[Inference doc] Next gen inference snippets #1643
Conversation
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
|
Handled by huggingface/doc-builder#549 see example: #1656 |
|
hmm https://moon-ci-docs.huggingface.co/docs/api-inference/pr_1643/en/index has not been updated no? |
It's now under https://moon-ci-docs.huggingface.co/docs/inference-providers/pr_1643/en/index (but in general this PR is not ready, I finishing it up) |
|
@julien-c @gary149 @hanouticelina @SBrandeis PR is finally ready for review. Best way to review is to go on https://moon-ci-docs.huggingface.co/docs/inference-providers/pr_1643/en/tasks/index |
| <InferenceSnippet | ||
| pipeline=automatic-speech-recognition | ||
| providersMapping={ {"fal-ai":{"modelId":"openai/whisper-large-v3","providerModelId":"fal-ai/whisper"},"hf-inference":{"modelId":"openai/whisper-large-v3-turbo","providerModelId":"openai/whisper-large-v3-turbo"}} } | ||
| /> |
There was a problem hiding this comment.
so in the end you don't embed the full snippets in the markdown. Correct?
There was a problem hiding this comment.
but we still need to regenerate the .md from time to time to pick up the mapping?
There was a problem hiding this comment.
Yes exactly. If we want we can still have an .md export for llms but not for now
There was a problem hiding this comment.
There was a problem hiding this comment.
so in the end you don't embed the full snippets in the markdown. Correct?
inference snippets strings are directly coming from https://github.com/huggingface/huggingface.js (single source of truth)
There was a problem hiding this comment.
^ I think that statement does not hold anymore - we moved snippets code to moon-landing
There was a problem hiding this comment.
I think there is a confusion here 😬
Inference snippets are indeed located in huggingface.js/inference (or I'm not aware of the change) and providers mapping is on the Hub => and pulled from the Hub by the script in this PR to make sure we always get up-to-date models in the docs => done in CI once a day
So all good, right?
There was a problem hiding this comment.
Sorry, I've mistaken 'snippets' and 'widgets' - my bad!
| </js> | ||
|
|
||
| </inferencesnippet> | ||
| No snippet available for this task. |
There was a problem hiding this comment.
there's no warm hf-inference for this seminal task? Que pena!!
How to review
Go to https://moon-ci-docs.huggingface.co/docs/inference-providers/pr_1643/en/tasks/index
Description
Opening this PR as a draft as it'll requires some deeper changes in doc-builder cc @julien-c @gary149 @mishig25 @SBrandeis @hanouticelina
This PR updates the script to generate the tasks pages of the Inference API docs (e.g. text-to-image page). The short-term goal is to revamp the "Inference API" docs as an "Inference Providers" one, making them more provider-centric. With this PR, each task page will have
<inferencesnippets>tag to have an interface similar to the one on the model pageWhat I did in this PR is to generate the snippets and serialize them inside some
<snippet provider="..." language="..." client="...">tags. This is where doc-builder should be updated to take them into account.EDIT: using the new
</InferenceSnippet>svelte component from huggingface/doc-builder#549 (thanks @mishig25!)