tensorflow-onnx
tensorflow-onnx copied to clipboard
onnx
Convert TensorFlow, Keras, Tensorflow.js and Tflite models to ONNX
**Describe the bug** Not sure if it is the intended behavior. I was using keras to build the model and then convert to ONNX through `tf2onnx.convert.from_keras` However, I notice the...
**Describe the bug** I'm sorry if this is a silly thing but I'm new at working with onnx. I converted a tf model to onnx and I have this issue:...
**Describe the bug** Fail to convert a batch normalization layer (written by `tf.keras`) when using tensorflow version `1.14.0` **System information** - OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Ubuntu...
Hello, I don't know if this is a bug or a feature in tf2onnx. When I convert a tensorflow model which contains multiple residual blocks, the constant in the Multiply...
Trying to convert LSTM to onnx WARNING - ONNX Failed to infer shapes and dtypes for [model_1/lstm/PartitionedCall/strided_slice_2, type: Slice] Traceback (most recent call last): File "/usr/local/lib/python3.7/dist-packages/tf2onnx/schemas.py", line 154, in infer_onnx_shape_dtype...
**Describe the bug** Converting TF1 .meta weights to ONNX yields to ONNX parsing error. **Urgency** Moderate **System information** - OS Platform and Distribution (e.g., Linux Ubuntu 16.04): NAME="Ubuntu" VERSION="20.04.3 LTS...
Hi, I encountered a problem when I converted a tf model to onnx format. 1. The tf model can be downloaded from: https://tfhub.dev/google/movenet/multipose/lightning/1 2. use tf2onnx to convert to onnx...
**Describe the bug** Now, tf2onnx only supports **RaggedGather** op rank == 1. But there are some NLP models have rank > 1 scenarios. **System information** - OS Platform and Distribution:...
When call tf2onnx.convert.from_graph_def API in a python process using tensorflow_gpu, from_graph_def will use all the gpu memory, make it difficult for someone else who shares the GPU card with me....
Currently, `CombinedNonMaxSuppression` converter does not support class-specific boxes, i.e. `q == num_class` case in https://www.tensorflow.org/api_docs/python/tf/image/combined_non_max_suppression. https://github.com/onnx/tensorflow-onnx/blob/01ec092384a8b2096425c2a5de7deb7b3215afdf/tf2onnx/onnx_opset/tensor.py#L1873-L1874 Apparently, "official" TF FasterRCNN / MaskRCNN models uses such variant of combined nms. So...
