kdeps-recipes-tgi
v1.0.0Hugging Face TGI with OpenAI-compatible /v1 (GPU-oriented)
set recipes / shelf model
LLM context: none - kdeps uses it without sending it to the model
Where it runs
Part of the recipes set, used in the kdeps CLI. How to build and run an LLM server as a Docker appliance: engine, base image, install steps, ports and health check.
- kdeps llm build or kdeps llm run
- the recipe is rendered into an imagerecipes/tgi
- the LLM server starts
- workflows call it
Loaded by: kdeps llm list, kdeps llm, kdeps llm build, kdeps llm export iso, kdeps llm export k8s, kdeps llm run, kdeps llm show, kdeps llm wizard
In the code: kdeps llm listcatalog.LoadDefaultcatalog.Load
Install
A kdeps asset: installed with the first-run profile that includes its set, or the first time a command needs it, and updated without a new release.
Pin a version with kdeps install recipes/tgi@<version>.
Contents
Hugging Face TGI with OpenAI-compatible /v1 (GPU-oriented)
- id
- tgi
- name
- Text Generation Inference (TGI)
api:
port: 8000
base_path: /v1
chat_path: /v1/chat/completions
health:
method: GET
path: /health
auth:
mode: none
env: LLM_API_KEY
engine:
kind: tgi
base_image: ghcr.io/huggingface/text-generation-inference:latest
packages: []
install: |
true
command:
- sh
- -c
- exec text-generation-launcher --hostname 0.0.0.0 --port 8000 --model-id
"${LLM_MODEL}"
env: {}
internal_port: 8000
openai_bridge: false
resources:
gpu: required
memory_hint: 16Gi
capabilities:
- chat
id: tgi
name: Text Generation Inference (TGI)
description: Hugging Face TGI with OpenAI-compatible /v1 (GPU-oriented)
version: 1.0.0
api:
port: 8000
base_path: /v1
chat_path: /v1/chat/completions
health:
method: GET
path: /health
auth:
mode: none
env: LLM_API_KEY
engine:
kind: tgi
base_image: "ghcr.io/huggingface/text-generation-inference:latest"
packages: []
install: |
true
# TGI entrypoint; model id/path via LLM_MODEL (HF hub id or local path)
command:
- sh
- -c
- exec text-generation-launcher --hostname 0.0.0.0 --port 8000 --model-id "${LLM_MODEL}"
env: {}
internal_port: 8000
openai_bridge: false
models:
strategy: pull
default: ""
resources:
gpu: required
memory_hint: "16Gi"
capabilities:
- chat
Versions
| Version | Published | Status |
|---|---|---|
| 1.0.0 | 10/11/2026 | active |
Details
- Author
- kdeps
- License
- Apache-2.0
- Latest Version
- 1.0.0
- Published
- 10/11/2026