fp16-optimization

Here are 5 public repositories matching this topic...

talhatahir-10xe / CuQwen

High-performance CUDA inference engine for Qwen models, written from scratch in C++/CUDA

deep-learning cplusplus hpc cuda transformers nvidia gpu-acceleration cuda-kernels nvidia-gpu cuda-programming llm llm-inference qwen qwen2-5 fp16-optimization

Updated Sep 26, 2026
Cuda

zhahoi / YOLOv11_TensorRT_fp16_int8

Star

针对YOLOv11进行fp16和ptq的int8量化，显著提升推理速度(C++) (包含完整模型转换流程和代码）

tensorrt int8-quantization yolov11 fp16-optimization

Updated Mar 13, 2026
C++

starrush26 / AhriEyes

Star

Ahrireyes 모델 및 파일들 저장소

python html ai deep-learning js public cnn free-software vit cnn-keras onnx fastapi efficientnet onnxruntime ai-model deepfakes-detection int8-quantization convnext fp16-optimization

Updated Sep 21, 2026
HTML

navaneet625 / ModernLPR_system

Star

High-performance LPR system optimized for Indian license plates, achieving 97% character accuracy. Features a hybrid pipeline using YOLOv11 and Mamba-SSM (State Space Models) with built-in regex correction and Beam Search decoding.

ocr computer-vision deep-learning pytorch beam-search anpr ctc-loss test-time-augmentation licence-plates-recognition trocr mamba-state-space-models yolov11 lora-fine-tuning fp16-optimization

Updated Dec 17, 2025
Jupyter Notebook

SayamAlt / Structural-Defects-Identification-using-YOLOv8n

Star

Successfully developed a lightweight YOLOv8n structural defect detector trained on 8,162 images across 7 defect classes, converted to ▎ TensorRT FP16 for edge deployment achieving 157.6 FPS — a 40.7% speedup over the FP32 baseline with under 1% accuracy loss.

computer-vision object-detection you-only-look-once object-localization model-quantization edge-inference yolov8n fp16-optimization

Updated Jun 7, 2026
Jupyter Notebook

Add this topic to your repo

To associate your repository with the fp16-optimization topic, visit your repo's landing page and select "manage topics."

Learn more