koko210Serve 9eb081efb1 llama-swap: use pre-built images (:cuda, :rocm) with GPU-specific flags ...

- Drop custom Dockerfiles; docker-compose uses ghcr.io pre-built images
  which ship llama-swap + llama-server with no pinned versions (always latest)
- NVIDIA GTX 1660 (6GB): add -fit off --no-kv-offload --cache-type-k q4_0 --cache-type-v q4_0
  to fix OOM segfault with new llama.cpp b9014's GPU-side KV cache default
- AMD RX 6800 (16GB): flags unchanged; KV cache stays on GPU for max speed
- Both running llama-swap v211 + llama.cpp b9014 (2026-05-05)

2026-05-05 16:53:34 +03:00

backups

refactor: Implement low-latency STT pipeline with speculative transcription

2026-01-22 22:08:07 +02:00

bot

fix: preserve collapsible subsection state across polling re-renders

2026-05-02 16:17:26 +03:00

cat-plugins

Fix vision pipeline: route images through Cat, pass user question to vision model

2026-03-05 21:59:27 +02:00

cheshire-cat

add: cheshire-cat configuration, tooling, tests, and documentation

2026-03-04 00:51:14 +02:00

face-detector

perf: reduce container sizes and build times

2026-02-25 14:41:04 +02:00

models

Initial commit: Miku Discord Bot

2025-12-07 17:15:09 +02:00

readmes

reorganize: consolidate all documentation into readmes/

2026-03-04 00:19:49 +02:00

soprano_to_rvc

add: absorb soprano_to_rvc as regular subdirectory

2026-03-04 00:24:53 +02:00

stt-realtime

perf: reduce container sizes and build times

2026-02-25 14:41:04 +02:00

tests

added test log with multiple various test scenarios between models and evil/regular miku

2026-03-05 22:04:26 +02:00

uno-online

add: absorb uno-online as regular subdirectory

2026-03-04 00:21:38 +02:00

.dockerignore

perf: reduce container sizes and build times

2026-02-25 14:41:04 +02:00

.env.example

cleanup: update .gitignore, sanitize .env.example, remove stale files

2026-03-04 00:17:05 +02:00

.gitignore

fix: update .gitignore to cover all bot/memory subdirs, untrack runtime data

2026-03-04 00:43:10 +02:00

config.yaml

fix: align config.yaml structure with Pydantic AppConfig schema

2026-04-08 14:14:56 +03:00

docker-compose.yml

llama-swap: use pre-built images (:cuda, :rocm) with GPU-specific flags

2026-05-05 16:53:34 +03:00

llama-swap-config.yaml

llama-swap: use pre-built images (:cuda, :rocm) with GPU-specific flags

2026-05-05 16:53:34 +03:00

llama-swap-rocm-config.yaml

feat: add Traefik proxy, custom chat template, improve Cheshire Cat memory

2026-03-04 00:48:58 +02:00

miku-cli.py

Add 'Detect and Join Conversation' button to Web UI and CLI

2025-12-10 14:57:59 +02:00

setup-dual-gpu.sh

Add dual GPU support with web UI selector

2026-01-09 00:03:59 +02:00

setup.sh

Implement comprehensive config system and clean up codebase

2026-02-15 19:51:00 +02:00

Description

Llama.cpp-powered Hatsune Miku Discord bot with autonomous features, chat, and image generation

458 MiB

Languages

Python 82.1%

JavaScript 10.1%

HTML 4.1%

Jupyter Notebook 1.2%

Shell 1%

Other 1.5%