tf.nn.embedding_lookup_sparse and tf.nn.safe_embedding_lookup_sparse now support ids and weights described by tf.RaggedTensors. This can cause breakages where existing calls pass Tensors with the wrong shape or omit certain non-Tensor arguments (including default values). ConcreteFunction ( tf.) as generated through get_concrete_function now performs holistic input validation similar to calling tf.function directly. add_op supports broadcasting up to 6 dimensions. Added reference implementation for 16-bit int and 32-bit unsigned int unquantized mul. 
Added reference implementation for 16-bit int unquantized add.Added 8-bit/16-bit/32-bit int/uint support for the built-in op right_shift.Added int16 indices support for built-in op gather and gather_nd.Added 8-bit/16-bit/32-bit int/uint support for the built-in op bitwise_xor.Added 16-bit and 32-bit int support for the built-in op bitcast.Added 8-bit and 16-bit support for floor_div and floor_mod.
Added 16-bit int type support for built-in op less, greater_than, equal. Added int16x8 support for the built-in ops space_to_batch_nd and batch_to_space_nd. Added int16x8 support for the built-in op mirror_pad. Added int16x8 support for the built-in op exp. The Python TF Lite Interpreter bindings now have an option experimental_disable_delegate_clustering to turn-off delegate clustering. Added 16-bit and 64-bit float type support for built-in op cast. The LMDB dependency that these kernels are bringing to TensorFlow has been dropped, thus making the build slightly faster and more secure. This is in preparation for completely removing them from TensorFlow. 
The LMDB kernels have been changed to return an error.Release 2.13.0 TensorFlow Breaking Changes
0 kommentar(er)
