grail/gf-lt
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Enha: docker-compose to not depend on locally downloaded models

Browse Source
This commit is contained in:
Grail Finder
2025-12-07 11:53:10 +03:00
parent 5582739e3c
commit 5bbb134aca
3 changed files with 22 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
Makefile
View File

@@ -36,6 +36,7 @@ download-whisper-model: ## Download Whisper model for STT in batteries directory
# Docker targets for STT/TTS services (in batteries directory) # Docker targets for STT/TTS services (in batteries directory)
docker-up: ## Start Docker Compose services for STT and TTS from batteries directory docker-up: ## Start Docker Compose services for STT and TTS from batteries directory
@echo "Starting Docker services for STT (whisper) and TTS (kokoro)..." @echo "Starting Docker services for STT (whisper) and TTS (kokoro)..."
@echo "Note: The Whisper model will be downloaded automatically inside the container on first run"
docker-compose -f batteries/docker-compose.yml up -d docker-compose -f batteries/docker-compose.yml up -d
@echo "Docker services started. STT available at http://localhost:8081, TTS available at http://localhost:8880" @echo "Docker services started. STT available at http://localhost:8081, TTS available at http://localhost:8880"
@@ -51,3 +52,4 @@ docker-logs: ## View logs from Docker services in batteries directory
# Convenience target to setup everything # Convenience target to setup everything
setup-complete: setup-whisper docker-up setup-complete: setup-whisper docker-up
@echo "Complete setup finished! STT and TTS services are running." @echo "Complete setup finished! STT and TTS services are running."
@echo "Note: Docker services will download the Whisper model automatically if not present."

9
README.md
View File

@@ -64,3 +64,12 @@ Ctrl+x: cycle through mentioned chars in chat, to pick persona to send next msg
cp config.example.toml config.toml cp config.example.toml config.toml
``` ```
set values as you need them to be. set values as you need them to be.
#### setting up STT/TTS services
For speech-to-text (STT) and text-to-speech (TTS) functionality:
1. The project uses Whisper.cpp for STT and Kokoro for TTS
2. Docker Compose automatically downloads the required Whisper model on first run
3. To start the services: `make docker-up`
4. To stop the services: `make docker-down`
5. The STT service runs on http://localhost:8081
6. The TTS service runs on http://localhost:8880

13
batteries/docker-compose.yml
View File

@@ -6,10 +6,17 @@ services:
ports: ports:
- "8081:8081" - "8081:8081"
volumes: volumes:
- ./whisper.cpp/models:/app/models - whisper_models:/app/models
working_dir: /app working_dir: /app
entrypoint: "" entrypoint: ""
command: ["./build/bin/whisper-server", "-m", "/app/models/ggml-large-v3-turbo.bin", "-t", "4", "-p", "1", "--port", "8081", "--host", "0.0.0.0"] command: >
sh -c "
if [ ! -f /app/models/ggml-large-v3-turbo.bin ]; then
echo 'Downloading ggml-large-v3-turbo model...'
./download-ggml-model.sh large-v3-turbo /app/models
fi &&
./build/bin/whisper-server -m /app/models/ggml-large-v3-turbo.bin -t 4 -p 1 --port 8081 --host 0.0.0.0
"
environment: environment:
- WHISPER_LOG_LEVEL=3 - WHISPER_LOG_LEVEL=3
# Restart policy in case the service fails # Restart policy in case the service fails
@@ -40,3 +47,5 @@ volumes:
driver: local driver: local
audio: audio:
driver: local driver: local
whisper_models:
driver: local