Spaces:

Samir87699
/

vanishly

Sleeping

App Files Files Community

vanishly

Commit History

May 10

04:55
Fix quality loss by avoiding patch downscaling and improving text-erase crop

Samir87699

e94e2ed

May 9

09:10
Add /remove_bg endpoint using rembg (u2net ONNX) for CPU background removal

Samir87699

e152ab7

May 2

10:19
Fix dependency version conflicts for Hugging Face deployment⁠

Samir87699

438bfc5
06:01
Add /text_erase: single-call text-prompt object removal (CLIPSeg+LaMa)

Samir87699

1b19dfc
05:41
Add CLIPSeg text-based segmentation endpoint /segment_text

Samir87699

04e4c05
05:02
Fix: remove duplicate torch import causing UnboundLocalError in _load_realesrgan

Samir87699

60d8cd9
04:53
Align with healthy gemini-eraser-ai space: remove CLIPSeg/transformers, use RealESRGAN_x4plus.pth from GitHub

Samir87699

c6cbe1c
04:41
Fix: download RealESRGAN x4 from GitHub instead of COPY missing 4x-UltraSharp.pth

Samir87699

2e2e7b0
04:33
Fix model loading: use absolute paths, retry logic, proper model assignment

Samir87699

e49de92

May 1

06:58
Update backend with text segmentation, UltraSharp, and lossless compositing

Samir87699

f2ba69e

Apr 18

04:33
fix: resolve runtime error in main.py by correcting variable initialization

Samir87699

c66ffdc
03:37
Optimize Real-ESRGAN CPU inference for Free Tier (1 thread, tile=128)

Samir87699

7ff983a
02:51
Fix basicsr torchvision compatibility by safely down-pinning TorchVision

Samir87699

02a5b3f
02:42
Upgrade backend to use Real-ESRGAN

Samir87699

522c74d

Apr 12

13:16
Update backend with Vanishly rebrand and AI upscaling capabilities

Samir87699

e708646
05:16
feat: FSRCNN AI upscaling endpoint via OpenCV dnn_superres (PRO feature)

Samir87699

ddfab46
02:12
perf: async LaMa offload, JPEG response, server-side downscale, faster FastSAM

Samir87699

18f6d62

Apr 10

12:42
Optimize FastSAM inference for extremely low latency

Samir87699

3ad0dbd
12:34
Bake FastSAM weights into Docker image to avoid RO filesystem error

Samir87699

65fed58
12:33
Fix HF glib dependencies

Samir87699

002d1a1
12:32
Fix form inputs for FastAPI

Samir87699

01ee9fd
12:24
Add FastSAM cloud segment endpoint

Samir87699

bcc1066
05:37
chore: update project metadata, dependencies, and Docker configuration

Samir87699

fcaebd2

Apr 9

12:07
Initial commit for HF Spaces backend

Samir87699

1c446a3

Commit History

May 10

Fix quality loss by avoiding patch downscaling and improving text-erase crop

May 9

Add /remove_bg endpoint using rembg (u2net ONNX) for CPU background removal

May 2

Fix dependency version conflicts for Hugging Face deployment⁠

Add /text_erase: single-call text-prompt object removal (CLIPSeg+LaMa)

Add CLIPSeg text-based segmentation endpoint /segment_text

Fix: remove duplicate torch import causing UnboundLocalError in _load_realesrgan

Align with healthy gemini-eraser-ai space: remove CLIPSeg/transformers, use RealESRGAN_x4plus.pth from GitHub

Fix: download RealESRGAN x4 from GitHub instead of COPY missing 4x-UltraSharp.pth

Fix model loading: use absolute paths, retry logic, proper model assignment

May 1

Update backend with text segmentation, UltraSharp, and lossless compositing

Apr 18

fix: resolve runtime error in main.py by correcting variable initialization

Optimize Real-ESRGAN CPU inference for Free Tier (1 thread, tile=128)

Fix basicsr torchvision compatibility by safely down-pinning TorchVision

Upgrade backend to use Real-ESRGAN

Apr 12

Update backend with Vanishly rebrand and AI upscaling capabilities

feat: FSRCNN AI upscaling endpoint via OpenCV dnn_superres (PRO feature)

perf: async LaMa offload, JPEG response, server-side downscale, faster FastSAM

Apr 10

Optimize FastSAM inference for extremely low latency

Bake FastSAM weights into Docker image to avoid RO filesystem error

Fix HF glib dependencies

Fix form inputs for FastAPI

Add FastSAM cloud segment endpoint

chore: update project metadata, dependencies, and Docker configuration

Apr 9

Initial commit for HF Spaces backend