aaron/rw-deepseek-ocr
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update GPU model from RTX 3090 to RTX 5090

Browse Source 
Updated GPU model in requirements and README.
This commit is contained in:
rdumasia303
2025-10-21 09:40:44 +01:00
committed by GitHub
parent aec04f6eb4
commit a7aeeaf109
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
README.md
View File

@@ -38,7 +38,7 @@ Then open:
- **Backend**: FastAPI + PyTorch + Transformers 4.46 + DeepSeek-OCR - **Backend**: FastAPI + PyTorch + Transformers 4.46 + DeepSeek-OCR
- **Server**: Nginx (reverse proxy) - **Server**: Nginx (reverse proxy)
- **Container**: Docker + Docker Compose with multi-stage builds - **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 ## Project Structure
@@ -79,7 +79,7 @@ npm run dev
## Requirements ## Requirements
- Docker & Docker Compose - Docker & Docker Compose
- NVIDIA GPU with CUDA support (tested on RTX 3090) - NVIDIA GPU with CUDA support (tested on RTX 5090)
- nvidia-docker runtime - nvidia-docker runtime
- ~8-12GB VRAM for model - ~8-12GB VRAM for model