item embedding and user sequence item embedding can not use the same embedding variable

[uzhy1987]

System information

• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): mac m1
• TensorFlow version and how it was installed (source or binary): binary
• TensorFlow-Recommenders-Addons version and how it was installed (source or binary): 0.7.2 binary
• Python version: 3.9.6
• Is GPU used? (yes/no): no

Describe the bug

File "/opt/python/tensorflow/lib/python3.9/site-packages/keras/src/utils/traceback_utils.py", line 70, in error_handler
raise e.with_traceback(filtered_tb) from None
File "/opt/python/tensorflow/lib/python3.9/site-packages/tensorflow_recommenders_addons/dynamic_embedding/python/ops/dynamic_embedding_optimizer.py", line 739, in apply_gradients_strategy_v2_lagacy
return self._distributed_apply(strategy, grads_and_vars, name,
File "/opt/python/tensorflow/lib/python3.9/site-packages/tensorflow_recommenders_addons/dynamic_embedding/python/ops/dynamic_embedding_optimizer.py", line 204, in _distributed_apply
update_op = distribution.extended.update(var,
File "/opt/python/tensorflow/lib/python3.9/site-packages/tensorflow_recommenders_addons/dynamic_embedding/python/ops/dynamic_embedding_optimizer.py", line 172, in apply_grad_to_update_var
_apply_op = self._resource_apply_sparse_duplicate_indices(
tensorflow.python.framework.errors_impl.InvalidArgumentError: {{function_node _wrapped__ResourceScatterAdd_device/job:localhost/replica:0/task:0/device:CPU:0}} indices[758] = 758 is not in [0, 758) [Op:ResourceScatterAdd] name:

A clear and concise description of what the bug is.

This problem occurs when the user behavior sequence features reuse embbeding from item, but not when used separately
当用户行为序列特征复用item的embbeding时会出现这个问题，分开使用时没有问题

Code to reproduce the issue

for f in sparse_cols:
    _input = sparse_inputs[f]
    embeddings[f] = de.keras.layers.Embedding(
        embed_dim, initializer=keras.initializers.RandomNormal(seed=25), name=f"emb_{f}",
        # init_capacity=lookups[f].vocabulary_size(), restrict_policy=de.FrequencyRestrictPolicy
    )
    input_embed[f] = Flatten()(embeddings[f](_input))

for f in varlens_inputs:
    _input = varlens_inputs[f] 
    mask = Masking(mask_value=0).compute_mask(_input)
    embedding = embeddings["bid"] if "_bids" in f else de.keras.layers.Embedding(
        embed_dim, initializer=keras.initializers.RandomNormal(seed=30), name=f"emb_{f}",
        # init_capacity=lookups[f].vocabulary_size(), restrict_policy=de.FrequencyRestrictPolicy
    )
    _embed = Reshape([-1, embed_dim])(embedding(_input))
    out_embed = MeanPoolLayer(axis=1)(_embed, mask)
    input_embed[f] = out_embed

Provide a reproducible test case that is the bare minimum necessary to generate the problem.

Other info / logs

Include any logs or source code that would be helpful to diagnose the problem. If including tracebacks, please include the full traceback. Large logs and files should be attached.

[MoFHeka]

TFRA Keras Embedding layers could not be called twice and more for now. Please concat all inputs in one tensor and then lookup. If you use GPU kernel, this would also be better performance.