Deploy from GitHub commit affdcd663b2d42fe3ddfd8139267f7e816e91713

Deploy from GitHub commit 9ea40055df1444189c87e54a990237f41e397404

Deploy from GitHub commit c41d8036f5d265065967accde19305ed620eb213

fix: update Dockerfile to use CUDA base image and install Python dependencies, adjust segmentation model in configuration

fix: change image encoding from RGBA to JPEG for improved performance and reduced size

fix: add GZipMiddleware for response compression and reduce MAX_DIM for image resizing

fix: optimize file reading in viz2d_job_file by reading entire file at once

fix: switch to aiofiles for asynchronous file reading in viz2d_job_file

fix: add image resizing to maintain dimensions within MAX_DIM for segmentation bundle

fix: replace GZipMiddleware with StreamingResponse for efficient file streaming

fix: add GZipMiddleware to compress responses for improved performance

fix: remove Content-Disposition header from viz2d_job_file response

fix: restrict CORS to specific origin and enable credentials

fix: ensure src array is converted to float32 in estimate_floor_plane function

fix: correct contour area calculation in estimate_floor_plane function

deploy floor visualizer backend