I will optimize and deploy your pytorch model with onnx and tensorrt
AI, Computer Vision Engineer
About this Gig
Your model works, but it's too slow or too heavy for production? I'll convert and optimize it with ONNX and TensorRT so it runs faster on GPUs and edge devices, and check that accuracy stays where it should.
I'm a Computer Vision Engineer with 3+ years shipping production CV systems. I optimize models with ONNX/TensorRT for kiosk and mobile devices and serve them through FastAPI in systems running in 60+ retail stores.
What I can do:
- Export PyTorch models to ONNX and verify outputs match the original
- Build TensorRT FP16 engines and measure the speed-up
- Benchmark latency and accuracy before vs after
- Fix common export problems (unsupported ops, dynamic shapes)
- Serve the optimized model with FastAPI + Docker
What you get:
- Optimized model files + inference code
- Benchmark report (latency, accuracy, hardware used)
- Clear instructions to run it on your machine
Please message me before ordering with your model type, framework version and target hardware (GPU model or edge device), so I can confirm the scope.
Expertise:
Classification
•
Software development
•
Object detection
Programming language:
Python
Tools:
OpenCV
•
TensorFlow
•
PyTorch
Frameworks:
PyTorch
