Add vLLM#221

Merged

doringeman merged 24 commits intomainfrom

vllm

Oct 22, 2025

Commits on Oct 21, 2025

build: add new vLLM CUDA image
doringeman
committed
fix(Dockerfile): add build-essential package

ilopezluna
authored and
doringeman
committed
fix(Dockerfile): add python3-venv and python3-dev packages

ilopezluna
authored and
doringeman
committed
fix(Dockerfile): use the final target for CPU and CUDA builds
doringeman
committed
feat(vllm): implement vLLM backend

ilopezluna
authored and
doringeman
committed
fix(client): remove unsupported model format checks for Safetensors

ilopezluna
authored and
doringeman
committed
perf: optimize Docker build by restructuring stage dependencies
doringeman
committed
feat(cli): allow vLLM backend
doringeman
committed
fix(vllm): use model bundle directory
doringeman
committed
refactor(backend): add modelRef parameter to Backend.Run interface
doringeman
committed
fix(client): remove unsupported model format checks for Safetensors on Linux
doringeman
committed
fix(vllm): check vllm.New error
doringeman
committed
vllm: Linux support only
doringeman
committed
vllm: add version detection from build-time capture
doringeman
committed
feat(scheduler): automatically identify models for vLLM
doringeman
committed
feat(vllm): enhance argument handling for vLLM backend configuration
ilopezluna
committed
fix(vllm): update model path handling to use directory for safetensors
ilopezluna
committed
fix(vllm): add model ID as served-model-name
doringeman
committed
feat(vllm): add argument sanitization for safe logging
ilopezluna
committed
update actions versin to be pinned to a full length commit SHA
ilopezluna
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Add vLLM#221

Add vLLM#221
doringeman merged 24 commits intomainfrom
vllm

Commits on Oct 21, 2025

build: add new vLLM CUDA image

fix(Dockerfile): add build-essential package

fix(Dockerfile): add python3-venv and python3-dev packages

fix(Dockerfile): use the final target for CPU and CUDA builds

feat(vllm): implement vLLM backend

fix(client): remove unsupported model format checks for Safetensors

perf: optimize Docker build by restructuring stage dependencies

feat(cli): allow vLLM backend

fix(vllm): use model bundle directory

refactor(backend): add modelRef parameter to Backend.Run interface

fix(client): remove unsupported model format checks for Safetensors on Linux

fix(vllm): check vllm.New error

vllm: Linux support only

vllm: add version detection from build-time capture

feat(scheduler): automatically identify models for vLLM

feat(vllm): enhance argument handling for vLLM backend configuration

fix(vllm): update model path handling to use directory for safetensors

fix(vllm): add model ID as served-model-name

feat(vllm): add argument sanitization for safe logging

update actions versin to be pinned to a full length commit SHA

Commits on Oct 22, 2025

fix(vllm): improve logging for backend errors and warn about unsupported safetensors format

fix(vllm): sanitize model reference in backend error logging

fix(vllm): sanitize model reference in warning log for unsupported safetensors format

fix(vllm): validate safetensors path before serving and update test cases