result = await response.json();
    return result;
}
query({
    inputs: "Can you please let us know more details about your "
}).then((response) => {
    console.log(JSON.stringify(response, null, 2));
});

I’m getting the error

{
  "error": [
    "Input should be a valid dictionary or instance of SentenceSimilarityInputsCheck: received `Can you please let us know more details about your ` in `parameters`"
  ]
}

Thanks in advance!

1 Like

Maybe the error is more troublesome than I thought. Maybe something is wrong with the model repo settings. Try another model if you can.

Thanks @John6666!!! I’ve read it’s due to the pipeline_tag property of the model. It seems it must have the value ‘feature-extraction’.

This model worked to me:

async function query(data) {
	const response = await fetch(
		"https://api-inference.huggingface.co/models/facebook/bart-base",
		{
			headers: {
				Authorization: "Bearer hf_XXXXXXXXXXXX",
				"Content-Type": "application/json",
			},
			method: "POST",
			body: JSON.stringify(data),
		}
	);
	const result = await response.json();
	return result;
}

query({"inputs": "Today is a sunny day and I will get some ice cream."}).then((response) => {
	console.log(JSON.stringify(response));
});
1 Like

Wow. Was it a hit?
It’s going to take a while for us to report back to the model authors and get them to fix it.
It’s faster if you have a usable model.

1 Like

Let me know their answer!! Thanks again!

1 Like

Oh, no, I’m not reporting it because it seems to have worked. It’s hard when it doesn’t work …