Support tf.nn.leaky_relu and fix advanced_activations (#514)

fixing the nightly build

a fixing for gpt-2 subclassing model. (#368)

* remove tf2onnx api

* migrate to tf2onnx

* remove tf2onnx test cases and package.

* minor corrections

* update the version and docs

* doc fixing

* a fixing for gpt-2 subclassing model.

Update onnx 1.6.0 to CI build (#284)

Add a unit test for MergePadConvSolution: [0, 0, 1, 1] + Conv2D (#254)

* Add a unit test: ZeroPadding2D (0, 0, 1, 1) + Conv2D to test MergePadConvSolution

* Add a unit test: ZeroPadding2D (0, 0, 1, 1) + Conv2D to test MergePadConvSolution

Unify the batch size dim in the model conversion. (#171)

Updating package version to 1.5.0 (#113)