ONNX Runtimeã§ãƒ—ãƒãƒ•ã‚¡ã‚¤ãƒ«ã‚’å–ã£ã¦ã¿ã‚‹ - QiitaRasperry Pi 4ã®CPUã§Deep Learningを高速化ã®ç¶šãã§ã™ã€‚ Deep Learningを高速化ã™ã‚‹ãŸã‚ã«ã¯ã€ã©ã®å‡¦ç†ãŒã©ã‚Œãらã„ã®æ™‚間を消費ã—ã¦ã„ã‚‹ã‹ã‚’調査ã—ã¦ã€å®Ÿéš›ã®å‡¦ç†æ™‚間を削減ã—ã¦ã„ãå¿…è¦ãŒã‚ã‚Šã¾ã™ã€‚ãã®ãŸã‚ã€ã¾ãšã¯ONNX Runtimeã®ãƒ—ãƒãƒ•ã‚¡ã‚¤ãƒ«æ©Ÿèƒ½ã‚’使用ã—ã¦ãƒ—ãƒãƒ•ã‚¡ã‚¤ãƒªãƒ³ã‚°ã‚’è¡Œã„ã¾ã™ã€‚ プãƒãƒ•ã‚¡ã‚¤ãƒ«æ©Ÿèƒ½ã‚’有効ã«ã™ã‚‹æ–¹æ³•ã¯ONNXå…¬å¼ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ã«è¨˜è¼‰ã•ã‚Œã¦ã„ã¾ã™ã€‚ import onnxruntime options = onnxruntime.SessionOptions() options.enable_profiling = True # <- プãƒãƒ•ã‚¡ã‚¤ãƒ«æ©Ÿèƒ½æœ‰åŠ¹åŒ– session = onnxruntime.InferenceSession(path_to_model, options) [プãƒãƒ•ã‚¡ã‚¤ãƒ«å¯¾è±¡] prof_file = se
ONNXモデルã®ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ãƒ†ã‚¯ãƒ‹ãƒƒã‚¯ (応用編２)
GitHub - PINTO0309/onnx2tf: Self-Created Tools to convert ONNX files (NCHW) to TensorFlow/TFLite/Keras format (NHWC). The purpose of this tool is to solve the massive Transpose extrapolation problem in onnx-tensorflow (onnx-tf). I don't need a Star, but g
GitHub - PINTO0309/PINTO_model_zoo: A repository for storing models that have been inter-converted between various frameworks. Supported frameworks are TensorFlow, PyTorch, ONNX, OpenVINO, TFJS, TFTRT, TensorFlowLite (Float32/16/INT8), EdgeTPU, CoreML.
Whisperã®ãƒ¢ãƒ‡ãƒ«ã‚’ONNXã«ã™ã‚‹ - TadaoYamaokaã®é–‹ç™ºæ—¥è¨˜
onnxconverter-common/onnxconverter_common/auto_mixed_precision_model_path.py at master · microsoft/onnxconverter-common
音声èªè˜AIã®Whisperã‚’Unreal Engineã§ãƒªã‚¢ãƒ«ã‚¿ã‚¤ãƒ ã«å‹•ã‹ã™ãŸã‚ã«ã‚„ã£ãŸã“ã¨
ONNXモデルã®ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ãƒ†ã‚¯ãƒ‹ãƒƒã‚¯ (基礎編)
TadaoYamaokaã®é–‹ç™ºæ—¥è¨˜
Windows - DirectML
onnxruntime-directml
CUDAExecutionProvider Not Available / GPU Not Visible on NVIDIA T4 · Issue #7748 · microsoft/onnxruntime
onnxruntime-gpu failing to find onnxruntime_providers_shared.dll when run from a pyinstaller-produced exe file of the project
{{#tags}}- {{label}}
{{/tags}}