Skip to Content

TGI support

TGI support spans three layers: open-source community help, Hugging Face enterprise offerings, and your own platform team’s runbooks. Know which layer owns your incident before you page someone.

Support paths

PathBest forResponse model
GitHub IssuesOSS bugs, feature requests, community questionsCommunity + HF maintainers
HF EnterpriseProduction SLAs, security reviews, dedicated supportContracted
Inference EndpointsManaged TGI without running K8sHF ops team
Your platform teamInternal deploys, ingress, guardrailsInternal on-call

Hugging Face Inference Endpoints

Inference Endpoints  is managed TGI on AWS, GCP, and Azure. You pick a model, instance type, and region — HF runs the server, upgrades, and scaling.

FactorSelf-hosted TGIInference Endpoints
Ops burdenYou own K8s, GPUs, upgradesHF owns infrastructure
Cost modelGPU-hours + engineer timePer-hour instance pricing
CustomizationFull flag controlDashboard-configured options
ComplianceYour cluster, your rulesHF-managed environment

Meter managed endpoints like cloud APIs. Self-hosted TGI meters GPU-hours. Compare both on cost per resolved outcome, not sticker price.

Enterprise support

HF Enterprise adds:

  • Dedicated support channels and SLAs
  • Security and compliance review for model serving
  • Priority access to TGI releases and roadmap input

Engage enterprise support when TGI is on your production critical path and GitHub Issues response times are too slow for your SLOs.

HF token management

Gated models require a Hugging Face token on every environment that pulls weights:

export HF_TOKEN=hf_...
EnvironmentPattern
Local Docker-e HF_TOKEN on docker run
KubernetesSecret mounted as env var — see Kubernetes
CIShort-lived token in CI secrets; rotate quarterly

Failure mode: gated model pulls fail silently in CI when tokens expire. Add a smoke test that verifies model access on every deploy.

Version pinning

Pin TGI container images in production — latest breaks API paths and flag names without warning:

image: ghcr.io/huggingface/text-generation-inference:2.4.0

Check the release notes  before upgrading. Test chat completions compatibility on your TGI host in staging.

When to escalate

SymptomFirst checkEscalate to
OOM on all replicasServer flags — quantize or shrink inputPlatform team
API 404 after upgradeAPI version differencesPin image, file GitHub Issue
Hub download failuresToken expiry, rate limitsRotate HF_TOKEN
Quality regressionQuantization or batch configRe-benchmark vs vLLM

Community resources

Last updated on