Getting Not Found for model google/flan-t5-small (and others)

Beginners
1

Hi everyone,

I am new to HF and I’m trying to use the Inference API to call the model google/flan-t5-small from the Hugging Face Hub.
I have a PRO subscription and I’m using a valid access token with Read permission.
Here’s the exact curl command I’m using:

curl -X POST https://huggingface.co/proxy/api-inference.huggingface.co/models/google/flan-t5-small
-H “Authorization: Bearer hf_…”
-H “Content-Type: application/json”
-d ‘{“inputs”: “Translate to French: How are you?”}’

But the response I get is simply:
Not Found

No JSON body, no additional error message — just that plain string.
I’ve tried:

  • Regenerating the token
  • Testing other models like bloomz-560m and t5-base-finetuned-common_gen (same result)
  • Using curl -i to inspect headers

I’m not sure if I’m doing something wrong or if these models aren’t available via Inference API.
but I tried it a few days ago and it worked…
Any guidance would be appreciated!

Thanks in advance :folded_hands:

2

Everyone is in pretty much the same situation right now…

3

Thanks a lot for the reply.
That clarifies things — I guess we just have to wait and hope it gets resolved soon.
I was planning to use it for a class tomorrow, so fingers crossed :crossed_fingers:
Good to know I’m not the only one experiencing this!

4

Hi! This is fixed now though let us know if you continue running into issues. Thanks for reporting!

5

Are you sure? I just tried it with the same curl command and it still doesn’t work.

6

Hi @KR0ld apologies - the model google/flan-t5-small is not deployed by any Inference Provider at the moment, but you can ask for provider support on the model page here: Ask for Provider Support.

Models that are available with Inference Providers can be found here: Models - Hugging Face.

7

I understand. So, what should I do so that my fine-tuned T5 model, which is in my own public/private repo, can have an inference provider implemented?

8

Hi @meganariley the models available for Inference Providers are free ?

9

Hi @Eddy872 Our pricing for Inference Providers is available here: Pricing and Billing.

Subscribe to PRO to receive $2 in free inference credits, among other features: Hugging Face – The AI community building the future..

10

I am not sure I am doing it correctly or not
I am using TypeScript and this is the code

import { HfInference } from “@huggingface/inference”;

const hf = new HfInference(“hf_fj”);

export async function explainCode(code: string): Promise {
const prompt = Explain this code line by line:\n\n${code};

try {
const result = await hf.textGeneration({
model: “google/flan-t5-base”,
inputs: prompt,
});

return result.generated_text || "No explanation returned.";

} catch (error: any) {
return ❌ Error: ${error.message};
}
}
View the image below for error

image
image1609×690 33.4 KB

11

The error message is correct because it does not appear to have been deployed. By the way, the token has been leaked, so let’s disable the leaked token.