NeuralThyself/fastvlm-private

🔧 Fix: Downgrade Gradio 5.9.1 → 4.44.0 (remove SSR)

🔧 Fix: Remove invalid ssr parameter

🔧 Fix: Disable SSR to enable API endpoints

🔧 Revert to Gradio SDK (ZeroGPU requirement)

🔧 Fix: Switch from Gradio to Docker SDK with Flask REST API

Upload app.py with huggingface_hub

Upload requirements.txt with huggingface_hub

Upload requirements.txt with huggingface_hub

Upload README.md with huggingface_hub

Upload app.py with huggingface_hub

Delete Dockerfile

Fix: Add wget for HF Dev mode openvscode-server

Fix: Use torch 2.4.0 + torchvision 0.19.0 (matching versions)

Fix: PyTorch version conflict - use torch 2.5.0 + torchvision 0.19.0

Fix: Restore Dockerfile and proper ZeroGPU config

Switch to Gradio SDK (no Docker)

Initial deployment: FastVLM private Space with ZeroGPU

initial commit