search

PINTO0309 / openvino2tensorflow

This script converts the ONNX/OpenVINO IR model to Tensorflow's saved_model, tflite, h5, tfjs, tftrt(TensorRT), CoreML, EdgeTPU, ONNX and pb. PyTorch (NCHW) -> ONNX (NCHW) -> OpenVINO (NCHW) -> openvino2tensorflow -> Tensorflow/Keras (NHWC/NCHW) -> TFLite (NHWC/NCHW). And the conversion from .pb to saved_model and from saved_model to .pb and from .pb to .tflite and saved_model to .tflite and saved_model to onnx. Support for building environments with Docker. It is possible to directly access the host PC GUI and the camera to verify the operation. NVIDIA GPU (dGPU) support. Intel iHD GPU (iGPU) support.
MIT License
338 stars 40 forks source link

tiny yolov4 #73

Closed Gaurav14cs17 closed 2 years ago

Gaurav14cs17 commented 3 years ago

tf_layers_dict: KerasTensor(type_spec=TensorSpec(shape=(1, 13, 13, 63), dtype=tf.float32, name=None), name='conv2d_17/Conv2D:0', description="created by layer 'conv2d_17'")

layer_type: Const layer_id: 316 tf_layers_dict: (1, 63, 1, 1)

layer_type: Add layer_id: 317 input_layer0: layer_id=315: KerasTensor(type_spec=TensorSpec(shape=(1, 13, 13, 63), dtype=tf.float32, name=None), name='conv2d_17/Conv2D:0', description="created by layer 'conv2d_17'") input_layer1: layer_id=316: Const(ndarray).shape (1, 63, 1, 1) tf_layers_dict: KerasTensor(type_spec=TensorSpec(shape=(1, 13, 13, 63), dtype=tf.float32, name=None), name='tf.math.add_17/Add:0', description="created by layer 'tf.math.add_17'")

The RegionYolo layer is not yet implemented.

PINTO0309 commented 2 years ago

Under investigation. It is easy to implement if you can replace the For loop with a generic layer. Suggestions are welcome. https://github.com/openvinotoolkit/openvino/blob/2021.4.2/ngraph/core/reference/include/ngraph/runtime/reference/region_yolo.hpp

PINTO0309 commented 2 years ago

There is no prospect of implementation, so it will be closed.