diff --git a/mindinsight/mindconverter/README.md b/mindinsight/mindconverter/README.md index 5273364b..73b6f4dd 100644 --- a/mindinsight/mindconverter/README.md +++ b/mindinsight/mindconverter/README.md @@ -156,7 +156,7 @@ Supported models list (Models in below table have been tested based on PyTorch 1 | DenseNet161 | [Link](https://github.com/pytorch/vision/blob/v0.5.0/torchvision/models/densenet.py) | / | | | NASNetMobile/Large | / | [Link](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/keras/applications/nasnet.py) | | | EfficientNetB0~B7 | [Link](https://github.com/lukemelas/EfficientNet-PyTorch) | [TF1.5Link](https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet) [TF2.3Link](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/keras/applications/efficientnet.py) | | -| Unet | [Link](https://github.com/milesial/Pytorch-UNet) | [Link](https://github.com/zhixuhao/unet) | Due to Operator `ResizeBilinear` not achieved on GPU device, Operator `ResizeBilinear` should be replaced by operator `ResizeNearest`, while running in GPU device | +| Unet | [Link](https://github.com/milesial/Pytorch-UNet) | [Link](https://github.com/zhixuhao/unet) | Due to Operator `mindspore.ops.ResizeBilinear` is not implemented on GPU device for now, operator `mindspore.ops.ResizeBilinear` should be replaced by operator `mindspore.ops.ResizeNearestNeighbor`, while running in GPU device | ## Example @@ -311,7 +311,7 @@ class ConvBNReLU(nn.Sequential): ## Requirements -For users using MindConverter, in addition to install the TensorFlow can satisfy the pb model loading, inference and training, users also need to pip install the following third party package (tf2onnx is not required for users that convert PyTorch model definition script to MindSpore): +For users using MindConverter, in addition to install the TensorFlow that can satisfy the model loading, inference and training requirements, users also need to pip install the following third party package (tf2onnx is not required for users that convert PyTorch model definition script to MindSpore): ```text onnx>=1.8.0 @@ -417,7 +417,7 @@ def convert_to_froze_graph(keras_model: tf.python.keras.models.Model, model_name | GraphInitFailError | Fail to trace the computational graph | 1000000 | Exception caused by 1000001~1000003 | | ModelNotSupportError | Fail to parse .pth/.pb file | 1000001 | Given `--input_nodes`, `--output_nodes don't match the input model; Meanwhile, the model file can not be loaded also can cause this error. | | TfRuntimeError | Fail to initialize the TF runtime | 1000002 | Resources required by TensorFlow are not available | -| ModelLoadingFailError | Fail to load the model | 1000003 | Maybe cause by the wrong `--input_shape` value | +| ModelLoadingError | Fail to load the model | 1000003 | Maybe cause by the wrong `--input_shape` value | | RuntimeIntegrityError | Fail to locate required third party dependency | 1000004 | Caused by required third party packages are not installed | | TreeCreateFailError | Fail to create code hierarchical tree | 2000000 | Mainly caused by usage of `torch.nn.functional.xxx`, `torch.xxx`, `torch.Tensor.xxx` in PyTorch | | NodeInputMissingError | Fail to get the input node info | 2000001 | Fail to get input node info | diff --git a/mindinsight/mindconverter/README_CN.md b/mindinsight/mindconverter/README_CN.md index 265a7633..59e0bca1 100644 --- a/mindinsight/mindconverter/README_CN.md +++ b/mindinsight/mindconverter/README_CN.md @@ -155,7 +155,7 @@ MindConverter提供两种技术方案，以应对不同脚本迁移场景： | DenseNet161 | [脚本链接](https://github.com/pytorch/vision/blob/v0.5.0/torchvision/models/densenet.py) | 暂未测试 | | | NASNetMobile/Large | 暂未测试 | [脚本链接](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/keras/applications/nasnet.py) | | | EfficientNetB0~B7 | [脚本链接](https://github.com/lukemelas/EfficientNet-PyTorch) | [TF1.5脚本链接](https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet) [TF2.3脚本链接](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/keras/applications/efficientnet.py) | | -| Unet | [脚本链接](https://github.com/milesial/Pytorch-UNet) | [脚本链接](https://github.com/zhixuhao/unet) | 由于算子`ResizeBilinear`在GPU上未实现，所以当运行在GPU设备上时，算子`ResizeBilinear`需要被替换为算子`ResizeNearest` | +| Unet | [脚本链接](https://github.com/milesial/Pytorch-UNet) | [脚本链接](https://github.com/zhixuhao/unet) | 由于算子`mindspore.ops.ResizeBilinear`在GPU上暂未实现，所以当运行在GPU设备上时，算子`mindspore.ops.ResizeBilinear`需要被替换为算子`mindspore.ops.ResizeNearestNeighbor` | ## 使用示例 @@ -431,7 +431,7 @@ def convert_to_froze_graph(keras_model: tf.python.keras.models.Model, model_name | GraphInitFailError | 依据网络模型构建计算图失败 | 1000000 | 由1000001，1000002，1000003导致的计算图无法解析。 | | ModelNotSupportError | 解析.pth/.pb文件失败 | 1000001 | 给定的`--input_nodes`, `--output_nodes`与实际模型不符；或模型文件存在问题导致模型无法加载。 | | TfRuntimeError | TensorFlow库执行出错 | 1000002 | TensorFlow启动申请所需资源失败导致无法正常启动，请检查系统资源（进程数、内存、显存占用、CPU占用）是否充足。 | -| ModelLoadingFailError | 模型加载失败 | 1000003 | 可能由于用户给定网络输入尺寸错误导致模型无法加载。 | +| ModelLoadingError | 模型加载失败 | 1000003 | 可能由于用户给定网络输入尺寸错误导致模型无法加载。 | | RuntimeIntegrityError | 三方依赖库不完整 | 1000004 | MindConverter运行时所需的三方依赖库未安装。 | | TreeCreateFailError | 依据计算图构建模型树失败 | 2000000 | Tree用于生成最终代码结构，通常由于PyTorch网络中存在`torch.nn.functional.xxx`, `torch.xxx`, `torch.Tensor.xxx`算子导致。 | | NodeInputMissingError | 网络节点输入信息丢失 | 2000001 | 节点的输入信息丢失。 |