Finalize image-feature-extraction support

[osanseviero]

Although we now have the ImageFeatureExtractionPipeline, we're still missing support in a couple of places

• In the task schemas, https://github.com/huggingface/huggingface.js/tree/main/packages/tasks/src/tasks
• In the Inference API, see https://github.com/search?q=repo%3Ahuggingface%2Fapi-inference%20feature-extraction&type=code for where it needs to be added
• The task page https://huggingface.co/tasks/image-feature-extraction
• In the widget (for this, we can reuse the feature-extraction widget but with an image input

cc @pcuenca @merveenoyan

[merveenoyan]

We have merged the task page, I'll open a PR to API Inference shortly, as for first item I didn't get what you meant. For the widget I'll see if I can do it.

[osanseviero]

For all tasks we have a formal spec that is used in downstream applications (e.g. inference API, API clients, etc). For example, here is the spec for feature-extraction https://github.com/huggingface/huggingface.js/tree/main/packages/tasks/src/tasks/feature-extraction/spec which is used to automatically generate https://github.com/huggingface/huggingface.js/blob/main/packages/tasks/src/tasks/feature-extraction/inference.ts as well

[merveenoyan]

@osanseviero I opened the PR

[osanseviero]

Please tag PRs so they are nicely linked: #614