TypeError: Input 'y' of 'AddV2' Op has type float16 that does not match type float32 of argument 'x'.

[ruizhidu]

Description

when use HPARAMS=universal_transformer_base_fp16 or HPARAMS=transformer_fairseq_fp16_activation_big , tensor2tensor not work

/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensor2tensor/utils/t2t_model.py:325 call  *
    sharded_logits, losses = self.model_fn_sharded(sharded_features)
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensor2tensor/utils/t2t_model.py:402 model_fn_sharded  *
    sharded_logits, sharded_losses = dp(self.model_fn, datashard_to_features)
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensor2tensor/utils/expert_utils.py:231 __call__  *
    outputs.append(fns[i](*my_args[i], **my_kwargs[i]))
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensor2tensor/utils/t2t_model.py:429 model_fn  *
    body_out = self.body(transformed_features)
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensor2tensor/models/transformer.py:246 body  *
    encoder_output, encoder_decoder_attention_bias = self.encode(
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensor2tensor/models/transformer.py:201 encode  *
    self._encoder_function, inputs, target_space, hparams,
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensor2tensor/models/transformer.py:103 transformer_encode  *
    encoder_output = encoder_function(
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensor2tensor/layers/transformer_layers.py:242 transformer_encoder  *
    x = common_layers.layer_postprocess(x, y, hparams)
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensor2tensor/layers/common_layers.py:993 layer_postprocess  *
    layer_output,
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensor2tensor/layers/common_layers.py:915 layer_prepostprocess  *
    x += previous_value
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensorflow/python/ops/math_ops.py:984 binary_op_wrapper
    return func(x, y, name=name)
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensorflow/python/ops/math_ops.py:1276 _add_dispatch
    return gen_math_ops.add_v2(x, y, name=name)
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensorflow/python/ops/gen_math_ops.py:483 add_v2
    "AddV2", x=x, y=y, name=name)
/home/username/.pyenv/versions/3.7.0/lib/python3.7/site-packages/tensorflow/python/framework/op_def_library.py:506 _apply_op_helper
    inferred_from[input_arg.type_attr]))

TypeError: Input 'y' of 'AddV2' Op has type float16 that does not match type float32 of argument 'x'.

Environment information

mesh-tensorflow 0.1.16 tensor2tensor 1.15.7 tensorboard 2.2.2 tensorboard-plugin-wit 1.7.0 tensorflow-addons 0.10.0 tensorflow-datasets 3.1.0 tensorflow-estimator 2.2.0 tensorflow-gan 2.0.0 tensorflow-gpu 2.2.0 tensorflow-hub 0.8.0 tensorflow-metadata 0.22.2 tensorflow-probability 0.7.0 centos7 python3.7