check again and start the python server in background
Implement GPU Inference with WebGPU
check again and start the python server in background
Both fixes applied. Let me check what's calling at the bottom and start the server.
Good, it's calling forward().
Server running. Open http://localhost:8000 in Chrome and check the console. You should see 10 probabilities, with the highest one at index 7 (the first test image is a 7).
0.0000017714637579047121, 1.3455591840738634e-7, 0.00026626381441019475, 0.00022798024292569607, 2.9781840882314725e-10, 0.000003364992153365165, 6.689376049612283e-12, 0.9994951486587524, 2.0764231578596082e-7, 0.0000051217348300269805