> This is the documentation for LibreYOLO 1.5.0. Current version: https://www.libreyolo.com/docs/changelog # Changelog A summary of the repository changelog: what 1.5.0 shipped, and what 1.4.0 shipped before it. Verified against LibreYOLO v1.5.0. ## 1.5.0, released 2026-08-09 The largest release so far: 28 new model families, four new tasks, five new export formats, two new serving backends, and an `import libreyolo` that no longer pulls PyTorch. 600 commits and 88 merged pull requests. Summarized from [`CHANGELOG.md`](https://github.com/LibreYOLO/libreyolo/blob/release/CHANGELOG.md), which carries the full entries. Two defaults move numbers and four arguments changed shape: [upgrading to 1.5.0](/docs/v1.5.0/upgrade) is the short list of what that asks of you. ### Breaking changes No public model class or function was removed, and `__all__` grew from 101 names to 142. Four things need a code edit: - `allow_experimental=True` is gone from every `.train()` gate. Delete the argument; a call that still passes it raises `TypeError`. - The export support tier `"experimental"` is gone. `Tier` is now `validated`, `available` or `blocked`. - `pretrained=False` combined with `resume` raises `ValueError` instead of proceeding incoherently. - CLI `--imgsz` widened from int to str so it can carry `480x640`. Typing `--imgsz 640` in a shell and calling `model.predict(imgsz=640)` are both unaffected; only direct Python calls into the CLI command functions need a string. Three changes move metrics at default settings: the COCO backend, YOLOX BatchNorm eps on checkpoints trained before this release, and D-FINE's per-size multi-scale recipe. All of it, with before and after code, is on [upgrading to 1.5.0](/docs/v1.5.0/upgrade). ### New tasks The task vocabulary went from 13 entries to 17, with nothing removed. Each new task ships an original-canvas result payload, visualization, dataset schema and filename suffix. [`edge`](/docs/v1.5.0/tasks/edge-detection) and [`normal`](/docs/v1.5.0/tasks/surface-normals) are dense-prediction contracts, with validators for edge ODS/OIS and normal angular error. [`embed`](/docs/v1.5.0/tasks/face-recognition) produces L2-normalized image and region embeddings. It ships `LibreFaceEmbedder`, a `Gallery` and `FaceGallery` API, and 1:N identification from `libreyolo predict` through `--gallery` and `--gallery-threshold`. The task is deliberately general: identity vectors are one use, region embeddings and re-identification are others. [`mesh`](/docs/v1.5.0/tasks/body-mesh) is body mesh recovery. `LibreSAM3DBody` is an optional model gated on the `sam_3d_body` dependency, so it is not exported from the top-level package. ### New model families 28 new user-facing classes, exported under 29 names. All are inference-only unless stated. Dome-DETR (`domedetr`) is the trainable addition: a tiny-object detector for aerial, drone and remote-sensing imagery, ported from the Dome-DETR release. It is D-FINE plus a density head (DeFE), encoder attention restricted to occupied windows (MWAS), and a query count set by local density rather than a fixed 300 (PAQI). Sizes s, m and l at 800x800, with a maximum absolute difference of 0.0 against upstream on all six published checkpoints. Training is wired against upstream's full objective, though the published 160-epoch schedule has not been reproduced, so the paper's AP numbers are unverified. Its advantage narrows as objects grow, so it sits beside D-FINE rather than replacing it. There is no COCO checkpoint upstream, only AI-TOD-V2 with 9 classes and VisDrone with 12, so every canonical filename carries a dataset suffix. Export raises instead of emitting a graph, because PAQI's per-image query count makes a traced graph valid only for the image it was traced on. Weights are not rehosted: the upstream card claims Apache-2.0 while also restricting use to academic research. [LingBot-Vision](/docs/v1.5.0/models/lingbot-vision) is the other trainable arrival, a ViT semantic-segmentation family at 512 px pairing boundary-centric self-supervised backbones with a 1x1 dense head. Sizes s, b and l are published; `g` is the 1.1B teacher and has no LibreYOLO-hosted checkpoint, so requesting it raises with the reason. The rest are inference-only ports of earlier published work, each checked against its pinned upstream source: - Detection: [DETR](/docs/v1.5.0/models/detr), [Deformable DETR](/docs/v1.5.0/models/deformable-detr), [DINO-DETR](/docs/v1.5.0/models/dino-detr), [LW-DETR](/docs/v1.5.0/models/lw-detr), [Faster R-CNN](/docs/v1.5.0/models/faster-rcnn), [Mask R-CNN](/docs/v1.5.0/models/mask-rcnn), [FCOS](/docs/v1.5.0/models/fcos), [RetinaNet](/docs/v1.5.0/models/retinanet), [SSD300](/docs/v1.5.0/models/ssd), [CenterNet](/docs/v1.5.0/models/centernet) and [EfficientDet](/docs/v1.5.0/models/efficientdet). - Classification: [AlexNet](/docs/v1.5.0/models/alexnet), [VGG](/docs/v1.5.0/models/vgg), [ViT](/docs/v1.5.0/models/vit), [DeiT](/docs/v1.5.0/models/deit) and [Swin](/docs/v1.5.0/models/swin). - Semantic segmentation: [FCN](/docs/v1.5.0/models/fcn) and [DeepLabv3](/docs/v1.5.0/models/deeplabv3). - Depth: [MiDaS](/docs/v1.5.0/models/midas). Pose: [HRNet](/docs/v1.5.0/models/hrnet), top-down on COCO-17 person crops. - Surface normals: [MoGe-2](/docs/v1.5.0/models/moge-2) at 518 px, a DINOv2 patch grid with letterbox preprocessing. - Edge: [TEED](/docs/v1.5.0/models/teed) and [DexiNed](/docs/v1.5.0/models/dexined), native MIT-licensed architectures with local checkpoint converters. Their upstream BIPED-trained checkpoints are not bundled, mirrored or auto-downloaded, because that dataset's terms are non-commercial. - Embeddings: [LibreFaceEmbedder](/docs/v1.5.0/models/librefacerec), ONNX Runtime only. [LibreMODUS](/docs/v1.5.0/models/libremodus) adds 14B-A7B analysis-only inference for depth, normals, edges and COCO detection from one checkpoint, plus phrase grounding and image-conditioned `any2any()` chaining. [LibreFeyNobg](/docs/v1.5.0/models/feynobg) adds a matte family at 1024 px, code and weights Apache-2.0. [RT-DETRv2](/docs/v1.5.0/models/rt-detr) gained [oriented-object detection](/docs/v1.5.0/tasks/oriented-detection) inference for the official DOTA 1.0 checkpoints in sizes n, s, m, l and x, with aspect-preserving preprocessing, native `Results.obb` output, validation, and validated ONNX and TorchScript export. ### Torch-free ONNX inference `import libreyolo` no longer pulls torch. Model and results names resolve through a lazy `__getattr__`, and a new top-level `libreyolo.preprocess` package provides numpy-native preprocessing for deim, deimv2, dfine, ec, rfdetr, rtdetr, yolo9, yolonas and yolox, so an ONNX Runtime install needs no torch at all. See [lightweight install](/docs/v1.5.0/lightweight-install). ### Export and serving Five new export formats: [RKNN](/docs/v1.5.0/export/rknn) for Rockchip, with `--name` for the target platform and `--verify` for PC-simulator versus ONNX Runtime parity; [MNN](/docs/v1.5.0/export/mnn); [Paddle](/docs/v1.5.0/export/paddle); [ExecuTorch](/docs/v1.5.0/export/executorch); and [Core AI](/docs/v1.5.0/export/coreai) on macOS. [DeepStream](/docs/v1.5.0/export/deepstream) sidecar configs are written by `deepstream=True` on [ONNX](/docs/v1.5.0/export/onnx) export. It is an ONNX-only option rather than a `format=` key, and is mutually exclusive with `nms=True`. ExecuTorch and MNN exports now require a checkpoint sidecar, `.pte.json` and `.mnn.json`. Two new inference backends: [Triton](/docs/v1.5.0/export/triton), with `create_triton_config` for NVIDIA Triton Inference Server, and Paddle Inference. ### New CLI commands `libreyolo enroll` builds an embedding gallery from a folder-per-person tree, and `libreyolo compare`, aliased as `libreyolo verify`, checks two images by cosine similarity. See the [CLI reference](/docs/v1.5.0/cli). ### Speed CUDA graph capture of the training step went from 2 families to 24, across detect, classify, semantic, point and restore. Measured on an RTX 5070 Ti under AMP: 3.63x on FOMO, 2.74x on MobileNetV4, 1.99x on YOLO9-t, and 1.04x to 1.26x for the rest, with the win tracking how much of a step is network rather than loss. End to end on a 20-epoch YOLO9-t fine-tune of 406 images, dataloader and validation included, 428 s became 368 s with identical mAP50-95 and per-epoch losses. Capture at prediction time reaches 39 families, spanning detect, segment, pose, point, classify, semantic, depth, restore, matte and OCR. Every enabled family is verified to replay bit-identically against two probe inputs. A family that cannot be captured whole is split at a verified seam, and the remainder runs eagerly with identical numbers. `pip install libreyolo[hub-kernels]` opts every Deformable-DETR-lineage family into a compiled Apache-2.0 CUDA kernel for multi-scale deformable attention, pinned to an audited revision. It applies to eager CUDA fp32 only; exports keep the portable path. Fused scaled dot product attention now runs across the transformer families using stock torch, with no optional dependency. Measured on an RTX 5070 Ti under fp16 autocast, roughly 1.8x on Swin window attention and 3.7x on OWLv2 vision attention. Families held to a byte-exact parity bar keep manual attention by default and opt in explicitly, and export graphs keep the primitive-op equation either way. COCO metrics moved to faster-coco-eval by default: 15.6x faster overall and 56x on detection-dense datasets, decided on measured parity across all 100 RF100-VL test splits, where 1381 of 1400 metric values were bit-identical to pycocotools, the maximum deviation was 2.22e-16, and headline deltas were exactly 0. `--no-faster-coco-eval` opts out, and pycocotools stays the automatic fallback when the package is missing. ### Training and validation Training from scratch (`pretrained=False`) now works for every g0, g1 and g2 family through a seeded random init. Previously only yolo9, rfdetr and dfine had a scratch path, and the other families silently loaded the pretrained checkpoint anyway. Rectangular input arrived: `imgsz=480x640` for prediction and validation everywhere it makes sense, and rectangular *training* for the seven CNN detect families, yolo9, yolo9_e2e, yolo9_p2, yolox, yolo7, rtmdet and picodet. Transformer families and non-detect tasks raise a clear error, and both dimensions must divide the family stride. Autobatch and DDP spawn honor the rectangular size when probing. `amp_dtype` selects bfloat16 for [training](/docs/v1.5.0/train) and [validation](/docs/v1.5.0/train/validation), on `TrainConfig`, `ValidationConfig` and as `--amp-dtype` on `train`, `val` and `profile`. GradScaler is skipped for bf16. Validation caps are configurable too: `max_det` defaults to 300, and `eval_max_det` decouples the COCO evaluator cap from NMS. `val_loss=True` reached every trainable family that can support it, across detect, classify, semantic and restore rather than the flagships alone. It moved from `YOLO9Config` and `RFDETRConfig` to `TrainConfig`, so a family that has not implemented it now raises a clear error instead of ignoring the flag. Denoising terms are never included, because validation forwards without ground truth. Comet, ClearML, Neptune and DVCLive joined the built-in training [loggers](/docs/v1.5.0/train/loggers), with the same canonical metrics and failure-isolation contract as the existing TensorBoard, MLflow and Weights & Biases integrations. The kernel registry moved to `libreyolo/kernels/`, organized by purpose. `LIBREYOLO_KERNELS` replaces `LIBREYOLO_QUANT_KERNELS`, which is still honored, and `libreyolo.quant.kernels` remains a working alias. ### Predict sources Webcams by index, RTSP and HTTP streams, screen capture, and `s3://` and `gs://` URLs, with `--stream`, `--stream-buffer`, `--vid-stride` and `--show`. Live sources implicitly enable streaming, which emits one JSON record per frame. See [prediction sources](/docs/v1.5.0/predict/sources). ### Fixes `train(..., cuda_graph=True)` had been a silent no-op for D-FINE, DEIM, DEIMv2, RT-DETRv4 and EC, whose trainers called the eager forward instead of the routed one, so capture never engaged. The end-to-end suite now asserts that capture actually engages per family. YOLOX applies its BatchNorm `eps=1e-3` and `momentum=0.03` at construction instead of as a fixup afterwards, so the values survive the class-count rebuild that `train()` performs when a dataset's class count differs from the checkpoint. A fine-tune previously trained and reported in-training validation at torch's default `eps=1e-5` but was reloaded for inference at `1e-3`. On RF100-VL `ball`, the same nano checkpoint scores 0.566 mAP50-95 evaluated at its trained eps and 0.151 after a stock reload. Checkpoints trained before this fix carry the old semantics and need the override described on [upgrading](/docs/v1.5.0/upgrade). Non-strict checkpoint loads now warn with the counts and first names of missing and unexpected state-dict keys. Shape mismatches always raised, but name mismatches let a partially matching checkpoint load and then predict with freshly initialized tensors, leaving no trace. Auto-conversion no longer risks the file it writes: the new `.pt` is staged and atomically renamed rather than written in place, and the original file mode is preserved instead of collapsing to owner-only. Interrupted weight downloads record a validator for `If-Range` resume and take a cross-process lock, so a changed remote file or a second process can no longer produce corrupt weights. CUDA graph capture no longer races with DataLoader pin-memory threads, which had killed runs with a capture error from the pin-memory thread, and graphs are captured and replayed on the right device. D-FINE training now applies upstream's per-size multi-scale recipe instead of a hardcoded `base_size_repeat=3`. Several families silently produced wrong geometry and no longer do: PicoDet clipped boxes against swapped axes, RTMDet segmentation resized masks to a square derived from width only, Faster R-CNN ONNX exports placed boxes wrong on non-square images, LW-DETR could drop real detections by running top-K over unmapped COCO columns, and an exported MoGe-2 model run at an aspect ratio other than its export canvas stretched the image into wrong normal directions and now raises instead. ### Contributors **juni3227**, first-time contributor: rectangular input resolution for the convolution-based detectors, and a fix for training with YOLO-format datasets at rectangular model sizes (#649, #658). **Xuban Ceccon**, maintainer: everything else. ### Release stats 600 commits, 902 files changed, +125,701 / -3,315 lines, 88 merged pull requests, 8 issues closed. 174 new test modules, 279 to 453. 38 new documentation pages. ## 1.4.0, released 2026-07-24 The release notes summarize it as 15 new model families, 3 new tasks, a quantization stack, two new trackers, and a multi-GPU training correctness overhaul. The new families were SegFormer, SwinIR, Real-ESRGAN, BiRefNet, ZipDepth, Depth Anything 3, PP-OCRv5, SigLIP 2, YOLOv1, SAM 3, EdgeTAM, PicoSAM3, OMDet-Turbo, OV-DEIM and SenseNova-Vision. The new tasks were `panoptic`, `matte` and `ocr`, each with its result type and validator, and EoMT gained instance and panoptic segmentation. Quantization arrived as the `libreyolo quantize` command and `model.quantize()`, with fp16, bf16, fp8, int8, w4a16, w4a8, nvfp4 and mxfp4 recipes, and quantization-aware training through `train()` on a quantized checkpoint, for yolo9 and rfdetr. Tracking gained BoT-SORT and Deep OC-SORT with an OSNet-AIN re-identification embedder. The multi-GPU work was about correctness. DDP now shards correctly for DEIM, D-FINE and YOLO-NAS-pose, where every rank had been training the full dataset at the full batch, and loss normalizers are globally all-reduced to match single-GPU gradients. SyncBatchNorm defaults on for the CNN detectors, and a non-divisible global batch or a custom loader that does not shard now raises at setup instead of running wrong. Several training fixes changed results outright: RTMDet fine-tune collapse from missing head init took an nc=1 rebuild from 0.26 to 0.709 mAP50-95, YOLOv7 training color-space and fp16 overflow bugs took a fine-tune from 0.0 to 0.92 mAP50, and PicoDet's and DEIM's learning-rate defaults were lowered because the old ones destroyed pretrained weights. Segmentation training no longer exhausts RAM on COCO-scale datasets with multiple workers. One compatibility note from that release: checkpoints written with the new task strings or with finalized quantization state cannot be loaded by 1.3.1. ## Earlier releases Releases before 1.4.0 are documented in the [GitHub releases](https://github.com/LibreYOLO/libreyolo/releases) only, not in `CHANGELOG.md`. Their documentation is still online, listed on [versions](/docs/v1.5.0/versions).