Commit History
-
fix Gradio 6 launch argument regression
Remove the unsupported show_api launch argument so the Space can boot on the upgraded Gradio runtime while keeping the named predict API endpoint in place. Made-with: Cursor67ce2d9 -
fix predict API using supported Gradio endpoint
Replace the unsupported gr.api.post decorator with a documented gr.api named endpoint, and switch the demo client to send data URLs to Gradio's direct /gradio_api/run/predict route. Made-with: Cursor69a3974 -
upgrade Gradio and restore predict API route
Bump the Space to Gradio 6.9 so custom POST endpoints are supported again, wire the JSON predict route through gr.api.post, and update the demo client to use the generated /post/predict path. Made-with: Cursor0190f43 -
fix Gradio Space startup path for predict route
Restore the normal demo.launch() flow for the Gradio Space and attach the custom predict endpoint to the Gradio app instead of replacing the server runtime. Made-with: Cursorb5e2759 -
fix early startup thread and ORT warnings
Normalize OpenMP-related env vars before importing torch and set ONNX Runtime logger severity before rembg initializes so the Space starts with fewer container-specific warnings. Made-with: Cursorc011205 -
fix Space startup for mounted API app
Remove the extra uvicorn runner that conflicted with the Space runtime, and normalize thread env vars before rembg imports while forcing CPU execution for background removal. Made-with: Cursor8fc7e47 -
fix JSON endpoint boot path and add demo client
Mount the Gradio app on FastAPI so the plain /predict route works on the Space's Gradio version, and add a small Node script for batch testing the endpoint from test_demo/inputs. Made-with: Cursor2ec5172 -
add JSON prediction endpoint for direct app integration
Expose a multipart POST /predict route so apps can send images with fetch() and receive pose results as JSON without using the Gradio client. Made-with: Cursordc74bcf
-
Enhance background preprocessing to handle Kubernetes CPU limits for rembg
727fc07 -
fix
cfb7d8c -
fix
2ce5c07
-
add duration
ccc1f92 -
add spaceGPU to func run_inference
23d2d0c -
no operate cuda
274424f -
test torch 2.8
a22d9ab -
update requirements.txt
f14abb6 -
set cuda
b94ed06 -
test float16
a6ef71e -
cublas bug
9ab87b8 -
stateless GPU env
5452770 -
fix NameError
fbb8705 -
global model
ee8949b -
Update inference.py
Verified signature9746db7 -
Update app.py
Verified signaturea45cf3c -
test ZeroGPU
Verified signatureb2c252d -
update Zero GPU
Verified signature0342306 -
Update requirements.txt
Verified signature27ede2b -
Update README.md
Verified signature3c7e4d0 -
update python=3.10 to match bpy=4.0
Verified signature59ea259 -
Update requirements.txt
Verified signaturebpy=4.03bc2e6a -
Update README.md
Verified signaturec1721ec
-
Update README.md
Verified signature469631e -
Update README.md
Verified signaturec875524
-
cuda
zhangziang939a4f3 -
update layout,model,example
zhangzianga1a407f
-
layout adjust
zhangziang9868529 -
layout adjust
zhangziang53a5272 -
mark_dtype
zhangziangf772fe2 -
config python version
zhangziang7c54c2e -
initial commit track binary
zhangziangf783161
-
initial commit
Verified signature706e128