编译深度学习模型

TVM 包括各种前端，可以导入不同格式的模型。这些教程演示了如何使用 Python API 导入模型。

• 编译 PyTorch 模型
  • 载入 PyTorch 预训练模型
  • 加载测试图片
  • 导入 Graph 到 Relay
  • Relay 构建
  • 在 TVM 上执行可移植 Graph
  • 查找 synset 名称
• Compile Tensorflow Models
  • Tutorials
  • Download required files
  • Import model
  • Decode image
  • Import the graph to Relay
  • Relay Build
  • Execute the portable graph on TVM
  • Process the output
  • Inference on tensorflow
• 编译 MXNet 模型
  • 从 gluon 模型动物园下载 Resnet18 模型
  • 编译 Graph
  • 在 TVM 上执行可移植 graph
  • 使用带有预训练权重的 MXNet symbol
• Compile ONNX Models
  • Load pretrained ONNX model
  • Load a test image
  • Compile the model with relay
  • Execute on TVM
  • Display results
  • Notes
• Compile Keras Models
  • Load pretrained keras model
  • Load a test image
  • Compile the model with Relay
  • Execute on TVM
  • Look up synset name
• Compile TFLite Models
  • Utils for downloading and extracting zip files
  • Load pretrained TFLite model
  • Load a test image
  • Compile the model with relay
  • Execute on TVM
  • Display results
• Compile CoreML Models
  • Load pretrained CoreML model
  • Load a test image
  • Compile the model on Relay
  • Execute on TVM
  • Look up synset name
• Compile YOLO-V2 and YOLO-V3 in DarkNet Models
  • Choose the model
  • Download required files
  • Import the graph to Relay
  • Load a test image
  • Execute on TVM Runtime
• Compile Caffe2 Models
  • Load pretrained Caffe2 model
  • Load a test image
  • Compile the model on Relay
  • Execute on TVM
  • Look up synset name
• Compile OneFlow Models
  • Load a pretrained OneFlow model and save model
  • Load a test image
  • Import the graph to Relay
  • Relay Build
  • Execute the portable graph on TVM
  • Look up synset name
• Compile PaddlePaddle Models
  • Load pretrained ResNet50 model
  • Load a test image
  • Compile the model with relay
  • Execute on TVM
  • Look up synset name