mirror of
https://github.com/LostRuins/koboldcpp.git
synced 2026-05-17 04:09:19 +00:00
26 lines
No EOL
1.2 KiB
YAML
26 lines
No EOL
1.2 KiB
YAML
version: "3.2"
|
|
services:
|
|
koboldcpp:
|
|
container_name: koboldcpp
|
|
image: koboldai/koboldcpp:latest
|
|
security_opt:
|
|
- apparmor:unconfined # AppArmor breaks sockets, don't worry everything is securely confined within docker
|
|
volumes:
|
|
- ./kcpp-data:/workspace/:rw
|
|
# devices: # Uncomment this section to add support for AMD/Intel GPU's
|
|
# - /dev/dri:/dev/dri
|
|
deploy: # You can remove this section if you do not wish to use an Nvidia GPU
|
|
resources:
|
|
reservations:
|
|
devices:
|
|
- driver: nvidia
|
|
device_ids: ['0']
|
|
capabilities: [gpu]
|
|
environment:
|
|
- KCPP_DONT_UPDATE=false
|
|
- KCPP_DONT_TUNNEL=true
|
|
- KCPP_ARGS=--model https://huggingface.co/unsloth/gemma-3-4b-it-GGUF/resolve/main/gemma-3-4b-it-Q4_K_S.gguf?download=true --gpulayers 99 --multiuser 20 --admin --admindir . --adminpassword ChangeMe # For split models in the 00001-of format specify the first file as the --model, remove --gpulayers 99 --multiuser 20 if you do not have a GPU
|
|
# Don't forget to change the model for production, the default model was chosen to fit on many systems rather than for its quality.
|
|
ports:
|
|
- "5001:5001"
|
|
restart: unless-stopped |