Create endpoint

documented

Create/deploy a new Inference Endpoint.

POST /endpointsInference Endpointsno implementation yet0 runs

Input

endpoint_name*

Endpoint name

my-endpoint

repository*

Model repository (e.g., 'gpt2' or 'username/repo')

gpt2

task*

Inference task (e.g., 'text-generation')

text-generation

framework

Framework (e.g., 'pytorch' or 'custom')

pytorch

accelerator*

Hardware accelerator (e.g., 'cpu' or 'gpu')

cpu

vendor*

Cloud vendor (e.g., 'aws' or 'azure')

aws

region*

Cloud region (e.g., 'us-east-1')

us-east-1

instance_size*

Instance size (e.g., 'x2')

instance_type*

Instance type (e.g., 'intel-icl')

intel-icl

type*

Endpoint type (e.g., 'protected' or 'public')

protected

namespace

Namespace (organization or user) [optional]

myorg

revision

Model revision (optional)

main

custom_image

Custom image (JSON payload)

{...}

secrets

Secrets (JSON)

{...}

Calls Hugging Face as you (your API key is your account); keychains.dev holds the credentials.