fix(core): wire text-only Gemma3 GGUF loading #338

	name: CI / cuda

	on:
	workflow_dispatch:
	pull_request:

	jobs:
	test-cuda:
	name: CI / cuda (${{ matrix.name }})

	# For security, only run automatically for:
	# - PRs from the same repo (not forks)
	# - manual workflow dispatch
	if: ${{ github.event_name == 'workflow_dispatch' \|\| github.event.pull_request.head.repo.full_name == github.event.pull_request.base.repo.full_name }}

	concurrency:
	group: ${{ github.workflow }}-${{ matrix.name }}-${{ github.head_ref \|\| github.run_id }}
	cancel-in-progress: true

	runs-on: [self-hosted, Linux, ARM64, gpu, cuda]

	strategy:
	matrix:
	include:
	- name: cuda
	features: cuda,nccl
	# - name: cuda+nccl
	# features: cuda,nccl

	steps:
	- name: Checkout repository
	uses: actions/checkout@v4

	- name: Install dependencies
	run: \|
	sudo apt-get update
	sudo apt-get install -y curl build-essential libssl-dev protobuf-compiler pkg-config

	- name: Install Rust Stable
	uses: dtolnay/rust-toolchain@stable

	- name: CUDA sanity
	run: \|
	nvidia-smi \|\| true
	nvcc --version \|\| true

	- name: Clippy
	run: cargo clippy --features ${{ matrix.features }}

	- name: Test
	run: cargo test --features ${{ matrix.features }}

Provide feedback