pyr0ball
740b0ea45a
- Dockerfile.finetune: PyTorch 2.3/CUDA 12.1 base + unsloth + training stack - finetune_local.py: auto-register model via Ollama HTTP API after GGUF export; path-translate between finetune container mount and Ollama's view; update config/llm.yaml automatically; DOCS_DIR env override for Docker - prepare_training_data.py: DOCS_DIR env override so make prepare-training works correctly inside the app container - compose.yml: add finetune service (cpu/single-gpu/dual-gpu profiles); DOCS_DIR=/docs injected into app + finetune containers - compose.podman-gpu.yml: CDI device override for finetune service - Makefile: make prepare-training + make finetune targets
43 lines
984 B
YAML
43 lines
984 B
YAML
# compose.podman-gpu.yml — Podman GPU override
|
|
#
|
|
# Replaces Docker-specific `driver: nvidia` reservations with CDI device specs
|
|
# for rootless Podman. Apply automatically via `make start PROFILE=single-gpu`
|
|
# when podman/podman-compose is detected, or manually:
|
|
# podman-compose -f compose.yml -f compose.podman-gpu.yml --profile single-gpu up -d
|
|
#
|
|
# Prerequisites:
|
|
# sudo nvidia-ctk cdi generate --output=/etc/cdi/nvidia.yaml
|
|
# (requires nvidia-container-toolkit >= 1.14)
|
|
#
|
|
services:
|
|
ollama-gpu:
|
|
devices:
|
|
- nvidia.com/gpu=0
|
|
deploy:
|
|
resources:
|
|
reservations:
|
|
devices: []
|
|
|
|
vision:
|
|
devices:
|
|
- nvidia.com/gpu=0
|
|
deploy:
|
|
resources:
|
|
reservations:
|
|
devices: []
|
|
|
|
vllm:
|
|
devices:
|
|
- nvidia.com/gpu=1
|
|
deploy:
|
|
resources:
|
|
reservations:
|
|
devices: []
|
|
|
|
finetune:
|
|
devices:
|
|
- nvidia.com/gpu=0
|
|
deploy:
|
|
resources:
|
|
reservations:
|
|
devices: []
|