From e3bf88a18bcdf0efbb91f8069a3d8b83a146ea90 Mon Sep 17 00:00:00 2001 From: Glenn Jocher Date: Fri, 29 May 2026 01:07:39 +0800 Subject: [PATCH] Fix 404 links in Comet and TensorRT docs (#24634) Signed-off-by: Glenn Jocher Co-authored-by: UltralyticsAssistant --- docs/en/integrations/comet.md | 4 ++-- docs/en/integrations/deepx.md | 10 +++++----- docs/en/integrations/tensorrt.md | 2 +- docs/mkdocs_github_authors.yaml | 3 +++ 4 files changed, 11 insertions(+), 8 deletions(-) diff --git a/docs/en/integrations/comet.md b/docs/en/integrations/comet.md index 9547eb00..0506b2a3 100644 --- a/docs/en/integrations/comet.md +++ b/docs/en/integrations/comet.md @@ -194,7 +194,7 @@ os.environ["COMET_EVAL_LOG_CONFUSION_MATRIX"] = "false" ### Online and Offline Mode -By default, Comet runs in online mode and streams experiment data to the Comet servers. If you need to train without internet access, set `COMET_START_ONLINE=0` before training starts. Experiment data is saved locally and can be uploaded later with the [`comet upload`](https://www.comet.com/docs/v2/guides/comet-cli/comet-upload-cli/) CLI. +By default, Comet runs in online mode and streams experiment data to the Comet servers. If you need to train without internet access, set `COMET_START_ONLINE=0` before training starts. Experiment data is saved locally and can be uploaded later with the [`comet upload`](https://www.comet.com/docs/v2/api-and-sdk/command-line/reference/) CLI. ```python import os @@ -368,4 +368,4 @@ import os os.environ["COMET_START_ONLINE"] = "0" ``` -Experiment data is saved on disk and can be uploaded to Comet later with the [`comet upload`](https://www.comet.com/docs/v2/guides/comet-cli/comet-upload-cli/) CLI when connectivity is available. The earlier `COMET_MODE="offline"` variable still works but emits a deprecation warning. For more details, see the [Online and Offline Mode](#online-and-offline-mode) section. +Experiment data is saved on disk and can be uploaded to Comet later with the [`comet upload`](https://www.comet.com/docs/v2/api-and-sdk/command-line/reference/) CLI when connectivity is available. The earlier `COMET_MODE="offline"` variable still works but emits a deprecation warning. For more details, see the [Online and Offline Mode](#online-and-offline-mode) section. diff --git a/docs/en/integrations/deepx.md b/docs/en/integrations/deepx.md index 2e128fd2..6b946398 100644 --- a/docs/en/integrations/deepx.md +++ b/docs/en/integrations/deepx.md @@ -193,12 +193,12 @@ Verify the runtime is installed correctly with `dxrt-cli --version`. You should ```sh DXRT v3.3.2 Minimum Driver Versions - Device Driver: v2.4.0 - PCIe Driver: v2.2.0 - Firmware: v2.5.2 +Device Driver: v2.4.0 +PCIe Driver: v2.2.0 +Firmware: v2.5.2 Minimum Compiler Versions - Compiler: v1.18.1 - .dxnn File Format: v6 +Compiler: v1.18.1 +.dxnn File Format: v6 ``` Once the runtime is installed, run inference and validation on your DeepX device exactly as shown in the [Usage](#usage) section above — the exported `_deepx_model` loads directly with `YOLO(...)`. diff --git a/docs/en/integrations/tensorrt.md b/docs/en/integrations/tensorrt.md index 731d06a4..a41484c6 100644 --- a/docs/en/integrations/tensorrt.md +++ b/docs/en/integrations/tensorrt.md @@ -177,7 +177,7 @@ When processing implicitly quantized networks TensorRT uses INT8 opportunistical #### Configuring INT8 Export -The arguments provided when using [export](../modes/export.md) for an Ultralytics YOLO model will **greatly** influence the performance of the exported model. They will also need to be selected based on the device resources available, however the default arguments _should_ work for most [Ampere (or newer) NVIDIA discrete GPUs](https://developer.nvidia.com/blog/nvidia-ampere-architecture-in-depth/). The calibration algorithm used is `"MINMAX_CALIBRATION"` and you can read more details about the options available [in the TensorRT Developer Guide](https://docs.nvidia.com/deeplearning/tensorrt/latest/_static/python-api/infer/Int8/MinMaxCalibrator.html). Ultralytics tests found that `"MINMAX_CALIBRATION"` was the best choice and exports are fixed to using this algorithm. +The arguments provided when using [export](../modes/export.md) for an Ultralytics YOLO model will **greatly** influence the performance of the exported model. They will also need to be selected based on the device resources available, however the default arguments _should_ work for most [Ampere (or newer) NVIDIA discrete GPUs](https://developer.nvidia.com/blog/nvidia-ampere-architecture-in-depth/). The calibration algorithm used is `"MINMAX_CALIBRATION"` and you can read more details about the options available [in the TensorRT Developer Guide](https://docs.nvidia.com/deeplearning/tensorrt/latest/inference-library/work-quantized-types.html). Ultralytics tests found that `"MINMAX_CALIBRATION"` was the best choice and exports are fixed to using this algorithm. - `workspace` : Controls the size (in GiB) of the device memory allocation while converting the model weights. - Adjust the `workspace` value according to your calibration needs and resource availability. While a larger `workspace` may increase calibration time, it allows TensorRT to explore a wider range of optimization tactics, potentially enhancing model performance and [accuracy](https://www.ultralytics.com/glossary/accuracy). Conversely, a smaller `workspace` can reduce calibration time but may limit the optimization strategies, affecting the quality of the quantized model. diff --git a/docs/mkdocs_github_authors.yaml b/docs/mkdocs_github_authors.yaml index 59ee2c87..95b71fe3 100644 --- a/docs/mkdocs_github_authors.yaml +++ b/docs/mkdocs_github_authors.yaml @@ -220,6 +220,9 @@ chr043416@gmail.com: davis.justin@mssm.org: avatar: https://avatars.githubusercontent.com/u/23462437?v=4 username: justincdavis +dgkim@deepx.ai: + avatar: https://avatars.githubusercontent.com/u/224481558?v=4 + username: deepx-dgkim esat@ultralytics.com: avatar: https://avatars.githubusercontent.com/u/43647848?v=4 username: artest08