Error during export of joint_conformer_fastspeech2_hifigan model

I am trying to export a joint_conformer_fastspeech2_hifigan model to be loaded as an ONNX model. The code I used:

```python
from espnet_onnx.export import TTSModelExport
m = TTSModelExport()
m.export(text2speech, 'custom_model', quantize=True)
```

I also tried the above with a pre-trained model:
```python
from espnet_onnx.export import TTSModelExport
tag_name = 'kan-bayashi/ljspeech_joint_train_conformer_fastspeech2_hifigan'
m = TTSModelExport()
m.export_from_pretrained(tag_name)
```

But in both cases I got the same error below:
```python
Traceback (most recent call last):
File "/Users/vigourav/development/tts_bitbucket/tts_inference/inference_api.py", line 853, in
m.export(text2speech, 'custom_model', quantize=True)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/espnet_onnx/export/tts/export_tts.py", line 56, in export
self._export_tts(tts_model, export_dir, verbose)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/espnet_onnx/export/tts/export_tts.py", line 173, in _export_tts
self._export_model(model, verbose, path)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/espnet_onnx/export/tts/export_tts.py", line 154, in _export_model
torch.onnx.export(
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/onnx/utils.py", line 516, in export
_export(
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/onnx/utils.py", line 1613, in _export
graph, params_dict, torch_out = _model_to_graph(
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/onnx/utils.py", line 1135, in _model_to_graph
graph, params, torch_out, module = _create_jit_graph(model, args)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/onnx/utils.py", line 1011, in _create_jit_graph
graph, torch_out = _trace_and_get_graph_from_model(model, args)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/onnx/utils.py", line 915, in _trace_and_get_graph_from_model
trace_graph, torch_out, inputs_states = torch.jit._get_trace_graph(
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/jit/_trace.py", line 1296, in _get_trace_graph
outs = ONNXTracedModule(
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
return forward_call(*args, **kwargs)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/jit/_trace.py", line 138, in forward
graph, out = torch._C._create_graph_by_tracing(
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/jit/_trace.py", line 129, in wrapper
outs.append(self.inner(*trace_inputs))
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
return forward_call(*args, **kwargs)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _slow_forward
result = self.forward(*input, **kwargs)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/espnet_onnx/export/tts/models/tts_models/fastspeech2.py", line 146, in forward
_, outs, d_outs, p_outs, e_outs = self._forward(
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/espnet_onnx/export/tts/models/tts_models/fastspeech2.py", line 195, in _forward
p_outs = self.pitch_predictor(hs, d_masks.unsqueeze(-1))
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
return forward_call(*args, **kwargs)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _slow_forward
result = self.forward(*input, **kwargs)
File "/Users/vigourav/anaconda3/envs/oracle_espnet/lib/python3.9/site-packages/espnet2/tts/fastspeech2/variance_predictor.py", line 84, in forward
xs = xs.masked_fill(x_masks, 0.0)
RuntimeError: masked_fill_ only supports boolean masks, but got mask with dtype float
```

Is Joint Fastspeech2 hifigan model not supported yet? If not, by when can we expect it?

Reproduce both export paths from the issue, then inspect espnet_onnx/export/tts/export_tts.py and espnet_onnx/export/tts/models/tts_models/fastspeech2.py around the failing export call. Trace the mask passed to espnet2/tts/fastspeech2/variance_predictor.py and compare the behavior for the joint_conformer_fastspeech2_hifigan model. Done means the supported export path no longer raises this mask-type error, or the model's unsupported status is clearly documented.

Written by the indexing model from the issue text.