Dockerfile improvements (#23090)

Signed-off-by: Glenn Jocher <glenn.jocher@ultralytics.com>

docker/Dockerfile

@@ -40,7 +40,7 @@ RUN sed -i '/^\[http "https:\/\/github\.com\/"\]/,+1d' .git/config && \
     sed -i'' -e 's/"opencv-python/"opencv-python-headless/' pyproject.toml
 ADD https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt .
 
-# Install pip packages
+# Install pip packages (uv already installed in base image)
 RUN uv pip install --system -e "." albumentations faster-coco-eval wandb && \
     # Remove extra build files \
     rm -rf tmp /root/.config/Ultralytics/persistent_cache.json

docker/Dockerfile-arm64

@@ -20,13 +20,11 @@ ADD https://github.com/ultralytics/assets/releases/download/v0.0.0/Arial.ttf \
     /root/.config/Ultralytics/
 
 # Install linux packages
-# pkg-config and libhdf5-dev (not included) are needed to build 'h5py==3.11.0' aarch64 wheel required by 'tensorflow'
-# gnupg required for Edge TPU install
-RUN apt update && \
-    apt upgrade -y && \
-    apt install -y --no-install-recommends \
+# TensorFlow on aarch64 may require pkg-config and libhdf5-dev if h5py builds from source.
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends \
     python3-pip git zip unzip wget curl htop gcc libgl1 libglib2.0-0 gnupg && \
-    apt clean && \
+    apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
 # Create working directory
@@ -39,7 +37,7 @@ RUN sed -i '/^\[http "https:\/\/github\.com\/"\]/,+1d' .git/config && \
 ADD https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt .
 
 # Install pip packages, create python symlink, and remove build files
-RUN pip install uv && \
+RUN python3 -m pip install uv && \
     uv pip install --system -e ".[export]" --break-system-packages && \
     # Creates a symbolic link to make 'python' point to 'python3'
     ln -sf /usr/bin/python3 /usr/bin/python && \

docker/Dockerfile-conda

@@ -20,11 +20,13 @@ ADD https://github.com/ultralytics/assets/releases/download/v0.0.0/Arial.ttf \
 # Install linux packages and conda packages
 RUN apt-get update && \
     apt-get install -y --no-install-recommends libgl1 && \
+    apt-get clean && \
     # Install conda packages
     # mkl required to fix 'OSError: libmkl_intel_lp64.so.2: cannot open shared object file: No such file or directory'
     conda config --set solver libmamba && \
-    conda install pytorch torchvision pytorch-cuda=12.1 -c pytorch -c nvidia && \
-    conda install -c conda-forge ultralytics mkl && \
+    conda install -y pytorch torchvision pytorch-cuda=12.1 -c pytorch -c nvidia && \
+    conda install -y -c conda-forge ultralytics mkl && \
+    conda clean -afy && \
     # Remove extra build files
     rm -rf /var/lib/apt/lists/* /root/.config/Ultralytics/persistent_cache.json
 

docker/Dockerfile-export

@@ -6,7 +6,7 @@
 FROM ultralytics/ultralytics:latest
 
 # Install export dependencies and run exports to AutoInstall packages
-# Numpy 1.26.4 required due to TF export bug with torch 2.8
+# Numpy 1.26.4 required for TensorFlow export compatibility
 # Note tensorrt installed on-demand as depends on runtime environment CUDA version
 RUN uv pip install --system -e ".[export]" "onnxruntime-gpu" paddlepaddle x2paddle numpy==1.26.4 && \
     # Run exports to AutoInstall packages \

docker/Dockerfile-jetson-jetpack4

@@ -25,8 +25,9 @@ RUN wget -q -O - https://repo.download.nvidia.com/jetson/jetson-ota-public.asc |
 # gnupg required for Edge TPU install
 RUN apt-get update && \
     apt-get install -y --no-install-recommends \
-    git python3.8 python3.8-dev python3-pip python3-libnvinfer libopenmpi-dev libopenblas-base libomp-dev gcc \
-    && rm -rf /var/lib/apt/lists/*
+    git python3.8 python3.8-dev python3-pip python3-libnvinfer libopenmpi-dev libopenblas-base libomp-dev gcc && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
 
 # Create symbolic links for python3.8 and pip3
 RUN ln -sf /usr/bin/python3.8 /usr/bin/python3 && \

docker/Dockerfile-jetson-jetpack5

@@ -21,6 +21,7 @@ ADD https://github.com/ultralytics/assets/releases/download/v0.0.0/Arial.ttf \
 RUN apt-get update && \
     apt-get install -y --no-install-recommends \
     git python3-pip libopenmpi-dev libopenblas-base libomp-dev \
+    && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 
 # Create working directory

docker/Dockerfile-jetson-jetpack6

@@ -23,6 +23,7 @@ RUN dpkg -i cuda-keyring_1.1-1_all.deb && \
     apt-get update && \
     apt-get install -y --no-install-recommends \
     git python3-pip libopenmpi-dev libopenblas-base libomp-dev libcusparselt0 libcusparselt-dev && \
+    apt-get clean && \
     rm -rf /var/lib/apt/lists/* cuda-keyring_1.1-1_all.deb
 
 # Create working directory

docker/Dockerfile-jupyter

@@ -15,7 +15,7 @@ RUN uv pip install --system jupyterlab && \
     rm -rf tmp /root/.config/Ultralytics/persistent_cache.json
 
 # Start JupyterLab with tutorial notebook
-ENTRYPOINT ["/usr/local/bin/jupyter", "lab", "--allow-root", "--ip=*", "/ultralytics/examples/tutorial.ipynb"]
+ENTRYPOINT ["/usr/local/bin/jupyter", "lab", "--allow-root", "--ip=0.0.0.0", "/ultralytics/examples/tutorial.ipynb"]
 
 # Usage Examples -------------------------------------------------------------------------------------------------------
 

docker/Dockerfile-runner

@@ -14,16 +14,17 @@ ENV RUNNER_ALLOW_RUNASROOT=1 \
 # Set the working directory
 WORKDIR /actions-runner
 
-# Download and unpack the latest runner from https://github.com/actions/runner and install dependencies
+# Download and unpack the runner from https://github.com/actions/runner and install dependencies
 RUN FILENAME=actions-runner-linux-x64-${RUNNER_VERSION}.tar.gz && \
-    curl -o "$FILENAME" -L "https://github.com/actions/runner/releases/download/v${RUNNER_VERSION}/${FILENAME}" && \
+    curl -fLso "$FILENAME" "https://github.com/actions/runner/releases/download/v${RUNNER_VERSION}/${FILENAME}" && \
     tar xzf "$FILENAME" && \
     rm "$FILENAME" && \
     # Install runner dependencies \
     uv pip install --system pytest-cov && \
     ./bin/installdependencies.sh && \
     apt-get update && \
-    apt-get -y install libicu-dev && \
+    apt-get install -y --no-install-recommends libicu-dev && \
+    apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
 # JSON ENTRYPOINT command to configure and start runner with default TOKEN and NAME

docs/en/help/FAQ.md

@@ -24,13 +24,13 @@ Ultralytics is a [computer vision](https://www.ultralytics.com/glossary/computer
 
 Installing the Ultralytics package is straightforward using pip:
 
-```
+```bash
 pip install ultralytics
 ```
 
 For the latest development version, install directly from the GitHub repository:
 
-```
+```bash
 pip install git+https://github.com/ultralytics/ultralytics.git
 ```
 
@@ -168,7 +168,7 @@ Ultralytics YOLO boasts a rich set of features for advanced computer vision task
 - Pretrained Models: Access a variety of [pretrained models](https://docs.ultralytics.com/models/) that balance speed and accuracy for different use cases.
 - Custom Training: Easily fine-tune models on custom datasets with the flexible [training pipeline](https://docs.ultralytics.com/modes/train/).
 - Wide [Deployment Options](https://docs.ultralytics.com/guides/model-deployment-options/): Export models to various formats like TensorRT, ONNX, and CoreML for deployment across different platforms.
-- Extensive Documentation: Benefit from comprehensive [documentation](https://docs.ultralytics.com/) and a supportive community to guide you through your computer vision journey.
+- Extensive Documentation: Benefit from comprehensive [documentation](https://docs.ultralytics.com/) and a supportive community for your computer vision workflows.
 
 ### How can I improve the performance of my YOLO model?
 

docs/en/integrations/jupyterlab.md

@@ -95,7 +95,7 @@ For a detailed understanding of the model training process and best practices, r
 
 ## Keep Learning about JupyterLab
 
-If you're excited to learn more about JupyterLab, here are some great resources to get you started:
+If you want to learn more about JupyterLab, here are resources to get you started:
 
 - [**JupyterLab Documentation**](https://jupyterlab.readthedocs.io/en/stable/getting_started/starting.html): Dive into the official JupyterLab Documentation to explore its features and capabilities. It's a great way to understand how to use this powerful tool to its fullest potential.
 - [**Try It With Binder**](https://mybinder.org/v2/gh/jupyterlab/jupyterlab-demo/HEAD?urlpath=lab/tree/demo): Experiment with JupyterLab without installing anything by using Binder, which lets you launch a live JupyterLab instance directly in your browser. It's a great way to start experimenting immediately.

docs/en/yolov5/environments/docker_image_quickstart_tutorial.md

@@ -202,4 +202,4 @@ Learn more about evaluation metrics like [Precision](https://www.ultralytics.com
 
 <p align="center"><img width="1000" src="https://github.com/ultralytics/docs/releases/download/0/gcp-running-docker.avif" alt="Running YOLOv5 inside a Docker container on GCP"></p>
 
-Congratulations! You have successfully set up and run YOLOv5 within a Docker container.
+You have successfully set up and run YOLOv5 within a Docker container.

examples/YOLO-Series-ONNXRuntime-Rust/README.md

@@ -1,6 +1,6 @@
 # YOLO-Series ONNXRuntime Rust Demo for Core YOLO Tasks
 
-This repository provides a [Rust](https://rust-lang.org/) demo showcasing key [Ultralytics YOLO](https://docs.ultralytics.com/) series tasks such as [Classification](https://docs.ultralytics.com/tasks/classify/), [Segmentation](https://docs.ultralytics.com/tasks/segment/), [Detection](https://docs.ultralytics.com/tasks/detect/), [Pose Estimation](https://docs.ultralytics.com/tasks/pose/), and Oriented Bounding Box ([OBB](https://docs.ultralytics.com/tasks/obb/)) detection using the [ONNXRuntime](https://github.com/microsoft/onnxruntime). It supports various YOLO models (v5 through 11) across multiple computer vision tasks.
+This repository provides a [Rust](https://rust-lang.org/) demo showcasing key [Ultralytics YOLO](https://docs.ultralytics.com/) series tasks such as [Classification](https://docs.ultralytics.com/tasks/classify/), [Segmentation](https://docs.ultralytics.com/tasks/segment/), [Detection](https://docs.ultralytics.com/tasks/detect/), [Pose Estimation](https://docs.ultralytics.com/tasks/pose/), and Oriented Bounding Box ([OBB](https://docs.ultralytics.com/tasks/obb/)) detection using the [ONNXRuntime](https://github.com/microsoft/onnxruntime). It supports various YOLO models (YOLOv5 through YOLO11) across multiple computer vision tasks.
 
 ## ✨ Introduction
 

examples/YOLO11-Triton-CPP/README.md

@@ -105,6 +105,6 @@ Contributions are welcome! If you find any issues or have suggestions for improv
 
 ---
 
-For more resources, explore the [Ultralytics documentation](https://docs.ultralytics.com/), [Ultralytics blog](https://www.ultralytics.com/blog), and [Ultralytics HUB](https://docs.ultralytics.com/hub/).
+For more resources, explore the [Ultralytics documentation](https://docs.ultralytics.com/) and [Ultralytics blog](https://www.ultralytics.com/blog).
 
 **We encourage your contributions to help improve this project.**

examples/YOLOv8-ONNXRuntime-CPP/inference.cpp

@@ -63,18 +63,15 @@ char* YOLO_V8::PreProcess(cv::Mat& iImg, std::vector<int> iImgSize, cv::Mat& oIm
     case YOLO_DETECT_V8_HALF:
     case YOLO_POSE_V8_HALF://LetterBox
     {
-        if (iImg.cols >= iImg.rows)
-        {
-            resizeScales = iImg.cols / (float)iImgSize.at(0);
-            cv::resize(oImg, oImg, cv::Size(iImgSize.at(0), int(iImg.rows / resizeScales)));
-        }
-        else
-        {
-            resizeScales = iImg.rows / (float)iImgSize.at(0);
-            cv::resize(oImg, oImg, cv::Size(int(iImg.cols / resizeScales), iImgSize.at(1)));
-        }
-        cv::Mat tempImg = cv::Mat::zeros(iImgSize.at(0), iImgSize.at(1), CV_8UC3);
-        oImg.copyTo(tempImg(cv::Rect(0, 0, oImg.cols, oImg.rows)));
+        int new_h = iImgSize.at(0);
+        int new_w = iImgSize.at(1);
+        float r = min(new_w / (float)iImg.cols, new_h / (float)iImg.rows);
+        int resized_w = static_cast<int>(iImg.cols * r);
+        int resized_h = static_cast<int>(iImg.rows * r);
+        resizeScales = 1.0f / r;
+        cv::resize(oImg, oImg, cv::Size(resized_w, resized_h));
+        cv::Mat tempImg = cv::Mat::zeros(new_h, new_w, CV_8UC3);
+        oImg.copyTo(tempImg(cv::Rect(0, 0, resized_w, resized_h)));
         oImg = tempImg;
         break;
     }
@@ -85,7 +82,7 @@ char* YOLO_V8::PreProcess(cv::Mat& iImg, std::vector<int> iImgSize, cv::Mat& oIm
         int m = min(h, w);
         int top = (h - m) / 2;
         int left = (w - m) / 2;
-        cv::resize(oImg(cv::Rect(left, top, m, m)), oImg, cv::Size(iImgSize.at(0), iImgSize.at(1)));
+        cv::resize(oImg(cv::Rect(left, top, m, m)), oImg, cv::Size(iImgSize.at(1), iImgSize.at(0)));
         break;
     }
     }
@@ -335,7 +332,7 @@ char* YOLO_V8::TensorProcess(clock_t& starttime_1, cv::Mat& iImg, N& blob, std::
 
 char* YOLO_V8::WarmUpSession() {
     clock_t starttime_1 = clock();
-    cv::Mat iImg = cv::Mat(cv::Size(imgSize.at(0), imgSize.at(1)), CV_8UC3);
+    cv::Mat iImg = cv::Mat(cv::Size(imgSize.at(1), imgSize.at(0)), CV_8UC3);
     cv::Mat processedImg;
     PreProcess(iImg, imgSize, processedImg);
     if (modelType < 4)