Streamer-Sales is a Python project that provides a live-commerce host model with data generation, RAG, TTS, digital human, agent web queries, ASR, Vue frontend, FastAPI backend, and Docker deployment. Latest release v0.10.0 introduces database integration and RESTful API refactor for PostgreSQL and improved deployment docs.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
Streamer-Sales 销冠 —— 卖货主播大模型 is a Python-based project that generates live-commerce narration scripts for products, with features including KV cache-based inference acceleration, RAG retrieval-augmented generation, TTS, digital human video, Agent web queries, ASR, a Vue frontend, FastAPI backend with PostgreSQL and JWT authentication, and Docker-compose deployment.
How it works
The stack includes a frontend built with Vue + pinia + element-plus, and a backend using FastAPI + Uvicorn + PostgreSQL. Inference optimizations use KV cache and Turbomind; RAG integrates a retrieval database; TTS converts text to speech; a digital human video generator produces presenter animations; an Agent module queries live information over the web; ASR converts speech to text. Deployment is supported via Docker-Compose for distributed services. Documentation references environments, deployment steps, and model zoo mappings.
Getting started
Getting started instructions are provided for two deployment modes:
- Docker-Compose (recommended):
git clone https://github.com/PeterH0323/Streamer-Sales.git
cd Streamer-Sales
docker build -t streamer-sales:v0.10.0 -f docker/Dockerfile .
docker-compose up
- Native host deployment (manual):
git clone https://github.com/PeterH0323/Streamer-Sales.git
cd Streamer-Sales
conda env create -f environment.yml
conda activate streamer-sales
pip install -r requirements.txt
Additional deployment steps detail per component (TTs, digital human, ASR, LLM, base services, frontend) with specific bash scripts like bash deploy.sh tts, bash deploy.sh dg, bash deploy.sh asr, bash deploy.sh llm, bash deploy.sh base, and bash deploy.sh frontend.
Recent releases
Latest: v0.10.0 (2024-11-04) — Highlights: integrates PostgreSQL database, refactors interfaces to RESTful API, adds database environment setup docs. Previous: v0.9.0 (2024-09-13) — Backend admin system, frontend rebuilt with Vue, JWT for APIs. Older: v0.8.0 (2024-07-26) — Frontend-backend separation with Docker-Compose support and customizable TTS models; v0.7 (2024-06-28) — ASR support and model file configuration via weights.
Traction
Stars: 3750, Forks: 571, Issues: 14
Behind the repo
Demo URL: https://openxlab.org.cn/apps/detail/HinGwenWong/Streamer-Sales
Caveats
License: AGPL-3.0. Created 2024-04-05; last_push 2025-03-08. Documentation references extensive deployment commands, environment variables, and model weights paths. Open issues: 14. Weight download guidance available in docsweights/README.md. Specific service memory usage and hardware requirements detailed per component (e.g., LLM uses 16G-40G GPUs, 24G+ for 4-bit models).