Guide/Tutorial to write an inference endpoint for custom models

[alvations]

Are there any docs/guides/tutorials to write custom endpoints to be hosted on Huggingface Hub’s inference endpoints?

We’re particularly looking at these models which does things a little differently from normal Huggingface AutoModels (that’ll need A100):

• Unbabel/XCOMET-XXL · Hugging Face
• google/metricx-23-xxl-v2p0 · Hugging Face
• GitHub - google-research/metricx
• GitHub - Unbabel/COMET: A Neural Framework for MT Evaluation

And potentially also these that runs on T4:

• GitHub - unslothai/unsloth: 2-5X faster 70% less memory QLoRA & LoRA finetuning