feat/bump #15

Hojland · 2022-10-17T12:57:50Z

Fix2 select_device() for Multi-GPU (Fix2 select_device() for Multi-GPU ultralytics/yolov5#6461)
Add Product Hunt social media icon (Add Product Hunt social media icon ultralytics/yolov5#6464)
Resolve dataset paths (Resolve dataset paths ultralytics/yolov5#6489)
Simplify TF normalized to pixels (Simplify TF normalized to pixels ultralytics/yolov5#6494)
Improved export.py usage examples (Improved export.py usage examples ultralytics/yolov5#6495)
CoreML inference fix list() -> sorted() (CoreML inference fix list() -> sorted() ultralytics/yolov5#6496)
Suppress torch.jit.TracerWarning on export (Suppress torch.jit.TracerWarning on export ultralytics/yolov5#6498)
Suppress export.run() TracerWarnings (Suppress export.run() TracerWarning ultralytics/yolov5#6499)
W&B: Remember batchsize on resuming (W&B: Remember batch_size on resuming ultralytics/yolov5#6512)
Update hyp.scratch-high.yaml (Update hyp.scratch-high.yaml lrf: 0.1 ultralytics/yolov5#6525)
TODO issues exempt from stale action (TODO issues exempt from stale action ultralytics/yolov5#6530)
Update val_batch*.jpg for Chinese fonts (Update val_batch*.jpg for Chinese fonts ultralytics/yolov5#6526)
Social icons after text (Social icons after text ultralytics/yolov5#6473)
Edge TPU compiler sudo fix (Edge TPU compiler sudo fix ultralytics/yolov5#6531)
Edge TPU export 'list index out of range' fix (Edge TPU export 'list index out of range' fix ultralytics/yolov5#6533)
Edge TPU tf.lite.experimental.load_delegate fix (Edge TPU tf.lite.experimental.load_delegate fix ultralytics/yolov5#6536)
Fixing minor multi-streaming issues with TensoRT engine (Fixing minor multi-streaming issues with TensoRT engine ultralytics/yolov5#6504)
Load checkpoint on CPU instead of on GPU (Load checkpoint on CPU instead of on GPU ultralytics/yolov5#6516)
flake8: code meanings (flake8: code meanings ultralytics/yolov5#6481)
Fix 6 Flake8 issues (Fix 6 Flake8 issues ultralytics/yolov5#6541)
Edge TPU TF imports fix (Edge TPU TF imports fix ultralytics/yolov5#6542)
Move trainloader functions to class methods (Move trainloader functions to class methods ultralytics/yolov5#6559)
Improved AutoBatch DDP error message (Improved AutoBatch DDP error message ultralytics/yolov5#6568)
Fix zero-export handling with if any(f): (Fix zero-export handling with if any(f): ultralytics/yolov5#6569)
Fix plot_labels() colored histogram bug (Fix plot_labels() colored histogram bug ultralytics/yolov5#6574)
Allow custom --evolve project names (Allow custom --evolve project names ultralytics/yolov5#6567)
Add DATASETS_DIR global in general.py (Add DATASETS_DIR global in general.py ultralytics/yolov5#6578)
return opt from train.run() (return opt from train.run() ultralytics/yolov5#6581)
Fix YouTube dislike button bug in pafy package (Fix YouTube dislike button bug in pafy package ultralytics/yolov5#6603)
Update train.py
Fix hyp_evolve.yaml indexing bug (Fix hyp_evolve.yaml indexing bug ultralytics/yolov5#6604)
Fix ROOT / data when running W&B log_dataset() (Fix ROOT / data when running W&B log_dataset() ultralytics/yolov5#6606)
YouTube dependency fix youtube_dl==2020.12.2 (YouTube dependency fix youtube_dl==2020.12.2 ultralytics/yolov5#6612)
Add YOLOv5n to Reproduce section (Add YOLOv5n to Reproduce section ultralytics/yolov5#6619)
W&B: Improve resume stability (W&B: Improve resume stability ultralytics/yolov5#6611)
W&B: don't log media in evolve (W&B: don't log media in evolve ultralytics/yolov5#6617)
YOLOv5 Export Benchmarks (YOLOv5 CPU Export Benchmarks ultralytics/yolov5#6613)
Fix ConfusionMatrix scale vmin=0.0 (Fix ConfusionMatrix scale vmin=0.0 ultralytics/yolov5#6638)
Fixed wandb logger KeyError (Fixed wandb logger KeyError ultralytics/yolov5#6637)
Fix yolov3.yaml remove list (Fix yolov3.yaml remove list ultralytics/yolov5#6655)
Validate with 2x --workers (Validate with 2x --workers ultralytics/yolov5#6658)
Validate with 2x --workers single-GPU/CPU fix (Validate with 2x --workers single-GPU/CPU fix ultralytics/yolov5#6659)
Add --cache val (Add --cache val option ultralytics/yolov5#6663)
Robust scipy.cluster.vq.kmeans too few points (Robust scipy.cluster.vq.kmeans too few points ultralytics/yolov5#6668)
Update Dockerfile torch==1.10.2+cu113 (Update Dockerfile torch==1.10.2+cu113 ultralytics/yolov5#6669)
FROM nvcr.io/nvidia/pytorch:22.01-py3 (FROM nvcr.io/nvidia/pytorch:22.01-py3 ultralytics/yolov5#6670)
FROM nvcr.io/nvidia/pytorch:21.10-py3 (FROM nvcr.io/nvidia/pytorch:21.10-py3 ultralytics/yolov5#6671)
Update Dockerfile reorder installs (Update Dockerfile reorder installs ultralytics/yolov5#6672)
FROM nvcr.io/nvidia/pytorch:21.10-py3 (FROM nvcr.io/nvidia/pytorch:22.01-py3 ultralytics/yolov5#6673)
FROM nvcr.io/nvidia/pytorch:21.10-py3 (FROM nvcr.io/nvidia/pytorch:21.10-py3 ultralytics/yolov5#6677)
Fix TF exports >= 2GB (Fix TF exports >= 2GB ultralytics/yolov5#6292)
Fix --evolve --bucket gs://... (Fix --evolve --bucket gs://... ultralytics/yolov5#6698)
Fix CoreML P6 inference (Fix CoreML P6 inference ultralytics/yolov5#6700)
Fix floating point in number of workers nw (Fix floating point in number of workers ultralytics/yolov5#6701)
Edge TPU inference fix (Edge TPU inference fix ultralytics/yolov5#6686)
Use export_formats() in export.py (Use export_formats() in export.py ultralytics/yolov5#6705)
Suppress torch AMP-CPU warnings (Suppress torch AMP-CPU warnings ultralytics/yolov5#6706)
Update nw to max(nd, 1) (Update nw to max(nd, 1) ultralytics/yolov5#6714)
GH: add PR template (GH: add PR template ultralytics/yolov5#6482)
Switch default LR scheduler from cos to linear (Switch default LR scheduler from cos to linear ultralytics/yolov5#6729)
Updated VOC hyperparameters (Updated VOC hyperparameters ultralytics/yolov5#6732)
YOLOv5 v6.1 release (YOLOv5 v6.1 release ultralytics/yolov5#6739)
Pre-commit table fix (Pre-commit table fix ultralytics/yolov5#6744)
Update tutorial.ipynb (2 CPUs, 12.7 GB RAM, 42.2/166.8 GB disk) (Update tutorial.ipynb (2 CPUs, 12.7 GB RAM, 42.2/166.8 GB disk) ultralytics/yolov5#6767)
Update min warmup iterations from 1k to 100 (Update min warmup iterations from 1k to 100 ultralytics/yolov5#6768)
Default OMP_NUM_THREADS=8 (Default OMP_NUM_THREADS=8 ultralytics/yolov5#6770)
Update tutorial.ipynb (Update tutorial.ipynb ultralytics/yolov5#6771)
Update hyp.VOC.yaml (Update hyp.VOC.yaml ultralytics/yolov5#6772)
Fix export for 1-channel images (Fix export for 1-channel images ultralytics/yolov5#6780)
Update EMA decay tau (Update EMA decay tau ultralytics/yolov5#6769)
YOLOv5s6 params FLOPs fix (YOLOv5s6 params FLOPs fix ultralytics/yolov5#6782)
Update PULL_REQUEST_TEMPLATE.md (Update PULL_REQUEST_TEMPLATE.md ultralytics/yolov5#6783)
Update autoanchor.py (Update autoanchor.py ultralytics/yolov5#6794)
Update sweep.yaml (Update sweep.yaml ultralytics/yolov5#6825)
AutoAnchor improved initialization robustness (AutoAnchor improved initialization robustness ultralytics/yolov5#6854)
Add *.ts to VID_FORMATS (Add *.ts to VID_FORMATS ultralytics/yolov5#6859)
Update --cache disk deprecate *_npy/ dirs (Update --cache disk deprecate *_npy/ dirs ultralytics/yolov5#6876)
Update yolov5s.yaml (Update yolov5s.yaml ultralytics/yolov5#6865)
Default FP16 TensorRT export (Default FP16 TensorRT export ultralytics/yolov5#6798)
Bump actions/setup-python from 2 to 3 (Bump actions/setup-python from 2 to 3 ultralytics/yolov5#6880)
Bump actions/checkout from 2 to 3 (Bump actions/checkout from 2 to 3 ultralytics/yolov5#6881)
Fix TRT max_workspace_size deprecation notice (Fix TRT max_workspace_size deprecation notice ultralytics/yolov5#6856)
Update bytes to GB with bitshift (Update bytes to GB with bit shift ultralytics/yolov5#6886)
Move git_describe() to general.py (Move git_describe() to general.py ultralytics/yolov5#6918)
PyTorch 1.11.0 compatibility updates (PyTorch 1.11.0 compatibility updates ultralytics/yolov5#6932)
Optimize PyTorch 1.11.0 compatibility update (Optimize PyTorch 1.11.0 compatibility update ultralytics/yolov5#6933)
Allow 3-point segments (Allow 3-point segments ultralytics/yolov5#6938)
Fix PyTorch Hub export inference shapes (Fix PyTorch Hub export inference shapes ultralytics/yolov5#6949)
DetectMultiBackend() --half handling (DetectMultiBackend() --half handling ultralytics/yolov5#6945)
Update Dockerfile torch==1.11.0+cu113 (Update Dockerfile torch==1.11.0+cu113 ultralytics/yolov5#6954)
New val.py cuda variable (New val.py cuda variable ultralytics/yolov5#6957)
DetectMultiBackend() return device update (DetectMultiBackend() return device update ultralytics/yolov5#6958)
Tensor initialization on device improvements (Tensor initialization on device improvements ultralytics/yolov5#6959)
EdgeTPU optimizations (EdgeTPU optimizations ultralytics/yolov5#6808)
Model ema key backward compatibility fix (Model ema key backward compatibility fix ultralytics/yolov5#6972)
pt model to cpu on TF export
YOLOv5 Export Benchmarks for GPU (YOLOv5 Export Benchmarks for GPU ultralytics/yolov5#6963)
Update TQDM bar format (Update TQDM bar format ultralytics/yolov5#6988)
Conditional Timeout() by OS (disable on Windows) (Conditional Timeout() by OS (disable on Windows) ultralytics/yolov5#7013)
fix: add default PIL font as fallback (fix: add default PIL font as fallback ultralytics/yolov5#7010)
Consistent saved_model output format (Consistent saved_model output format ultralytics/yolov5#7032)
ComputeLoss() indexing/speed improvements (ComputeLoss() indexing/speed improvements ultralytics/yolov5#7048)
Update Dockerfile to git clone instead of COPY (Update Dockerfile to git clone instead of COPY ultralytics/yolov5#7053)
Create SECURITY.md (Create SECURITY.md ultralytics/yolov5#7054)
Fix incomplete URL substring sanitation (Fix incomplete URL substring sanitation ultralytics/yolov5#7056)
Use PIL to eliminate chroma subsampling in crops (Use PIL to eliminate chroma subsampling in crops ultralytics/yolov5#7008)
Fix check_anchor_order() in pixel-space not grid-space (Fix check_anchor_order() in pixel-space not grid-space ultralytics/yolov5#7060)
Update detect.py non-inplace with y.tensor_split() (Update detect.py non-inplace with y.tensor_split() ultralytics/yolov5#7062)
Update common.py lists for tuples (Update common.py lists for tuples ultralytics/yolov5#7063)
Update W&B message to LOGGER.info() (Update W&B message to LOGGER.info() ultralytics/yolov5#7064)
Update init.py (Update __init__.py ultralytics/yolov5#7065)
Add non-zero da check_anchor_order() condition (Add non-zero da check_anchor_order() condition ultralytics/yolov5#7066)
Fix2 check_anchor_order() in pixel-space not grid-space (Fix2 check_anchor_order() in pixel-space not grid-space ultralytics/yolov5#7067)
Revert "Update detect.py non-inplace with y.tensor_split() (Update detect.py non-inplace with y.tensor_split() ultralytics/yolov5#7062)" (Revert "Update detect.py non-inplace with y.tensor_split()" ultralytics/yolov5#7074)
Update loss.py with if self.gr < 1: (Update loss.py with if self.gr < 1: ultralytics/yolov5#7087)
Update loss for FP16 tobj (Update loss for FP16 tobj ultralytics/yolov5#7088)
Update model summary to display model name (Display model name in model summary ultralytics/yolov5#7101)
torch.split() 1.7.0 compatibility fix (torch.split() 1.7.0 compatibility fix ultralytics/yolov5#7102)
Update benchmarks significant digits (Update benchmarks significant digits ultralytics/yolov5#7103)
Model summary pathlib fix (Model summary pathlib fix ultralytics/yolov5#7104)
Remove named arguments where possible (Remove named arguments where possible ultralytics/yolov5#7105)
Multi-threaded VisDrone and VOC downloads (Multi-threaded VisDrone and VOC downloads ultralytics/yolov5#7108)
np.fromfile() Chinese image paths fix (np.fromfile() Chinese image paths fix ultralytics/yolov5#6979)
Add PyTorch Hub results.save(labels=False) option (Add PyTorch Hub results.save(labels=False) option ultralytics/yolov5#7129)
Fix cv2.imwrite on non-ASCII paths (Fix cv2.imwrite on non-ASCII paths ultralytics/yolov5#7139)
Fix detect.py --view-img for non-ASCII paths (Fix detect.py --view-img for non-ASCII paths ultralytics/yolov5#7093)
Add Architecture Summary to README Tutorials (Add Architecture Summary to README Tutorials ultralytics/yolov5#7146)
Adjust NMS time limit warning to batch size (Adjust NMS time limit warning to batch size ultralytics/yolov5#7156)
Sidestep os.path.relpath() Windows bug (Sidestep os.path.relpath() Windows bug ultralytics/yolov5#7158)
NMS unused variable fix (NMS unused variable fix ultralytics/yolov5#7161)
yolo.py --profile default GPU batch size 16
yolo.py --profile updates (yolo.py --profile updates ultralytics/yolov5#7170)
Update
Revert C3() change (Revert C3() change ultralytics/yolov5#7172)
Bump actions/cache from 2.1.7 to 3 (Bump actions/cache from 2.1.7 to 3 ultralytics/yolov5#7175)
yolo.py profiling updates (yolo.py profiling updates ultralytics/yolov5#7178)
Update tutorial.ipynb (Update tutorial.ipynb ultralytics/yolov5#7212)
ENV OMP_NUM_THREADS=8 (ENV OMP_NUM_THREADS=8 ultralytics/yolov5#7215)
Add train.py --name cfg option (Add train.py --name cfg option ultralytics/yolov5#7202)
precommit: yapf (precommit: yapf ultralytics/yolov5#5494)
CLI fire prep updates (CLI fire prep updates ultralytics/yolov5#7229)
Update .pre-commit-config.yaml (Update .pre-commit-config.yaml ultralytics/yolov5#7230)
SavedModel TF Serve Fix (SavedModel TF Serve Fix ultralytics/yolov5#7228)
Create CODE_OF_CONDUCT.md (Create CODE_OF_CONDUCT.md ultralytics/yolov5#7233)
Fix www.youtube.com hostname (Fix www.youtube.com hostname ultralytics/yolov5#7242)
Update minimum Python>=3.7.0 (Update minimum Python>=3.7.0 ultralytics/yolov5#7247)
Update setup.cfg to description_file field (Update setup.cfg to description_file field ultralytics/yolov5#7248)
Update tutorial.ipynb (Update tutorial.ipynb ultralytics/yolov5#7254)
Update tutorial.ipynb (Update tutorial.ipynb ultralytics/yolov5#7255)
Fix Flask REST API (Fix Flask REST API ultralytics/yolov5#7210)
Export with official nn.SiLU() (Export with official nn.SiLU() ultralytics/yolov5#7256)
Refactor out-of-place Detect() for reduced ops (Refactor out-of-place Detect() for reduced ops ultralytics/yolov5#7257)
torch.split() replace slicing on out-of-place inference (torch.split() replace slicing on out-of-place inference ultralytics/yolov5#7258)
Update
Export single output only (Export single output only ultralytics/yolov5#7259)
TorchScript single-output fix (TorchScript single-output fix ultralytics/yolov5#7261)
Integrate offset into grid (Integrate offset into grid ultralytics/yolov5#7262)
[pre-commit.ci] pre-commit suggestions ([pre-commit.ci] pre-commit suggestions ultralytics/yolov5#7279)
Update Dockerfile (Update Dockerfile ultralytics/yolov5#7282)
Enable TensorFlow ops for --nms and --agnostic-nms (Enable TensorFlow ops for --nms and --agnostic-nms ultralytics/yolov5#7281)
Update cv2.imread() patch with flags argument (Update cv2.imread() patch with flags argument ultralytics/yolov5#7287)
Context manager open(file) as f fixes (Context manager open(file) as f fixes ultralytics/yolov5#7289)
val.py --weights and --data compatibility check (val.py --weights and --data compatibility check ultralytics/yolov5#7292)
Add dataset sizes (zipped) (Add dataset sizes (zipped) ultralytics/yolov5#7293)
Add check_requirements(('pycocotools>=2.0',)) (Add check_requirements(('pycocotools>=2.0',)) ultralytics/yolov5#7295)
fix: disable usage of root logger (fix: disable usage of root logger ultralytics/yolov5#7296)
Update export.py (Update export.py ultralytics/yolov5#7301)
Use tqdm.auto (Use tqdm.auto ultralytics/yolov5#7311)
Add retry=3 to download() (Add retry=3 to download() ultralytics/yolov5#7313)
Add callbacks (Add callbacks ultralytics/yolov5#7315)
Copy wandb param dict before training to avoid overwrites (Copy wandb param dict before training to avoid overwrites. ultralytics/yolov5#7317)
Update Objects365.yaml (Update Objects365.yaml ultralytics/yolov5#7323)
Fix Tf export for BottleneckCSP (Fix TF export for BottleneckCSP layer ultralytics/yolov5#7330)
Objects365 dataset breakdown images vs zips (Objects365 images GB vs zips GB ultralytics/yolov5#7335)
Simplify callbacks.py return (Simplify callbacks.py return ultralytics/yolov5#7333)
Print dataset scan only if RANK in (-1, 0) (Print dataset scan only if RANK in (-1, 0) ultralytics/yolov5#7337)
Update _make_grid() (Update _make_grid() to model dtype ultralytics/yolov5#7346)
Rename 'MacOS' to 'macOS' (Rename 'MacOS' to 'macOS' ultralytics/yolov5#7349)
Add python benchmarks.py --test for export-only (Add python benchmarks.py --test for export-only ultralytics/yolov5#7350)
Add ONNX export metadata (Add ONNX export metadata ultralytics/yolov5#7353)
DetectMultiBackend() default stride=32 (DetectMultiBackend() default stride=32 ultralytics/yolov5#7342)
Loss and IoU speed improvements (Loss and IoU speed improvements ultralytics/yolov5#7361)
Swap unsafe_chunk() for chunk() (Swap unsafe_chunk() for chunk() ultralytics/yolov5#7362)
Delete FUNDING.yml (Delete FUNDING.yml ultralytics/yolov5#7363)
Replace Slack with Community Forum in issues (Replace Slack with Community Forum in issues ultralytics/yolov5#7364)
Update ci-testing.yml (Update ci-testing.yml ultralytics/yolov5#7365)
Bump actions/stale from 4 to 5 (Bump actions/stale from 4 to 5 ultralytics/yolov5#7371)
Update optimizer param group strategy (Update optimizer param group strategy ultralytics/yolov5#7376)
Add support for different normalization layers (Add support for different normalization layers ultralytics/yolov5#7377)
Dynamic normalization layer selection (Dynamic normalization layer selection ultralytics/yolov5#7392)
Add version warning for wandb (Add version warning for wandb ultralytics/yolov5#7385)
Remove OpenVINO ONNX opset<=12 check (Remove OpenVINO ONNX opset<=12 check ultralytics/yolov5#7398)
Fix EdgeTPU output directory (Fix EdgeTPU output directory ultralytics/yolov5#7399)
Update git_describe() (Update git_describe() ultralytics/yolov5#7402)
Remove tensorrt pip install check (Remove tensorrt pip install check ultralytics/yolov5#7439)
Disable pbar for DDP ranks > 0 (Disable pbar for DDP ranks > 0 ultralytics/yolov5#7440)
Add --half support for FP16 CoreML exports with (Add --half support for FP16 CoreML exports with ultralytics/yolov5#7446)
Bump cirrus-actions/rebase from 1.5 to 1.6 (Bump cirrus-actions/rebase from 1.5 to 1.6 ultralytics/yolov5#7462)
Update val.py (Update val.py ultralytics/yolov5#7478)
Improved non-latin Annotator() plotting (Improved non-latin Annotator() plotting ultralytics/yolov5#7488)
check_fonts() download to CONFIG_DIR fix (check_fonts() download to CONFIG_DIR fix ultralytics/yolov5#7489)
Fix val.py Ensemble() (Fix val.py Ensemble() ultralytics/yolov5#7490)
Added YOLOv5_AUTOINSTALL environment variable (Added YOLOv5_AUTOINSTALL environment variable ultralytics/yolov5#7505)
Refactor Dockerfiles to utils/docker (Refactor Dockerfiles to utils/docker ultralytics/yolov5#7510)
Add yesqa to precommit checks (Add yesqa to precommit checks ultralytics/yolov5#7511)
Fix val plots=plots (Fix val plots=plots ultralytics/yolov5#7524)
Reduce val device transfers (Reduce val device transfers ultralytics/yolov5#7525)
Add Docker --file argument to build (Add Docker --file argument to build ultralytics/yolov5#7527)
Empty val batch CUDA device fix (Empty val batch CUDA device fix ultralytics/yolov5#7539)
Autoinstall TensorRT if missing (Autoinstall TensorRT if missing ultralytics/yolov5#7537)
Add mdformat to precommit checks and update other version (Add mdformat to precommit checks and update other version ultralytics/yolov5#7529)
Update precommit monthly python 3.7+ (Update precommit monthly python 3.7+ ultralytics/yolov5#7542)
Update downloads.py current release (Update downloads.py current release ultralytics/yolov5#7541)
Update check_requirements() with cmds=() argument (Update check_requirements() with cmds=() argument ultralytics/yolov5#7543)
Add --noplots flag to suppress figures and images logging (Add --noplots flag to suppress figures and images logging ultralytics/yolov5#7534)
Improve availability of YOLOv5 in Russia (Improve availability of YOLOv5 in Russia ultralytics/yolov5#7545)
Update restapi.py (Update restapi.py ultralytics/yolov5#7309)
Restrict TRT autoinstall to Linux-only (Restrict TRT autoinstall to Linux-only ultralytics/yolov5#7549)
PyTorch Hub _verbose=False fix2 (PyTorch Hub _verbose=False fix2 ultralytics/yolov5#7550)
Enable results.print() when _verbose=False (Enable results.print() when _verbose=False ultralytics/yolov5#7558)
Add print(results) override for PyTorch Hub results (Add print(results) override for PyTorch Hub results ultralytics/yolov5#7559)
TensorRT PyTorch Hub inference fix (TensorRT PyTorch Hub inference fix ultralytics/yolov5#7560)
Add PyTorch-only benchmark arg (Add PyTorch-only benchmark arg ultralytics/yolov5#7564)
Fix from yolov5 import utils statement (Fix from yolov5 import utils statement ultralytics/yolov5#7578)
Remove tqdm.auto (Remove tqdm.auto ultralytics/yolov5#7599)
Add --half support for OpenVINO exports (Add --half support for OpenVINO exports ultralytics/yolov5#7615)
increment_path() robustness improvements (increment_path() robustness improvements ultralytics/yolov5#7628)
Reduce opencv-python>=4.1.1 for Jetson Nano (Reduce opencv-python>=4.1.1 for Jetson Nano ultralytics/yolov5#7645)
Bump scipy>=1.5 to meet numpy constraints (Bump scipy>=1.5 to meet numpy constraints ultralytics/yolov5#7646)
Update plot.py (Update plot.py ultralytics/yolov5#7654)
Bump github/codeql-action from 1 to 2 (Bump github/codeql-action from 1 to 2 ultralytics/yolov5#7665)
Disallow --dynamic when --half is passed (Disallow --dynamic when --half is passed ultralytics/yolov5#7669)
Update Dockerfile-cpu to force python3.9 (Update Dockerfile-cpu to force python3.9 ultralytics/yolov5#7675)
Pin Docker-cpu FROM ubuntu:20.04 (Pin Docker-cpu FROM ubuntu:20.04 ultralytics/yolov5#7677)
FROM nvcr.io/nvidia/pytorch:22.04-py3 (FROM nvcr.io/nvidia/pytorch:22.04-py3 ultralytics/yolov5#7680)
Update README.md (Update README.md ultralytics/yolov5#7686)
Update tutorial.ipynb (Update tutorial.ipynb ultralytics/yolov5#7715)
Update on_train_end callback (Update on_train_end callback ultralytics/yolov5#7716)
Report fused model summary by default (Report fused model summary by default ultralytics/yolov5#7722)
Ability to dowlnoad older assets (Ability to download older assets ultralytics/yolov5#7767)
Pin downloads to release version (Pin downloads to release version ultralytics/yolov5#7790)
Implement DDP static_graph=True (Implement DDP static_graph=True ultralytics/yolov5#6940)
Add smoothing to curve plots and max F1 index (Add smoothing to curve plots and max F1 index ultralytics/yolov5#7798)
Add Dockerfile-M1 (Add Dockerfile-M1 ultralytics/yolov5#7720)
Add random interpolation method augmentation (Add random interpolation method augmentation ultralytics/yolov5#6826)
Rename utils/datasets.py > utils/dataloaders.py (Rename utils/datasets.py -> utils/dataloaders.py ultralytics/yolov5#7799)
Add @threaded decorator (Add @threaded decorator ultralytics/yolov5#7813)
Add Python version output (Add Python version output ultralytics/yolov5#7814)
Refactor collections and fstrings (Refactor collections and fstrings ultralytics/yolov5#7821)
Refactor modules (Refactor modules ultralytics/yolov5#7823)
YOLOv5 OpenVINO PyTorch Hub inference fix (YOLOv5 OpenVINO PyTorch Hub inference fix ultralytics/yolov5#7826)
New TensorFlow TFDWConv() module (New TensorFlow TFDWConv() module ultralytics/yolov5#7824)
Bump cirrus-actions/rebase from 1.6 to 1.7 (Bump cirrus-actions/rebase from 1.6 to 1.7 ultralytics/yolov5#7831)
New TensorFlow TFCrossConv() module (New TensorFlow TFCrossConv() module ultralytics/yolov5#7827)
Fix TFDWConv() c1 == c2 check (Fix TFDWConv() c1 == c2 check ultralytics/yolov5#7842)
Replace openvino-dev with OpenVINO Runtime inference (Replace openvino-dev with OpenVINO Runtime inference ultralytics/yolov5#7843)
TFDWConv() depthwise_initializer fix (TFDWConv() depthwise_initializer fix ultralytics/yolov5#7845)
Update Dockerfile --no-install-recommends (Update Dockerfile --no-install-recommends ultralytics/yolov5#7846)
Update CI CPU badge (Update CI CPU badge ultralytics/yolov5#7855)
Create docker.yml (Create docker.yml ultralytics/yolov5#7856)
Update Dockerfile-cpu install libpython3.8-dev (Update Dockerfile-cpu install libpython3.8-dev ultralytics/yolov5#7857)
Add TFDWConv() depth_multiplier (Add TFDWConv() depth_multiplier arg ultralytics/yolov5#7858)
Update Dockerfile-arm64 (Update Dockerfile-arm64 ultralytics/yolov5#7860)
Invert Docker Image publishing order (Invert Docker Image publishing order ultralytics/yolov5#7877)
Removed shell=True from subprocess commands that require user inputs (Removed shell=True from subprocess commands that require user inputs ultralytics/yolov5#7875)
Added Windows cmd to count GPU devices (Added Windows cmd to count GPU devices ultralytics/yolov5#7891)
Bug fix mAP0.5-0.95 (Bug fix mAP0.5-0.95 ultralytics/yolov5#6787)
cv2.resize interpolation fix (cv2.resize interpolation fix ultralytics/yolov5#7903)
Add DWConvTranspose2d() module (Add DWConvTranspose2d() module ultralytics/yolov5#7881)
Add check_yaml() to benchmarks.py (Add check_yaml() to benchmarks.py ultralytics/yolov5#7916)
Add --keras argument for TF exports (Add --keras argument for TF exports ultralytics/yolov5#7921)
Add PyTorch AMP check (Add PyTorch AMP check ultralytics/yolov5#7917)
Code refactor (Code refactor ultralytics/yolov5#7923)
AMP check image download backup (AMP check image download backup ultralytics/yolov5#7936)
Fix AMP check tolerance (Fix AMP check tolerance ultralytics/yolov5#7937)
Windows check_file() fix (Windows check_file() fix ultralytics/yolov5#7938)
Add OpenVINO metadata to export (Add OpenVINO metadata to export ultralytics/yolov5#7947)
OpenVINO metadata fix (OpenVINO metadata fix ultralytics/yolov5#7952)
OpenVINO metadata fix2 (OpenVINO metadata fix2 ultralytics/yolov5#7954)
YOLOv5 Apple Metal Performance Shader (MPS) support (YOLOv5 Apple Metal Performance Shader (MPS) support ultralytics/yolov5#7878)
Increase NMS time limit to 50 ms/img (Increase NMS time limit to 50 ms/img ultralytics/yolov5#7956)
Increase NMS fixed time limit 300ms + 30ms/img (Increase NMS fixed time limit 300ms + 30ms/img ultralytics/yolov5#7957)
AMP check improvements backup YOLOv5n pretrained (AMP check improvements backup YOLOv5n pretrained ultralytics/yolov5#7959)
Update greetings.yml (Update greetings.yml ultralytics/yolov5#7965)
Update requirements.txt with Usage example (Update requirements.txt with Usage example ultralytics/yolov5#7966)
Remove formats variable to avoid pd conflict (Remove formats variable to avoid pd conflict ultralytics/yolov5#7993)
check_requirements() Windows fix (check_requirements() Windows fix ultralytics/yolov5#7997)
Add psutil and ipython to requirements.txt (Add psutil and ipython to requirements.txt ultralytics/yolov5#7998)
cURL download double quotes (cURL download double quotes ultralytics/yolov5#7999)
Update and add benchmarks to ci-testing.yml (Update and add benchmarks to ci-testing.yml ultralytics/yolov5#7996)
Add warning emoji ⚠️ on --conf > 0.001 (Add warning emoji ⚠️ on --conf > 0.001 ultralytics/yolov5#8005)
Update CI benchmarks to ONNX autoinstall (Update CI benchmarks to ONNX autoinstall ultralytics/yolov5#8008)
Pin CI protobuf==3.20.1 (Pin CI protobuf==3.20.1 ultralytics/yolov5#8015)
check_requirements() "" Windows double quote (check_requirements() "" Windows double quote ultralytics/yolov5#8016)
Add requirements.txt protobuf<=3.20.1 (Add requirements.txt protobuf<=3.20.1 ultralytics/yolov5#8017)
Add best.pt PyTorch Hub inference to CI (Add best.pt PyTorch Hub inference to CI ultralytics/yolov5#8024)
Add Dockerfile descriptions (Add Dockerfile descriptions ultralytics/yolov5#8031)
Simplify and fix --save-period epoch 0 (Simplify and fix --save-period epoch 0 ultralytics/yolov5#8042)
.detach() on bias init (.detach() on bias init ultralytics/yolov5#8044)
attempt_load() deserialize fix (attempt_load() deserialize fix ultralytics/yolov5#8051)
Dataset autodownload fstring updates
Fix FP32 TensorRT model export (Fix FP32 TensorRT model export ultralytics/yolov5#8046)
Make Docker actions conditional on ultralytics/yolov5 repo (Make Docker actions conditional on ultralytics/yolov5 repo ultralytics/yolov5#8060)
Update stale.yml to 300 ops per run (Update stale.yml to 300 ops per run ultralytics/yolov5#8061)
Fix torch long to float tensor on HUB macOS (Fix torch long to float tensor on HUB macOS ultralytics/yolov5#8067)
Improved dataset error introspection (Improved dataset error introspection ultralytics/yolov5#8091)
experimental.py Apple MPS device fix (experimental.py Apple MPS device fix ultralytics/yolov5#8121)
[pre-commit.ci] pre-commit suggestions ([pre-commit.ci] pre-commit suggestions ultralytics/yolov5#8119)
fix(utils): missing edge when resample segments (fix(utils): missing edge when resample segments ultralytics/yolov5#8092)
AutoBatch checks against failed solutions (AutoBatch checks against failed solutions ultralytics/yolov5#8159)
FROM nvcr.io/nvidia/pytorch:22.05-py3 (FROM nvcr.io/nvidia/pytorch:22.05-py3 ultralytics/yolov5#8162)
Docker GPU pip install opencv-python<4.6.0.66 (Docker GPU pip install opencv-python<4.6.0.66 ultralytics/yolov5#8164)
Improved dataset_stats() YAML checks (Improved dataset_stats() YAML checks ultralytics/yolov5#8125)
Bump actions/setup-python from 3 to 4 (Bump actions/setup-python from 3 to 4 ultralytics/yolov5#8186)
Reject webcam inference on Colab/Kaggle (Reject webcam inference on Colab/Kaggle ultralytics/yolov5#8209)
Remove streaming source sleep period (Remove streaming source sleep period ultralytics/yolov5#8243)
Prefer MPS over CPU if available (Prefer MPS over CPU if available ultralytics/yolov5#8210)
HUB dataset_stats() error reporting (HUB dataset_stats() error reporting ultralytics/yolov5#8192)
process_batch() as numpy arrays (process_batch() as numpy arrays ultralytics/yolov5#8254)
Created using Colaboratory (Created using Colaboratory ultralytics/yolov5#8255)
logging.ERROR on RANK not in (0, 1) (logging.ERROR on RANK not in (0, 1) ultralytics/yolov5#8284)
device_name="MYRIAD" for Intel NCS2 comment (device_name="MYRIAD" for Intel NCS2 comment ultralytics/yolov5#8327)
Allow preview resize in detect.py (Allow preview resize in detect.py ultralytics/yolov5#8318)
Create README_cn.md (Create README_cn.md ultralytics/yolov5#8344)
Allow detect.py to use video size for initial window size (Allow detect.py to use video size for initial window size ultralytics/yolov5#8330)
Revamp Chinese docs (Revamp Chinese docs ultralytics/yolov5#8350)
Fix bias warmup LR init (Fix bias warmup LR init ultralytics/yolov5#8356)
Add File Size (MB) column to benchmarks (Add File Size (MB) column to benchmarks ultralytics/yolov5#8359)
Update protobuf requirement from <=3.20.1 to <4.21.3 (Update protobuf requirement from <=3.20.1 to <4.21.3 ultralytics/yolov5#8346)
Fix ONNX --dynamic export on GPU (Fix ONNX --dynamic export on GPU ultralytics/yolov5#8378)
Update tutorial.ipynb (Update tutorial.ipynb ultralytics/yolov5#8380)
Implementation of Early Stopping for DDP training (Implementation of Early Stopping for DDP training ultralytics/yolov5#8345)
Improve --local_rank arg comment (Improve --local_rank arg comment ultralytics/yolov5#8409)
Update cache comments (Update cache comments ultralytics/yolov5#8414)
TRT --half fix autocast images to FP16 (TRT --half fix autocast images to FP16 ultralytics/yolov5#8435)
Expose OpenVINO batch_size similarly to TensorRT (Expose OpenVINO batch_size similarly to TensorRT ultralytics/yolov5#8437)
Do not prefer Apple MPS (Do not prefer Apple MPS ultralytics/yolov5#8446)
Update stale.yml (Update stale.yml ultralytics/yolov5#8465)
[pre-commit.ci] pre-commit suggestions ([pre-commit.ci] pre-commit suggestions ultralytics/yolov5#8470)
Exclude torch==1.12.0, torchvision==0.13.0 (Fix YOLOv5 issues with torch==1.12 on Multi-GPU systems ultralytics/yolov5#8395) (Exclude torch==1.12.0, torchvision==0.13.0 (Fix #8395) ultralytics/yolov5#8497)
Update tutorial.ipynb (Update tutorial.ipynb ultralytics/yolov5#8507)
Training reproducibility improvements (Training reproducibility improvements ultralytics/yolov5#8213)
Revert "Expose OpenVINO batch_size similarly to TensorRT" (Revert "Expose OpenVINO batch_size similarly to TensorRT" ultralytics/yolov5#8510)
Avoid FP64 ops for MPS support in train.py (Avoid FP64 ops for MPS support in train.py ultralytics/yolov5#8511)
Fix AP calculation bug Val.py: No prediction on image follows in wrong AP calculation ultralytics/yolov5#8464 (Fix ap calculation bug #8464 ultralytics/yolov5#8484)
Add --hard-fail argument to benchmarks for CI errors (Add --hard-fail argument to benchmarks for CI errors ultralytics/yolov5#8513)
Simplify benchmarks.py assertions (Simplify benchmarks.py assertions ultralytics/yolov5#8515)
Properly expose batch_size from OpenVINO similarly to TensorRT (Properly expose batch_size from OpenVINO similarly to TensorRT ultralytics/yolov5#8514)
Add --half arguments to export.py Usage examples (Add --half arguments to export.py Usage examples ultralytics/yolov5#8516)
XML export --half fix (XML export --half fix ultralytics/yolov5#8522)
Fix LoadImages() with dataset YAML lists (Fix LoadImages() with dataset YAML lists ultralytics/yolov5#8517)
val.py assert ncm == nc fix (val.py assert ncm == nc fix ultralytics/yolov5#8545)
CIoU protected divides (CIoU protected divides ultralytics/yolov5#8546)
Update metrics.py with IoU protected divides (Update metrics.py with IoU protected divides ultralytics/yolov5#8550)
Add TensorRT dependencies (Add TensorRT dependencies ultralytics/yolov5#8553)
Add thop>=0.1.0 (Add thop>=0.1.0 ultralytics/yolov5#8558)
Raise error on suffix-less model path (Raise error on suffix-less model path ultralytics/yolov5#8561)
Assert --optimize not used with cuda device (Assert --optimize not used with cuda device ultralytics/yolov5#8569)
Update requirements.txt comment spacing (Update requirements.txt comment spacing ultralytics/yolov5#8562)
Explicitly set weight_decay value (Explicitly set weight_decay value ultralytics/yolov5#8592)
Update scipy>=1.7.3 (Update scipy>=1.7.3 ultralytics/yolov5#8595)
Update tqdm>=4.64.0 and thop>=0.1.1 (Update tqdm>=4.64.0 and thop>=0.1.1 ultralytics/yolov5#8596)
Add pip install wheel to avoid legacy setup.py install (Add pip install wheel to avoid legacy setup.py install ultralytics/yolov5#8597)
Link fuse() to AutoShape() for Hub models (Link fuse() to AutoShape() for Hub models ultralytics/yolov5#8599)
FROM nvcr.io/nvidia/pytorch:22.06-py3 (FROM nvcr.io/nvidia/pytorch:22.06-py3 ultralytics/yolov5#8600)
Refactor optimizer initialization (Refactor optimizer initialization ultralytics/yolov5#8607)
assert torch!=1.12.0 for DDP training (assert torch!=1.12.0 for DDP training ultralytics/yolov5#8621)
Fix OMP_NUM_THREADS=1 for macOS (Fix OMP_NUM_THREADS=1 for macOS ultralytics/yolov5#8624)
Upgrade onnxsim to v0.4.1 (Upgrade onnxsim to v0.4.1 ultralytics/yolov5#8632)
Check TensorBoard logger before adding graph (Check TensorBoard logger before adding graph ultralytics/yolov5#8664)
Use contextlib's suppress method to silence an error (Use contextlib's suppress method to silence an error ultralytics/yolov5#8668)
Update hubconf.py to reset LOGGER.level after load (Update hubconf.py to reset LOGGER.level after load ultralytics/yolov5#8674)
Update warning emojis (Update warning emojis ultralytics/yolov5#8678)
Update hubconf.py to reset logging level to INFO (Update hubconf.py to reset logging level to INFO ultralytics/yolov5#8680)
Add generator and worker seed (Add generator and worker seed ultralytics/yolov5#8602)
Set torch.cuda.manual_seed_all() for DDP (Set torch.cuda.manual_seed_all() for DDP ultralytics/yolov5#8688)
Move .dockerignore to root (Move .dockerignore to root ultralytics/yolov5#8690)
Faster crop saving (Faster crop saving ultralytics/yolov5#8696)
Remove else: from load_image() (Remove else: from load_image() ultralytics/yolov5#8692)
Avoid cv2 window init code on Windows (Avoid cv2 window init code on Windows ultralytics/yolov5#8712)
Update dataloaders.py (Update dataloaders.py ultralytics/yolov5#8714)
New HUBDatasetStats() class (New HUBDatasetStats() class ultralytics/yolov5#8716)
Fix BGR->RGB Bug in albumentations Albumentations-Pipeline is applied to BGR not to RGB ultralytics/yolov5#8641 (Fix BGR->RGB Bug in albumentations #8641 ultralytics/yolov5#8695)
coremltools>=5.2 for CoreML export (coremltools>=5.2 for CoreML export ultralytics/yolov5#8725)
Revert "Fix BGR->RGB Bug in albumentations Albumentations-Pipeline is applied to BGR not to RGB ultralytics/yolov5#8641" (Revert "Fix BGR->RGB Bug in albumentations #8641" ultralytics/yolov5#8727)
fix: broken is_docker check (fix: broken is_docker check ultralytics/yolov5#8711)
Revert protobuf<=3.20.1 (Revert protobuf<=3.20.1 ultralytics/yolov5#8742)
Dynamic batch size support for TensorRT (Dynamic batch size support for TensorRT ultralytics/yolov5#8526)
Fix confusion matrix update when no predictions are made (Fix confusion matrix update when no predictions are made ultralytics/yolov5#8748)
Add val.py no label warning (Add val.py no label warning ultralytics/yolov5#8782)
Fix detect.py --update list bug (Fix detect.py --update list bug ultralytics/yolov5#8783)
ci-testing.yml Windows-friendly ENV variables (ci-testing.yml Windows-friendly ENV variables ultralytics/yolov5#8794)
Add hubconf.py argparser (Add hubconf.py argparser ultralytics/yolov5#8799)
Print hubconf.py args (Print hubconf.py args ultralytics/yolov5#8800)
Update Colab Notebook CI (Update Colab Notebook CI ultralytics/yolov5#8798)
Deprecate torch 1.6.0 compat _non_persistent_buffers_set (Deprecate torch 1.6.0 compat _non_persistent_buffers_set ultralytics/yolov5#8797)
Detect.inplace=False for multithread-safe inference (Detect.inplace=False for multithread-safe inference ultralytics/yolov5#8801)
Update train.py for val.run(half=amp) (Update train.py for val.run(half=amp) ultralytics/yolov5#8804)
Fix val.py 'no labels found bug' (Fix val.py 'no labels found bug' ultralytics/yolov5#8806)
Update requirements.txt with tf-cpu and tf-aarch64 (Update requirements.txt with tf-cpu and tf-aarch64 ultralytics/yolov5#8807)
FROM nvcr.io/nvidia/pytorch:22.07-py3 (FROM nvcr.io/nvidia/pytorch:22.07-py3 ultralytics/yolov5#8808)
Update ci-testing.yml streamlined tests (Update ci-testing.yml streamlined tests ultralytics/yolov5#8809)
Check git status on upstream ultralytics or origin dynamically (Check git status on upstream ultralytics or origin dynamically ultralytics/yolov5#8694)
Fix Colab-update pre-commit EOF bug (Fix Colab-update pre-commit EOF bug ultralytics/yolov5#8810)
Update .pre-commit-config.yaml (Update .pre-commit-config.yaml ultralytics/yolov5#8811)
Created using Colaboratory
Update .pre-commit-config.yaml (Update .pre-commit-config.yaml ultralytics/yolov5#8812)
Remove assert not is_docker() from GitHub checks (Remove assert not is_docker() from GitHub checks ultralytics/yolov5#8813)
Add .git to .dockerignore (Add .git to .dockerignore ultralytics/yolov5#8815)
Add tensor hooks and 10.0 gradient clipping (Add tensor hooks and 10.0 gradient clipping ultralytics/yolov5#8598)
Update README.md with contributors.png (Update README.md with contributors.png ultralytics/yolov5#8820)
Remove hook torch.nan_to_num(x) (Remove hook torch.nan_to_num(x) ultralytics/yolov5#8826)
RUN git clone instead of COPY to /usr/src/app (RUN git clone instead of COPY to /usr/src/app ultralytics/yolov5#8827)
[pre-commit.ci] pre-commit suggestions ([pre-commit.ci] pre-commit suggestions ultralytics/yolov5#8828)
Fix missing attr model.model when loading custom yolov model (Fix missing attr model.model when loading custom yolov model ultralytics/yolov5#8830)
New smart_resume() (New smart_resume() ultralytics/yolov5#8838)
Created using Colaboratory
Created using Colaboratory
Self-contained checkpoint --resume (Self-contained checkpoint --resume ultralytics/yolov5#8839)
Add check_file(data) i.e. --data coco128.yaml (Add check_file(data) i.e. --data coco128.yaml ultralytics/yolov5#8851)
export.py replace check_file -> check_yaml (export.py replace check_file -> check_yaml ultralytics/yolov5#8852)
Update dataloaders.py remove float64 shapes (Update dataloaders.py remove float64 shapes ultralytics/yolov5#8865)
Fix TensorRT --dynamic excess outputs bug (Fix TensorRT --dynamic excess outputs bug ultralytics/yolov5#8869)
Single-line TRT dynamic assertion (Single-line TRT dynamic assertion ultralytics/yolov5#8871)
HUBDatasetStats() preview images to 50 quality (HUBDatasetStats() preview images to 50 quality ultralytics/yolov5#8880)
--resume training from URL weights (--resume training from URL weights ultralytics/yolov5#8882)
--resume training from URL weights fix (--resume training from URL weights fix ultralytics/yolov5#8884)
Update CI to default Python 3.10 (Update CI to default Python 3.10 ultralytics/yolov5#8883)
ClearML experiment tracking integration (ClearML experiment tracking integration ultralytics/yolov5#8620)
Created using Colaboratory
Created using Colaboratory
smart_optimizer() improved reporting (smart_optimizer() improved reporting ultralytics/yolov5#8887)
Created using Colaboratory
Created using Colaboratory
emoji-safe default logging (emoji-safe default logging ultralytics/yolov5#8888)
Update README_cn.md
Fix/improve social media bar (Fix/improve social media bar ultralytics/yolov5#8893)
Update README.md Environments icons (Update README.md Environments icons ultralytics/yolov5#8895)
Fix anchor incorrectly initialized (Fix anchor incorrectly initialized ultralytics/yolov5#8891)
Edge TPU add --search_delegate fix (Edge TPU add --search_delegate fix ultralytics/yolov5#8902)
Capitalize YouTube (Capitalize YouTube ultralytics/yolov5#8903)
About "os.R_OK" in general.py (About "os.R_OK" in general.py ultralytics/yolov5#8909)
Scope torchvision.utils.draw_bounding_boxes import (Scope torchvision.utils.draw_bounding_boxes import ultralytics/yolov5#8915)
Add torch==1.7.0 to CI tests (Add torch==1.7.0 to CI tests ultralytics/yolov5#8916)
Update README.md
Direct download table links (Direct download table links ultralytics/yolov5#8917)
Fix --source 'path/*.jpg' glob example (Fix --source 'path/*.jpg' glob example ultralytics/yolov5#8927)
Deci.ai updates (Deci.ai optimization and deployment Integration ultralytics/yolov5#8928)
Allow train to use zipped content (Allow train to use zipped content ultralytics/yolov5#8940)
Remove DDP destroy_process_group() on train end (Remove DDP destroy_process_group() on train end ultralytics/yolov5#8935)
GFLOPs computation fix for classification models (GFLOPs computation fix for classification models ultralytics/yolov5#8954)
New smart_inference_mode() conditional decorator (New smart_inference_mode() conditional decorator ultralytics/yolov5#8957)
Fix windows LOGGER with emojis output (Fix windows LOGGER with emojis output ultralytics/yolov5#8958)
Add bash script arguments (Add bash script arguments ultralytics/yolov5#8962)
Created using Colaboratory
Add .tar support for datasets (Add .tar support for datasets ultralytics/yolov5#8963)
Allow multiple-model serving from Flask REST API (Allow multiple-model serving from Flask REST API ultralytics/yolov5#8973)
Created using Colaboratory
Created using Colaboratory
Colab and Kaggle loggers removeHandler (Colab and Kaggle loggers removeHandler ultralytics/yolov5#8985)
Created using Colaboratory
Fix bug with resume (Fix bug with resume ultralytics/yolov5#8912)
Fix torchvision dependency for ClearML logging (Fix torchvision dependency for ClearML logging ultralytics/yolov5#8993)
New YOLOv5 Classification Models (New YOLOv5 Classification Models ultralytics/yolov5#8956)
Improve classification comments (Improve classification comments ultralytics/yolov5#8997)
Update attempt_download(release='v6.2') (Update attempt_download(release='v6.2') ultralytics/yolov5#8998)
Created using Colaboratory
Update README_cn.md (Update README_cn.md ultralytics/yolov5#9001)
Update dataset names from array to dictionary (Update dataset names from array to dictionary ultralytics/yolov5#9000)
[Classify]: Allow inference on dirs and videos ([Classify]: Allow inference on dirs and videos ultralytics/yolov5#9003)
DockerHub tag update Usage example (DockerHub tag update Usage example ultralytics/yolov5#9005)
Add weight decay to argparser (Add weight decay to argparser ultralytics/yolov5#9006)
Add glob quotes to detect.py usage example (Add glob quotes to detect.py usage example ultralytics/yolov5#9007)
requires grad after reset params
Fix TorchScript JSON string key bug (Fix TorchScript JSON string key bug ultralytics/yolov5#9015)
EMA FP32 assert classification bug fix (EMA FP32 assert classification bug fix ultralytics/yolov5#9016)
Faster pre-processing for gray image input (Faster pre-processing for gray image input ultralytics/yolov5#9009)
Improved Profile() inference timing (Improved Profile() inference timing ultralytics/yolov5#9024)
torch.empty() for speed improvements (torch.empty() for speed improvements ultralytics/yolov5#9025)
Created using Colaboratory
Remove unused time_sync import (Remove unused time_sync import ultralytics/yolov5#9026)
Add PyTorch Hub classification CI checks (Add PyTorch Hub classification CI checks ultralytics/yolov5#9027)
Created using Colaboratory
Attach transforms to model (Attach transforms to model ultralytics/yolov5#9028)
Created using Colaboratory
Default --data imagenette160 training (fastest) (Default --data imagenette160 training (fastest) ultralytics/yolov5#9033)
VOC names dictionary fix (VOC names dictionary fix ultralytics/yolov5#9034)
Update train.py import val as validate (Update train.py import val as validate ultralytics/yolov5#9037)
Simplified notebook
Created using Colaboratory
AutoBatch protect from negative batch sizes (AutoBatch protect from negative batch sizes ultralytics/yolov5#9048)
Temporarily remove macos-latest from CI (Temporarily remove macos-latest from CI ultralytics/yolov5#9049)
Add --save-hybrid mAP warning (Add --save-hybrid mAP warning ultralytics/yolov5#9050)
Refactor for simplification (Refactor for simplification ultralytics/yolov5#9054)
Refactor for simplification 2 (Refactor for simplification 2 ultralytics/yolov5#9055)
zero-mAP fix return .detach() to EMA (zero-mAP fix return .detach() to EMA ultralytics/yolov5#9056)
zero-mAP fix 3 (zero-mAP fix 3 ultralytics/yolov5#9058)
Daemon plot_labels() for faster start (Daemon plot_labels() for faster start ultralytics/yolov5#9057)
TensorBoard fix in tutorial.ipynb (TensorBoard fix in tutorial.ipynb ultralytics/yolov5#9064)
Created using Colaboratory
Created using Colaboratory
Created using Colaboratory
zero-mAP fix remove torch.empty() forward pass in .train() mode (zero-mAP fix remove torch.empty() forward pass in .train() mode ultralytics/yolov5#9068)
Rename 'labels' to 'instances' (Rename 'labels' to 'instances' ultralytics/yolov5#9066)
Threaded TensorBoard graph logging (Threaded TensorBoard graph logging ultralytics/yolov5#9070)
Created using Colaboratory
De-thread TensorBoard graph logging (De-thread TensorBoard graph logging ultralytics/yolov5#9071)
Two dimensional size=(h,w) AutoShape support (Two dimensional size=(h,w) AutoShape support ultralytics/yolov5#9072)
Remove unused Timeout import (Remove unused Timeout import ultralytics/yolov5#9073)
Improved Usage example docstrings (Improved Usage example docstrings ultralytics/yolov5#9075)
Install torch latest stable (Install torch latest stable ultralytics/yolov5#9092)
New @try_export decorator (New @try_export decorator ultralytics/yolov5#9096)
Add optional transforms argument to LoadStreams() (Add optional transforms argument to LoadStreams() ultralytics/yolov5#9105)
Streaming Classification support (Streaming Classification support ultralytics/yolov5#9106)
Fix numpy to torch cls streaming bug (Fix numpy to torch cls streaming bug ultralytics/yolov5#9112)
Infer Loggers project name (Infer Loggers project name ultralytics/yolov5#9117)
Add CSV logging to GenericLogger (Add CSV logging to GenericLogger ultralytics/yolov5#9128)
New TryExcept decorator (New TryExcept decorator ultralytics/yolov5#9154)
Fixed Classify offsets (Fixed Classify offsets ultralytics/yolov5#9155)
New YOLOv5 v6.2 splash images (New YOLOv5 v6.2 splash images ultralytics/yolov5#9142)
Created using Colaboratory
Rename onnx_dynamic -> dynamic (Rename onnx_dynamic -> dynamic ultralytics/yolov5#9168)
Inline _make_grid() meshgrid (Inline _make_grid() meshgrid ultralytics/yolov5#9170)
Comment EMA assert (Comment EMA assert ultralytics/yolov5#9173)
Fix confidence threshold for ClearML debug images (Fix confidence threshold for ClearML debug images ultralytics/yolov5#9174)
Update Dockerfile-cpu (Update Dockerfile-cpu ultralytics/yolov5#9184)
Update Dockerfile-cpu to libpython3-dev (Update Dockerfile-cpu to libpython3-dev ultralytics/yolov5#9185)
Update Dockerfile-arm64 to libpython3-dev (Update Dockerfile-arm64 to libpython3-dev ultralytics/yolov5#9187)
Fix AutoAnchor MPS bug (Fix AutoAnchor MPS bug ultralytics/yolov5#9188)
Skip AMP check on MPS (Skip AMP check on MPS ultralytics/yolov5#9189)
ClearML's set_report_period's time is defined in minutes not seconds. (ClearML's set_report_period's time is defined in minutes not seconds. ultralytics/yolov5#9186)
Add check_git_status(..., branch='master') argument (Add check_git_status(..., branch='master') argument ultralytics/yolov5#9199)
check_font() on notebook init (check_font() on notebook init ultralytics/yolov5#9200)
Comment protobuf in requirements.txt (Comment protobuf in requirements.txt ultralytics/yolov5#9207)
check_font() fstring update (check_font() fstring update ultralytics/yolov5#9208)
AutoBatch protect from extreme batch sizes (AutoBatch protect from extreme batch sizes ultralytics/yolov5#9209)
Default AutoBatch 0.8 fraction (Default AutoBatch 0.8 fraction ultralytics/yolov5#9212)
Delete rebase.yml (Delete rebase.yml ultralytics/yolov5#9202)
Duplicate segment verification fix (Duplicate segment verification fix ultralytics/yolov5#9225)
New LetterBox(size) CenterCrop(size), ToTensor() transforms (New LetterBox(size) CenterCrop(size), ToTensor() transforms (#9213) ultralytics/yolov5#9213)
Add ClassificationModel TF export assert (Add ClassificationModel TF export assert ultralytics/yolov5#9226)
Remove usage of pathlib.Path.unlink(missing_ok=...) (Remove usage of pathlib.Path.unlink(missing_ok=...) ultralytics/yolov5#9227)
Add support for *.pfm images (Add support for *.pfm images ultralytics/yolov5#9230)
Python check warning emoji (Python check warning emoji ultralytics/yolov5#9238)
Add url_getsize() function (Add url_getsize() function ultralytics/yolov5#9247)
Update dataloaders.py (Update dataloaders.py ultralytics/yolov5#9250)
Refactor Loggers : Move code outside train.py (Refactor Loggers : Move code outside train.py ultralytics/yolov5#9241)
Update general.py (Update general.py ultralytics/yolov5#9252)
Add LoadImages._cv2_rotate() (Add LoadImages._cv2_rotate() ultralytics/yolov5#9249)
Move cudnn.benchmarks(True) to LoadStreams (Move cudnn.benchmarks(True) to LoadStreams ultralytics/yolov5#9258)
cudnn.benchmark = True on Seed 0 (cudnn.benchmark = True on Seed 0 ultralytics/yolov5#9259)
Update `TryExcept(msg='...')`` (Update TryExcept(msg='...')` ultralytics/yolov5#9261)
Make sure best.pt model file is preserved ClearML (Make sure best.pt model file is preserved ClearML ultralytics/yolov5#9265)
DetectMultiBackend improvements (DetectMultiBackend improvements ultralytics/yolov5#9269)
Update DetectMultiBackend for tuple outputs (Update DetectMultiBackend for tuple outputs ultralytics/yolov5#9274)
Update DetectMultiBackend for tuple outputs 2 (Update DetectMultiBackend for tuple outputs 2 ultralytics/yolov5#9275)
Update benchmarks CI with --hard-fail min metric floor (Update benchmarks CI with --hard-fail min metric floor ultralytics/yolov5#9276)
Add new --vid-stride inference parameter for videos (Add new --vid-stride inference parameter for videos ultralytics/yolov5#9256)
[pre-commit.ci] pre-commit suggestions ([pre-commit.ci] pre-commit suggestions ultralytics/yolov5#9295)
Replace deprecated np.int with int (Replace deprecated np.int with int ultralytics/yolov5#9307)
Comet Logging and Visualization Integration (Comet Logging and Visualization Integration ultralytics/yolov5#9232)
Comet changes (Comet changes ultralytics/yolov5#9328)
Train.py line 486 typo fix (Train.py line 486 typo fix ultralytics/yolov5#9330)
Add dilated conv support (Add dilated conv support ultralytics/yolov5#9347)
Update check_requirements() single install (Update check_requirements() single install ultralytics/yolov5#9353)
Update check_requirements(args, cmds='') (Update check_requirements(args, cmds='') ultralytics/yolov5#9355)
Update check_requirements() multiple string (Update check_requirements() multiple string ultralytics/yolov5#9356)
Add PaddlePaddle export and inference (Add PaddlePaddle export and inference ultralytics/yolov5#9240)
PaddlePaddle Usage examples (PaddlePaddle Usage examples ultralytics/yolov5#9358)
labels.jpg names fix (labels.jpg names fix ultralytics/yolov5#9361)
Exclude ipython from hubconf.py check_requirements() (Exclude ipython from hubconf.py check_requirements() ultralytics/yolov5#9362)
torch.jit.trace() fix (torch.jit.trace() fix ultralytics/yolov5#9363)
AMP Check fix (AMP Check fix ultralytics/yolov5#9367)
Remove duplicate line in setup.cfg (Remove duplicate line in setup.cfg ultralytics/yolov5#9380)
Remove .train() mode exports (Remove .train() mode exports ultralytics/yolov5#9429)
Continue on Docker arm64 failure (Continue on Docker arm64 failure ultralytics/yolov5#9430)
Continue on Docker failure (all backends) (Continue on Docker failure (all backends) ultralytics/yolov5#9432)
Continue on Docker fail (all backends) fix (Continue on Docker fail (all backends) fix ultralytics/yolov5#9433)
YOLOv5 segmentation model support (YOLOv5 segmentation model support ultralytics/yolov5#9052)
Fix val.py zero-TP bug (Fix val.py zero-TP bug ultralytics/yolov5#9431)
New model.yaml activation: field (New model.yaml activation: field ultralytics/yolov5#9371)
Fix tick labels for background FN/FP (Fix tick labels for background FN/FP ultralytics/yolov5#9414)
Fix TensorRT exports to ONNX opset 12 (Fix TensorRT exports to ONNX opset 12 ultralytics/yolov5#9441)
AutoShape explicit arguments fix (AutoShape explicit arguments fix ultralytics/yolov5#9443)
Update Detections() instance printing (Update Detections() instance printing ultralytics/yolov5#9445)
AutoUpdate TensorFlow in export.py (AutoUpdate TensorFlow in export.py ultralytics/yolov5#9447)
AutoBatch cudnn.benchmark=True fix (AutoBatch cudnn.benchmark=True fix ultralytics/yolov5#9448)
Do not move downloaded zips (Do not move downloaded zips ultralytics/yolov5#9455)
Update general.py (Update general.py ultralytics/yolov5#9454)
Detect() and Segment() fixes for CoreML and Paddle (Detect() and Segment() fixes for CoreML and Paddle ultralytics/yolov5#9458)
Add Paddle exports to benchmarks (Add Paddle exports to benchmarks ultralytics/yolov5#9459)
Add macos-latest runner for CoreML benchmarks (Add macos-latest runner for CoreML benchmarks ultralytics/yolov5#9453)
Fix cutout bug (Fix cutout bug ultralytics/yolov5#9452)
Optimize imports (Optimize imports ultralytics/yolov5#9464)
TensorRT SegmentationModel fix (TensorRT SegmentationModel fix ultralytics/yolov5#9465)
Conv() dilation argument fix (Conv() dilation argument fix ultralytics/yolov5#9466)
Update ClassificationModel default training imgsz=224 (Update ClassificationModel default training imgsz=224 ultralytics/yolov5#9469)
Standardize warnings with WARNING ⚠️ ... (Standardize warnings with WARNING ⚠️ ... ultralytics/yolov5#9467)
TensorFlow macOS AutoUpdate (TensorFlow macOS AutoUpdate ultralytics/yolov5#9471)
classify/predict --save-txt fix (classify/predict --save-txt fix ultralytics/yolov5#9478)
TensorFlow SegmentationModel support (TensorFlow SegmentationModel support ultralytics/yolov5#9472)
AutoBatch report include reserved+allocated (AutoBatch report include reserved+allocated ultralytics/yolov5#9491)
Update Detect() grid init for loop (Update Detect() grid init for loop ultralytics/yolov5#9494)
Accelerate video inference (Accelerate video inference ultralytics/yolov5#9487)
Comet Image Logging Fix (Comet Image Logging Fix ultralytics/yolov5#9498)
Fix visualization title bug (Fix visualization title bug ultralytics/yolov5#9500)
Add paddle tips (Add paddle tips ultralytics/yolov5#9502)
Segmentation polygons2masks_overlap() in np.int32 (Segmentation polygons2masks_overlap() in np.int32 ultralytics/yolov5#9493)
Fix random_perspective param bug in segment (Fix random_perspective param bug in segment ultralytics/yolov5#9512)
Remove check_requirements('flatbuffers==1.12') (Remove check_requirements('flatbuffers==1.12') ultralytics/yolov5#9514)
Fix TF Lite exports (Fix TF Lite exports ultralytics/yolov5#9517)
TFLite fix 2 (TFLite fix 2 ultralytics/yolov5#9518)
FROM nvcr.io/nvidia/pytorch:22.08-py3 (FROM nvcr.io/nvidia/pytorch:22.08-py3 ultralytics/yolov5#9520)
Remove scikit-learn constraint on coremltools 6.0 (Remove scikit-learn constraint on coremltools 6.0 ultralytics/yolov5#9530)
Update scikit-learn constraint per coremltools 6.0 (Update scikit-learn constraint per coremltools 6.0 ultralytics/yolov5#9531)
Update coremltools>=6.0 (Update coremltools>=6.0 ultralytics/yolov5#9532)
Update albumentations (Update albumentations ultralytics/yolov5#9503)
import re (import re ultralytics/yolov5#9535)
TF.js fix (TF.js fix ultralytics/yolov5#9536)
Refactor dataset batch-size (Refactor dataset batch-size ultralytics/yolov5#9551)
Add --source screen for screenshot inference (Add --source screen for screenshot inference ultralytics/yolov5#9542)
Update is_url() (Update is_url() ultralytics/yolov5#9566)
Detect.py supports running against a Triton container (Detect.py supports running against a Triton container ultralytics/yolov5#9228)
New scale_segments() function (New scale_segments() function ultralytics/yolov5#9570)
generator seed fix for DDP mAP drop (generator seed fix for DDP mAP drop ultralytics/yolov5#9545)
Update default GitHub assets (Update default GitHub assets ultralytics/yolov5#9573)
Update requirements.txt comment https://pytorch.org/get-started/locally/ (Update requirements.txt comment https://pytorch.org/get-started/locally/ ultralytics/yolov5#9576)
Add segment line predictions (Add segment line predictions ultralytics/yolov5#9571)
TensorRT detect.py inference fix (TensorRT detect.py inference fix ultralytics/yolov5#9581)
Update Comet links (Update Comet links ultralytics/yolov5#9587)
Add global YOLOv5_DATASETS_DIR (Add global YOLOv5_DATASETS_DIR ultralytics/yolov5#9586)
Add Paperspace Gradient badges (Add Paperspace Gradient badges ultralytics/yolov5#9588)
#YOLOVISION22 announcement (#YOLOVISION22 announcement ultralytics/yolov5#9590)
Bump actions/stale from 5 to 6 (Bump actions/stale from 5 to 6 ultralytics/yolov5#9595)
#YOLOVISION22 update (#YOLOVISION22 update ultralytics/yolov5#9598)
Apple MPS -> CPU NMS fallback strategy (Apple MPS -> CPU NMS fallback strategy ultralytics/yolov5#9600)
Updated Segmentation and Classification usage (Updated Segmentation and Classification usage ultralytics/yolov5#9607)
Update export.py Usage examples (Update export.py Usage examples ultralytics/yolov5#9609)
Fix is_url('https://ultralytics.com') (Fix is_url('https://ultralytics.com') ultralytics/yolov5#9610)
Add results.save(save_dir='path', exist_ok=False) (Add results.save(save_dir='path', exist_ok=False) ultralytics/yolov5#9617)
NMS MPS device wrapper (NMS MPS device wrapper ultralytics/yolov5#9620)
Add SegmentationModel unsupported warning (Add SegmentationModel unsupported warning ultralytics/yolov5#9632)
Disabled upload_dataset flag temporarily due to an artifact related bug (Disabled upload_dataset flag temporarily due to an artifact related bug ultralytics/yolov5#9652)
Add NVIDIA Jetson Nano Deployment tutorial (Add NVIDIA Jetson Nano Deployment tutorial ultralytics/yolov5#9656)
Added cutout import from utils/augmentations.py to use Cutout Aug in … (Added cutout import from utils/augmentations.py to use Cutout Aug in … ultralytics/yolov5#9668)
Simplify val.py benchmark mode with speed mode (Simplify val.py benchmark mode with speed mode ultralytics/yolov5#9674)
Allow list for Comet artifact class 'names' field (Allow list for Comet artifact class 'names' field ultralytics/yolov5#9654)
[pre-commit.ci] pre-commit suggestions ([pre-commit.ci] pre-commit suggestions ultralytics/yolov5#9685)
TensorRT --dynamic fix (TensorRT --dynamic fix ultralytics/yolov5#9691)
FROM nvcr.io/nvidia/pytorch:22.09-py3 (FROM nvcr.io/nvidia/pytorch:22.09-py3 ultralytics/yolov5#9711)
Error in utils/segment/general masks2segments() (Error in utils/segment/general masks2segments() ultralytics/yolov5#9724)
Fix segment evolution keys (Fix segment evolution keys ultralytics/yolov5#9742)
Remove #YOLOVISION22 notice (Remove #YOLOVISION22 notice ultralytics/yolov5#9751)
Update Loggers (Update Loggers ultralytics/yolov5#9760)
update mask2segments and saving results (update mask2segments and saving results ultralytics/yolov5#9785)
HUB VOC fix (HUB VOC fix ultralytics/yolov5#9792)
Update hubconf.py local repo Usage example (Update hubconf.py local repo Usage example ultralytics/yolov5#9803)
Fix xView dataloaders import (Fix xView dataloaders import ultralytics/yolov5#9807)
Argoverse HUB fix (Argoverse HUB fix ultralytics/yolov5#9809)
smart_optimizer() revert to weight with decay (smart_optimizer() revert to weight with decay ultralytics/yolov5#9817)
Allow PyTorch Hub results to display in notebooks (Allow PyTorch Hub results to display in notebooks ultralytics/yolov5#9825)

Update

updates: - [github.com/asottile/pyupgrade: v2.34.0 → v2.37.3](asottile/pyupgrade@v2.34.0...v2.37.3) - [github.com/PyCQA/flake8: 4.0.1 → 5.0.2](PyCQA/flake8@4.0.1...5.0.2) Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

…tics#8830) * Update hubconf.py Loading a custom yolov model causes this line to fail. Adding a test to check if the model actually has a model.model field. With this check, I'm able to load the model no prob. Loading model via ```py model = torch.hub.load( 'ultralytics/yolov5', 'custom', 'models/frozen_backbone_coco_unlabeled_best.onnx', autoshape=True, force_reload=False ) ``` Causes traceback: ``` Traceback (most recent call last): File "/Users/jackson/Documents/GitHub/w210-capstone/.venv/lib/python3.10/site-packages/flask/app.py", line 2077, in wsgi_app response = self.full_dispatch_request() File "/Users/jackson/Documents/GitHub/w210-capstone/.venv/lib/python3.10/site-packages/flask/app.py", line 1525, in full_dispatch_request rv = self.handle_user_exception(e) File "/Users/jackson/Documents/GitHub/w210-capstone/.venv/lib/python3.10/site-packages/flask/app.py", line 1523, in full_dispatch_request rv = self.dispatch_request() File "/Users/jackson/Documents/GitHub/w210-capstone/.venv/lib/python3.10/site-packages/flask/app.py", line 1509, in dispatch_request return self.ensure_sync(self.view_functions[rule.endpoint])(**req.view_args) File "/Users/jackson/Documents/GitHub/w210-capstone/api/endpoints/predictions.py", line 26, in post_predictions yolov_predictions = predict_bounding_boxes_for_collection(collection_id) File "/Users/jackson/Documents/GitHub/w210-capstone/api/predictions/predict_bounding_boxes.py", line 43, in predict_bounding_boxes_for_collection model = torch.hub.load( File "/Users/jackson/Documents/GitHub/w210-capstone/.venv/lib/python3.10/site-packages/torch/hub.py", line 404, in load model = _load_local(repo_or_dir, model, *args, **kwargs) File "/Users/jackson/Documents/GitHub/w210-capstone/.venv/lib/python3.10/site-packages/torch/hub.py", line 433, in _load_local model = entry(*args, **kwargs) File "/Users/jackson/.cache/torch/hub/ultralytics_yolov5_master/hubconf.py", line 72, in custom return _create(path, autoshape=autoshape, verbose=_verbose, device=device) File "/Users/jackson/.cache/torch/hub/ultralytics_yolov5_master/hubconf.py", line 67, in _create raise Exception(s) from e Exception: 'DetectMultiBackend' object has no attribute 'model'. Cache may be out of date, try `force_reload=True` or see ultralytics#36 for help. Exception on /api/v1/predictions [POST] Traceback (most recent call last): File "/Users/jackson/.cache/torch/hub/ultralytics_yolov5_master/hubconf.py", line 58, in _create model.model.model[-1].inplace = False # Detect.inplace=False for safe multithread inference File "/Users/jackson/Documents/GitHub/w210-capstone/.venv/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1185, in __getattr__ raise AttributeError("'{}' object has no attribute '{}'".format( AttributeError: 'DetectMultiBackend' object has no attribute 'model' ``` * Update hubconf.py * Update common.py Co-authored-by: Glenn Jocher <[email protected]>

* New `smart_resume()` * Update torch_utils.py * Update torch_utils.py * Update torch_utils.py * fix

* Single checkpoint resume * Update train.py * Add hyp * Add hyp * Add hyp * FIX * avoid resume on url data * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * avoid resume on url data * avoid resume on url data * Update Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

* Add check_file(data) i.e. `--data coco128.yaml` * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

* export.py replace `check_file` -> `check_yaml` * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

May help ultralytics#8862

* Fix TensorRT --dynamic excess outputs bug Potential fix for ultralytics#8790 * Cleanup * Update common.py * Update common.py * New fix

@kalenmike

@kalenmike should represent a 30% filesize reduction vs 75 quality

@kalenmike

@kalenmike

@kalenmike

--resume training from URL weights fix @kalenmike should fix data error on HUB resume

* Update CI to default Python 3.10 * Update ci-testing.yml * Update ci-testing.yml

* Add titles to matplotlib plots * Add ClearML Experiment Tracking integration. * Add ClearML Data Version Management automatic download when requested * Add ClearML Hyperparameter Optimization * ClearML save period integration * Fix wandb breaking when used with ClearML dataset * Fix wandb breaking when used with ClearML resume and dataset * Add ClearML documentation * fixed small bug in clearml integration that misreports epoch number * Final ClearMl additions before refactor * Add correct epoch reporting * Add remote execution and autoscaling docs for ClearML integration * Added images to clearml integration docs * fixed logo alignment bug and added hpo screenshot clearml * Fixed small epoch number bug in clearml integration * Remove saved model flush clearml * Cleanup clearml readme section * Cleaned up clearml logger docstring * Remove resume readme section clearml * Clearml integration cleanup * Updated ClearML documentation * Added dark vs light icons ClearML Readme * Clearml Readme styling * Add better gifs * Fixed gif file size * Add better images in tutorial notebook * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Addressed comments in PR ultralytics#8620 * Fixed circular import * Fixed circular import * Update tutorial.ipynb * Update tutorial.ipynb * Inline comment * Restructured tutorial notebook * Add correct ClearML link to README * Update tutorial.ipynb * Update general.py * Update __init__.py * Update __init__.py * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update __init__.py * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update __init__.py * Update README.md * Update __init__.py * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * spelling * Update tutorial.ipynb * notebook cutt.ly links * Update README.md * Update README.md * cutt.ly links in tutorial * Removed labels as they show up on last subplot only Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Glenn Jocher <[email protected]>

Update smart_optimizer() weight_decay reporting

* emoji-safe default logging * implement everywhere

* fix: added transparent image and empty alt to social bar * fix: remove underlines * fix: hide underlines for all social links * Update image URLs to ultralytics/assets * Update bottom 3% icons * Update README.md Co-authored-by: Glenn Jocher <[email protected]>

* Fix anchor incorrectly initialized Using --noautoanchor and --evolve simultaneously leads to anchor incorrectly initialized. --noautoanchor denotes anchors don't need to evolve, thus removing anchors from hyp will fix it. * Update train.py * Update train.py Co-authored-by: Glenn Jocher <[email protected]>

Resolves ultralytics#8842

The notes says "Return True if directory has write permissions", however, the code below is "os.R_OK", I think "os.W_OK" is preferred.

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Until more ops are fully supported this update will allow for seamless MPS inference (but slower MPS to CPU transfer before NMS, so slower NMS times). Partially resolves ultralytics#9596 Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

* Updated Segmentation and Classification usage Signed-off-by: Glenn Jocher <[email protected]> * Update export.py Signed-off-by: Glenn Jocher <[email protected]> * Update export.py Signed-off-by: Glenn Jocher <[email protected]> * Update export.py Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

* Update export.py Usage examples Signed-off-by: Glenn Jocher <[email protected]> * Update export.py Signed-off-by: Glenn Jocher <[email protected]> * Update export.py Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Failing on missing path, i.e. no 'www.' Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

* NMS MPS device wrapper May resolve ultralytics#9613 Signed-off-by: Glenn Jocher <[email protected]> * Update general.py Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

* Add SegmentationModel unsupported warning Signed-off-by: Glenn Jocher <[email protected]> * Update hubconf.py Signed-off-by: Glenn Jocher <[email protected]> * Update hubconf.py Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

…ug (ultralytics#9652) * disabled upload_dataset flag temporarily due to an artifact related bug * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

ultralytics#9668) * Added cutout import from utils/augmentations.py to use Cutout Aug in data loader by un-commenting line 679, 680, 681 * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Update

* Update __init__.py In the Comet logger, when I run train.py, it wants to download the data artifact. It was requiring me to format the 'names' field in the data artifact metadata as a dictionary, so I've changed this so that it also accepts a list. Signed-off-by: KristenKehrer <[email protected]> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update utils/loggers/comet/__init__.py Co-authored-by: Dhruv Nair <[email protected]> Signed-off-by: KristenKehrer <[email protected]> Signed-off-by: KristenKehrer <[email protected]> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Dhruv Nair <[email protected]> Co-authored-by: Glenn Jocher <[email protected]>

updates: - [github.com/asottile/pyupgrade: v2.37.3 → v2.38.2](asottile/pyupgrade@v2.37.3...v2.38.2) Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

* Update export.py Signed-off-by: Glenn Jocher <[email protected]> * Update export.py Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

When running segmentation predict on gpu, the conversion from tensor to numpy fails. Calling `.cpu()` solves this problem. Signed-off-by: Paul Guerrie <[email protected]> Signed-off-by: Paul Guerrie <[email protected]>

* Update * Cleanup

Update README.md Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

* Update * Update tutorial.ipynb Signed-off-by: Glenn Jocher <[email protected]> * Update tutorial.ipynb Signed-off-by: Glenn Jocher <[email protected]> * Update tutorial.ipynb Signed-off-by: Glenn Jocher <[email protected]> * Update requirements.txt Signed-off-by: Glenn Jocher <[email protected]> * Update * Update README.md Signed-off-by: Glenn Jocher <[email protected]> * Update Signed-off-by: Glenn Jocher <[email protected]>

* update mask2segments and saving results * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Update general.py Signed-off-by: Glenn Jocher <[email protected]> * Update predict.py Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Glenn Jocher <[email protected]>

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

* Update hubconf.py Signed-off-by: Glenn Jocher <[email protected]> * Update hubconf.py Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

If a parameter does not fall into any other category Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

* Allow PyTorch Hub results to display in notebooks * fix CI * fix CI * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * fix CI * fix CI * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * fix CI * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * fix CI * fix CI * fix CI * fix CI * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * fix CI * fix CI * fix CI Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

glenn-jocher and others added 30 commits August 1, 2022 21:45

RUN git clone instead of COPY to /usr/src/app (ultralytics#8827)

ba140e5

Update

New smart_resume() (ultralytics#8838)

08c8c3e

* New `smart_resume()` * Update torch_utils.py * Update torch_utils.py * Update torch_utils.py * fix

Created using Colaboratory

e5991c9

Created using Colaboratory

4d8d84b

Update dataloaders.py remove float64 shapes (ultralytics#8865)

84e7748

May help ultralytics#8862

Fix TensorRT --dynamic excess outputs bug (ultralytics#8869)

38a6eb6

* Fix TensorRT --dynamic excess outputs bug Potential fix for ultralytics#8790 * Cleanup * Update common.py * Update common.py * New fix

Single-line TRT dynamic assertion (ultralytics#8871)

731a2f8

HUBDatasetStats() preview images to 50 quality (ultralytics#8880)

bc9fcb1

@kalenmike should represent a 30% filesize reduction vs 75 quality

--resume training from URL weights (ultralytics#8882)

e073658

@kalenmike

--resume training from URL weights fix (ultralytics#8884)

daed7a8

--resume training from URL weights fix @kalenmike should fix data error on HUB resume

Update CI to default Python 3.10 (ultralytics#8883)

2794483

* Update CI to default Python 3.10 * Update ci-testing.yml * Update ci-testing.yml

Created using Colaboratory

5a0faef

Created using Colaboratory

030586a

smart_optimizer() improved reporting (ultralytics#8887)

20f1b7e

Update smart_optimizer() weight_decay reporting

Created using Colaboratory

d631a69

Created using Colaboratory

73ed7a5

emoji-safe default logging (ultralytics#8888)

8f28ea4

* emoji-safe default logging * implement everywhere

Update README_cn.md

fcfaf0a

Update README.md Environments icons (ultralytics#8895)

48a8531

Edge TPU add --search_delegate fix (ultralytics#8902)

4bb3052

Resolves ultralytics#8842

Capitalize YouTube (ultralytics#8903)

f5335f2

About "os.R_OK" in general.py (ultralytics#8909)

b551098

The notes says "Return True if directory has write permissions", however, the code below is "os.R_OK", I think "os.W_OK" is preferred.

glenn-jocher and others added 28 commits September 26, 2022 13:27

#YOLOVISION22 update (ultralytics#9598)

bd9c0c4

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Fix is_url('https://ultralytics.com')` (ultralytics#9610)

1460e57

Failing on missing path, i.e. no 'www.' Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Add results.save(save_dir='path', exist_ok=False) (ultralytics#9617)

7314363

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

NMS MPS device wrapper (ultralytics#9620)

2373d54

* NMS MPS device wrapper May resolve ultralytics#9613 Signed-off-by: Glenn Jocher <[email protected]> * Update general.py Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Add NVIDIA Jetson Nano Deployment tutorial (ultralytics#9656)

82bec4c

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Simplify val.py benchmark mode with speed mode (ultralytics#9674)

1158a50

Update

[pre-commit.ci] pre-commit suggestions (ultralytics#9685)

68d654d

updates: - [github.com/asottile/pyupgrade: v2.37.3 → v2.38.2](asottile/pyupgrade@v2.37.3...v2.38.2) Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

TensorRT --dynamic fix (ultralytics#9691)

e4398cf

* Update export.py Signed-off-by: Glenn Jocher <[email protected]> * Update export.py Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

FROM nvcr.io/nvidia/pytorch:22.09-py3 (ultralytics#9711)

7f097dd

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Error in utils/segment/general masks2segments() (ultralytics#9724)

5ef69ef

When running segmentation predict on gpu, the conversion from tensor to numpy fails. Calling `.cpu()` solves this problem. Signed-off-by: Paul Guerrie <[email protected]> Signed-off-by: Paul Guerrie <[email protected]>

Fix segment evolution keys (ultralytics#9742)

209be93

* Update * Cleanup

Remove #YOLOVISION22 notice (ultralytics#9751)

2f1eb21

Update README.md Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

HUB VOC fix (ultralytics#9792)

16f87bb

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Update hubconf.py local repo Usage example (ultralytics#9803)

15b7565

* Update hubconf.py Signed-off-by: Glenn Jocher <[email protected]> * Update hubconf.py Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Fix xView dataloaders import (ultralytics#9807)

2a19d07

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

Argoverse HUB fix (ultralytics#9809)

df80e7c

Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

smart_optimizer() revert to weight with decay (ultralytics#9817)

e42c89d

If a parameter does not fall into any other category Signed-off-by: Glenn Jocher <[email protected]> Signed-off-by: Glenn Jocher <[email protected]>

feat: bump our version to latest

5721dee

Hojland merged commit e148d54 into master Oct 17, 2022

Hojland deleted the feat/bump branch October 17, 2022 13:01

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat/bump #15

feat/bump #15

Hojland commented Oct 17, 2022

feat/bump #15

feat/bump #15

Conversation

Hojland commented Oct 17, 2022