DebertaV3MaskedLM example don't work

[mrektor]

Describe the bug

The example from the (documentation)[https://keras.io/api/keras_nlp/models/deberta_v3/deberta_v3_masked_lm/] does not work

To Reproduce

import os

os.environ["KERAS_BACKEND"] = "tensorflow"  # or "tensorflow" or "torch" or "jax"

import keras_nlp
import keras

# Use mixed precision to speed up all training in this guide.
keras.mixed_precision.set_global_policy("mixed_float16")

BATCH_SIZE = 16


features = ["The quick brown fox jumped.", "I forgot my homework."]

masked_lm = keras_nlp.models.DebertaV3MaskedLM.from_preset(
    "deberta_v3_base_en",
)
masked_lm.fit(x=features, batch_size=2)

Expected behavior
The training loop should start

Additional context
Throws instead this nasty error:

Traceback (most recent call last):
  File "/home/ettore.mariotti/deberta_gal/deberta_keras.py", line 19, in <module>
    masked_lm.fit(x=features, batch_size=2)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras_nlp/src/utils/pipeline_model.py", line 181, in fit
    x = x.map(
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/data/ops/dataset_ops.py", line 2299, in map
    return map_op._map_v2(
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/data/ops/map_op.py", line 40, in _map_v2
    return _ParallelMapDataset(
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/data/ops/map_op.py", line 148, in __init__
    self._map_func = structured_function.StructuredFunctionWrapper(
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/data/ops/structured_function.py", line 265, in __init__
    self._function = fn_factory()
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py", line 1251, in get_concrete_function
    concrete = self._get_concrete_function_garbage_collected(*args, **kwargs)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py", line 1221, in _get_concrete_function_garbage_collected
    self._initialize(args, kwargs, add_initializers_to=initializers)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py", line 696, in _initialize
    self._concrete_variable_creation_fn = tracing_compilation.trace_function(
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compilation.py", line 178, in trace_function
    concrete_function = _maybe_define_function(
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compilation.py", line 283, in _maybe_define_function
    concrete_function = _create_concrete_function(
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compilation.py", line 310, in _create_concrete_function
    traced_func_graph = func_graph_module.func_graph_from_py_func(
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/framework/func_graph.py", line 1059, in func_graph_from_py_func
    func_outputs = python_func(*func_args, **func_kwargs)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/eager/polymorphic_function/polymorphic_function.py", line 599, in wrapped_fn
    out = weak_wrapped_fn().__wrapped__(*args, **kwds)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/data/ops/structured_function.py", line 231, in wrapped_fn
    ret = wrapper_helper(*args)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/data/ops/structured_function.py", line 161, in wrapper_helper
    ret = autograph.tf_convert(self._func, ag_ctx)(*nested_args)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/autograph/impl/api.py", line 693, in wrapper
    raise e.ag_error_metadata.to_exception(e)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/autograph/impl/api.py", line 690, in wrapper
    return converted_call(f, args, kwargs, options=options)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/autograph/impl/api.py", line 439, in converted_call
    result = converted_f(*effective_args, **kwargs)
  File "/tmp/__autograph_generated_file80f1z6yd.py", line 36, in tf__preprocess_samples
    ag__.if_stmt(ag__.ld(self).preprocessor is not None, if_body, else_body, get_state, set_state, ('do_return', 'retval_'), 2)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/autograph/operators/control_flow.py", line 1217, in if_stmt
    _py_if_stmt(cond, body, orelse)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/autograph/operators/control_flow.py", line 1270, in _py_if_stmt
    return body() if cond else orelse()
  File "/tmp/__autograph_generated_file80f1z6yd.py", line 23, in if_body
    retval_ = ag__.converted_call(ag__.ld(self).preprocessor, (ag__.ld(x),), dict(y=ag__.ld(y), sample_weight=ag__.ld(sample_weight)), fscope)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/autograph/impl/api.py", line 439, in converted_call
    result = converted_f(*effective_args, **kwargs)
  File "/tmp/__autograph_generated_fileyo02hq6s.py", line 12, in tf____call__
    outputs = ag__.converted_call(ag__.converted_call(ag__.ld(super), (), None, fscope).__call__, tuple(ag__.ld(args)), dict(**ag__.ld(kwargs)), fscope)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/autograph/impl/api.py", line 377, in converted_call
    return _call_unconverted(f, args, kwargs, options)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/tensorflow/python/autograph/impl/api.py", line 459, in _call_unconverted
    return f(*args, **kwargs)
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras/src/utils/traceback_utils.py", line 122, in error_handler
    raise e.with_traceback(filtered_tb) from None
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras_nlp/src/models/deberta_v3/deberta_v3_masked_lm_preprocessor.py", line 151, in build
    vocabulary_size=self.tokenizer.vocabulary_size(),
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras_nlp/src/models/deberta_v3/deberta_v3_tokenizer.py", line 135, in vocabulary_size
    sentence_piece_size = super().vocabulary_size()
  File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras_nlp/src/tokenizers/sentence_piece_tokenizer.py", line 185, in vocabulary_size
    return int(self._sentence_piece.vocab_size().numpy())
AttributeError: in user code:

    File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras_nlp/src/models/task.py", line 98, in preprocess_samples  *
        return self.preprocessor(x, y=y, sample_weight=sample_weight)
    File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras_nlp/src/layers/preprocessing/preprocessing_layer.py", line 49, in __call__  *
        outputs = super().__call__(*args, **kwargs)
    File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras/src/utils/traceback_utils.py", line 122, in error_handler  **
        raise e.with_traceback(filtered_tb) from None
    File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras_nlp/src/models/deberta_v3/deberta_v3_masked_lm_preprocessor.py", line 151, in build
        vocabulary_size=self.tokenizer.vocabulary_size(),
    File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras_nlp/src/models/deberta_v3/deberta_v3_tokenizer.py", line 135, in vocabulary_size
        sentence_piece_size = super().vocabulary_size()
    File "/home/ettore.mariotti/deepenv/lib64/python3.9/site-packages/keras_nlp/src/tokenizers/sentence_piece_tokenizer.py", line 185, in vocabulary_size
        return int(self._sentence_piece.vocab_size().numpy())

    AttributeError: 'SymbolicTensor' object has no attribute 'numpy'

Would you like to help us fix it?
Yes please

[SuryanarayanaY]

Hi @mrektor ,

I have replicated the reported issue and attached gist here.

[mrektor]

So I guess this has not been fixed yet?