Convert YOLO ONNX models to CVIMODEL for Sophgo CV181x TPU deployment.
npx skills add https://github.com/Seeed-Studio/ai-skills --skill onnx-to-cvimodel
Please help me install this Agent Skill. Skill: onnx-to-cvimodel Source: https://github.com/Seeed-Studio/ai-skills/tree/main/skills/onnx-to-cvimodel Command: npx skills add https://github.com/Seeed-Studio/ai-skills --skill onnx-to-cvimodel
Converts YOLO ONNX models into CVIMODEL format for deployment on Sophgo CV181x TPU platforms, simplifying the end-to-end conversion workflow from ONNX to deployable CVIMODEL artifacts.
Run the universal ONNX-to-CVIMODEL conversion script with your ONNX model and calibration dataset to generate a CVIMODEL artifact ready for TPU deployment.
High-intent search queries and answers about installing and using this skill.
To convert a YOLO ONNX model to CVIMODEL, you need a Docker-based TPU-MLIR environment, your ONNX model file, and a calibration dataset. The conversion script processes these inputs to generate a deployable CVIMODEL artifact for the Sophgo CV181x TPU platform.
Yes, the ONNX to CVIMODEL conversion supports YOLO11 and YOLO26 workflows. It includes task-specific scripts covering detect, pose, seg, and cls, allowing you to convert various YOLO model types for CV181x TPU deployment.
Before converting ONNX models to CVIMODEL, you must set up a Docker-based TPU-MLIR environment. You also need the source ONNX model file and a calibration dataset to perform the quantization and conversion process successfully.
Yes, you can batch convert multiple YOLO ONNX models to CVIMODEL format. The conversion process includes batch processing capabilities and optional quantization support with included qtables to produce deployment-ready outputs.
The best way to apply quantization during YOLO ONNX to CVIMODEL conversion is by using the included optional qtables. The conversion script integrates these quantization tables with your calibration dataset to generate a deployment-ready CVIMODEL.
YOLO ONNX to CVIMODEL conversion requires a Docker environment because it relies on the TPU-MLIR toolchain. Docker ensures the necessary TPU-MLIR dependencies and runtime environment are consistently configured to process the ONNX model and calibration dataset.
zenobi-usDeploy scalable ML models with real-time inference, monitoring, and CI/CD pipelines.
gizixPackage trained machine learning models with versioning, ONNX export, and FastAPI endpoints.
amdExecute end-to-end post-training quantization workflows for ONNX models with AMD Quark.
Pico-DeveloperConvert ONNX, TFLite, PB, and PyTorch models to QNN context binaries.
