Commits on Jan 6, 2022

1. Sort signature inputs by input key instead of tensor name when loadin… …

   …g a TF1 model in TF2.
   
   Signatures are saved using nest.flatten, which sorts the keys alphabetically for determinism. This can cause a mismatch between the lifted signature function's arg_keywords and structured_input_signature, resulting in users not being able to re-save the signature.
   
   Detailed example below.
   
   Say there is a SavedModel signature with the inputs:
   ```
     inputs['a'] tensor_info:
           dtype: DT_FLOAT
           shape: (-1, 1)
           name: input_2:0
       inputs['b'] tensor_info:
           dtype: DT_FLOAT
           shape: (-1, 2)
           name: input_1:0
   ```
   
   When this signature is loaded using the TF2 API, it becomes a concrete function with arg keywords = `['b', 'a']` (sorted by tensor name)
   
   The `structured_input_signature` becomes `([], {'b': spec_for_b, 'a': spec_for_a})`
   
   When this function is exported as a signature in a new SavedModel, the structured inputs are flattened and zipped with the arg keywords. However, the flattened `structured_input_signature` is `[spec_for_a, spec_for_b]`, since `nest.flatten` sorts the keys for determinism. This results in the wrong spec being mapped to the wrong arguments.
   
   PiperOrigin-RevId: 417889747
   Change-Id: Icdce72a43033027c18f5530068f2d6fd47e8e32d

   

   k-w-w committed Jan 6, 2022
   Copy the full SHA

   a7a884c View commit details

   Browse the repository at this point in the history