Add weights/best_int8.onnx

Add weights/best.onnx.data

Add weights/best.onnx

Add weights/last.pt

Add weights/best.pt

Do not let the ignore rules hide weights/ from upload_folder

Add weights/best_int8.onnx

Add weights/best.onnx.data

Add weights/best.onnx

Add weights/last.pt

Add weights/best.pt

Add the fine-tuned YOLO11n checkpoint, its exports and training log

added all the files

Drop the stock-weights note from the Pi section

Add the Pi 5 results and ready the submission repo

Add the Tesla T4 GPU results

Drop barrel, add held-out stop-sign distance, add the GPU runbook

Make setup.sh say what it is doing

Add a Pi 5 runbook, and correct the NCNN diagnosis

Quantise the ONNX export, and record how the model performs

Score checkpoints end to end and keep the results

Add a portable evaluation package builder

Add a portable evaluation package, and drop the transcription files

Phase 3: per-format benchmark with stage timing

Document src/depth for transcription

Phase 2: distance from focal length and pixel geometry

Put repo cloning back into the local notebook

Describe both training notebooks in the README

Add a local training notebook

Mirror checkpoints to Drive, and move the logic out of the notebook

Fix the Colab clone cell: shell lines cannot see Python variables

Make the gradient-norm probe safe to leave on

Log train loss, val loss and gradient norm to Weights & Biases

Keep the training recipe identical to the reference notebook

Stop tracking the editable-install egg-info

Host the code on the Hub alongside the dataset

Publish the dataset to the Hub and verify the round trip

Assemble the full dataset on nuScenes trainval

Fix two cross-dataset convention bugs found by scoring against ground truth

Add COCO and AV2 fetchers that pull only what the pipeline reads

Verify the pipeline on nuScenes v1.0-mini

Phase 1: unified data pipeline, QA and training