[DEP-291] NMS params passed at runtime for NMS-fused ONNX graph by dkosowski87 · Pull Request #2195 · roboflow/inference

Add function to retrieve comma-separated list of strings from environment variables in utils/environment.py

dkosowski87

committed

Add function to retrieve a string from environment variables with optional default value in utils/environment.py

dkosowski87

committed

Add default input name configurations for YOLO Ultralytics in configuration.py, utilizing new environment variable retrieval functions for strings and comma-separated lists.

dkosowski87

committed

Enhance YOLOv8 ONNX model input handling by adding support for multiple input names and validating against declared fused NMS input names. Update forward method to dynamically build input tensors b…

dkosowski87

committed

Refactor YOLOv8 ONNX model input validation to enhance error handling and logging. Update input name checks for fused NMS and modify forward method to conditionally build input tensors based on pro…

dkosowski87

committed

Add a new script for running fused NMS inference with YOLOv8 model. The script includes command-line options for specifying image and model paths, as well as parameters for confidence, IOU threshol…

dkosowski87

committed

Add a new script for running fused NMS inference with YOLOv8 model. The script includes command-line options for specifying image and model paths, as well as parameters for confidence, IOU threshol…

dkosowski87

committed

Integrate fused NMS post-processing in YOLOv8 ONNX model. Update the inference method to utilize the new post_process_nms_fused_model_output function for improved detection results when fused NMS i…

dkosowski87

committed

Add benchmarking options to fused NMS inference script. Introduced command-line parameters for specifying the number of benchmark iterations and warmup runs, allowing users to measure inference lat…

dkosowski87

committed

Enhance fused NMS inference script with JSON reporting. Added functions to generate and write latency and prediction reports in JSON format, including detailed statistics on inference latencies. Up…

dkosowski87

committed

Add command-line options for run name and target directory in fused NMS inference script. Updated output file paths to include run name as a subdirectory, enhancing organization of results and impr…

dkosowski87

committed

Add ONNX execution provider options to fused NMS inference script. Introduced a new command-line option for selecting ONNX Runtime execution providers (cpu, cuda, tensorrt) and updated the model lo…

dkosowski87

committed

Add batch processing capability to fused NMS inference script. Introduced a new command-line option for specifying batch size, allowing users to duplicate the input image for batched inference. Upd…

dkosowski87

committed

Refactor fused NMS inference script to support directory input for images. Updated command-line options to accept an image directory instead of a single image path, enabling batch processing of mul…

dkosowski87

committed

Update fused NMS inference script to improve batch processing and latency reporting. Introduced a constant for test batch size, enhanced JSON output to include image paths, and updated command-line…

dkosowski87

committed

Refactor input handling in YOLOv8 ONNX object detection to use single-element tensors for confidence, IOU threshold, and max detections. This change improves compatibility with the input builders a…

dkosowski87

committed

Add new script for running fused NMS inference with YOLOv8 model. Includes command-line options for image and model paths, as well as parameters for confidence, IOU threshold, and maximum detection…

dkosowski87

committed

Add new script for benchmarking fused NMS inference with YOLOv8 model. Includes command-line options for run name, image directory, model path, target directory, and parameters for confidence, IOU …

dkosowski87

committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[DEP-291] NMS params passed at runtime for NMS-fused ONNX graph#2195

[DEP-291] NMS params passed at runtime for NMS-fused ONNX graph#2195
dkosowski87 wants to merge 20 commits intomainfrom
DEP-291/configurable-NMS-threshold-for-YOLO

Commits on Apr 1, 2026

Add function to retrieve comma-separated list of strings from environment variables in utils/environment.py

Add function to retrieve a string from environment variables with optional default value in utils/environment.py

Add default input name configurations for YOLO Ultralytics in configuration.py, utilizing new environment variable retrieval functions for strings and comma-separated lists.

Enhance YOLOv8 ONNX model input handling by adding support for multiple input names and validating against declared fused NMS input names. Update forward method to dynamically build input tensors b…

Refactor YOLOv8 ONNX model input validation to enhance error handling and logging. Update input name checks for fused NMS and modify forward method to conditionally build input tensors based on pro…

Add a new script for running fused NMS inference with YOLOv8 model. The script includes command-line options for specifying image and model paths, as well as parameters for confidence, IOU threshol…

Add a new script for running fused NMS inference with YOLOv8 model. The script includes command-line options for specifying image and model paths, as well as parameters for confidence, IOU threshol…

Integrate fused NMS post-processing in YOLOv8 ONNX model. Update the inference method to utilize the new post_process_nms_fused_model_output function for improved detection results when fused NMS i…

Add benchmarking options to fused NMS inference script. Introduced command-line parameters for specifying the number of benchmark iterations and warmup runs, allowing users to measure inference lat…

Enhance fused NMS inference script with JSON reporting. Added functions to generate and write latency and prediction reports in JSON format, including detailed statistics on inference latencies. Up…

Add command-line options for run name and target directory in fused NMS inference script. Updated output file paths to include run name as a subdirectory, enhancing organization of results and impr…

Add ONNX execution provider options to fused NMS inference script. Introduced a new command-line option for selecting ONNX Runtime execution providers (cpu, cuda, tensorrt) and updated the model lo…

Add batch processing capability to fused NMS inference script. Introduced a new command-line option for specifying batch size, allowing users to duplicate the input image for batched inference. Upd…

Refactor fused NMS inference script to support directory input for images. Updated command-line options to accept an image directory instead of a single image path, enabling batch processing of mul…

Update fused NMS inference script to improve batch processing and latency reporting. Introduced a constant for test batch size, enhanced JSON output to include image paths, and updated command-line…

Refactor input handling in YOLOv8 ONNX object detection to use single-element tensors for confidence, IOU threshold, and max detections. This change improves compatibility with the input builders a…

Add new script for running fused NMS inference with YOLOv8 model. Includes command-line options for image and model paths, as well as parameters for confidence, IOU threshold, and maximum detection…

Add new script for benchmarking fused NMS inference with YOLOv8 model. Includes command-line options for run name, image directory, model path, target directory, and parameters for confidence, IOU …

Commits on Apr 2, 2026

Refactor ONNX input handling to support broadcast scalars and improve error messaging. Updated input validation to allow tensors with dimension 0 equal to 1 and enhanced error messages for incompat…

Add ONNX execution provider options to the fused NMS inference script. Introduced a new command-line option for selecting execution providers (cpu, cuda, tensorrt) and updated model loading to util…