fix: Import Full/Empty from queue to prevent NameError on backpressure

feat: Continuous object tracking, speed estimation, and overlay syncing

Fix inference concurrency bugs and enable default first-frame GPT

fix: Move CUDA_VISIBLE_DEVICES clearing to module import time

perf: Tune batch sizes and queues for A10 GPUs

feat: Implement AsyncVideoReader for parallel video decoding

Uncomment CUDA_VISIBLE_DEVICES clearing to enable all GPUs

Optimize streaming buffer and implement parallel depth pipeline

Refactor: Decouple depth estimator and suppress noisy logs

Implement Batch Inference & Queue Backpressure Fixes

Enable MJPEG streaming for live video feedback

fix: add wget to Dockerfile, cleanup js errors, and remove GPU env override

Integrated GPT Reasoning and Object Track Cards

Fix frontend layout and make depth estimation optional backend-wide

Feature: GPT distance integration, Radar Map, and optional legacy depth

Fix: GPU detection order and distance consistency

fix:distance consistency

fix:multi-gpu support

update:inference pipeline optimization

Fix GPU visibility, update depth scale, and improve radar display timing

Reduce log verbosity and update default depth scale to 100

Optimize model loading to prevent reload loop

Fix draw_boxes arguments

Fix IndentationError and undefined depth_raw variable in inference.py

feat: fix depth inversion and add video overlays

chore: add detailed GPU diagnostics logs

feat: fix multi-gpu visibility and improve logging

added debug level logging for GPU inferenciing and distance

feat(inference): enable full multi-GPU support for all models

added estimated/relative distance

added depth view

fixed overflow

fixed depth norm

added fallback for Nan

added apple depth pro

wired endpoints to new html

fixed numpy error

modified segmentation mask display

modified segmentation bbox display

modified bbox display

added async first frame/video detection

removed owlv2

inital commit