Update GPU model from RTX 3090 to RTX 5090
Updated GPU model in requirements and README.
This commit is contained in:
@@ -38,7 +38,7 @@ Then open:
|
||||
- **Backend**: FastAPI + PyTorch + Transformers 4.46 + DeepSeek-OCR
|
||||
- **Server**: Nginx (reverse proxy)
|
||||
- **Container**: Docker + Docker Compose with multi-stage builds
|
||||
- **GPU**: NVIDIA CUDA support (tested on RTX 3090)
|
||||
- **GPU**: NVIDIA CUDA support (tested on RTX 5090)
|
||||
|
||||
## Project Structure
|
||||
|
||||
@@ -79,7 +79,7 @@ npm run dev
|
||||
## Requirements
|
||||
|
||||
- Docker & Docker Compose
|
||||
- NVIDIA GPU with CUDA support (tested on RTX 3090)
|
||||
- NVIDIA GPU with CUDA support (tested on RTX 5090)
|
||||
- nvidia-docker runtime
|
||||
- ~8-12GB VRAM for model
|
||||
|
||||
|
||||
Reference in New Issue
Block a user