Repository Issue Activity (beta)

#1391 Integrate `turbojpeg` for accelerated CPU image decoding

#1433 TPU: XLA recompiles on the ground-truth box-count tuple, not just image shape

#1412 N-dataset_builder_option_defaults

#1272 Query embedding extraction

#1408 Docs/CI: Self-Contained Weekly Metrics SVG Generator

#1382 Could not get the algorithm running

#1349 CLAHE clip_limit: Kornia backend pins a scalar where Albumentations samples a range

#1414 Issue with drawing overlapping bounding boxes during the prediction phase after training a multiclass detection using RF-DETR

#1410 Investigation: faster training throughput without accuracy regression — starting with `torch.compile`

#529 training RF-DETR from scratch

#1133 Proposal: Gather exceptions and name them

#1401 Roadmap for sub-10M and sub-5M parameter models for Edge CPU deployment

#564 [Refactor]: Improve typing and sanitize docs

#1347 Suggestion: merge dataset in `val` and `train` folder to train for final epochs

#1351 Kornia GaussianBlur and GaussNoise defaults diverge from the Albumentations defaults they stand in for

#1058 TPU Training Support

#1392 Implement `WebDataset` pipeline for sequential I/O

#1365 TFLite export fails when onnxsim is not on PATH

#1344 Suggestion: store decoded dataset images in RAM one time only instead of before every epoch

#1348 v1.10.0.dev: the program takes the best EMA from early quick sanity check, not from real trained epoch, and generates fake `checkpoint_best_ema.pth`

#1363 TFLite INT8 docs and warning incorrectly require calibration data

#1116 `deploy_to_roboflow()` reports success but model upload fails server-side ("Model Upload Failed" in UI)

#1153 The slow progress of RF-DETR training and the post-training validation process

#1326 Missed chart legends on charts generated by function `plot_metrics()`

#1354 Support training on a selected subset of classes