seed from hparams for TextMelLoader

[bug-fix] pillow dependency in Dockerfile

build(deps): bump tensorflow from 1.12.0 to 1.15.2

Bumps [tensorflow](https://github.com/tensorflow/tensorflow) from 1.12.0 to 1.15.2.
- [Release notes](https://github.com/tensorflow/tensorflow/releases)
- [Changelog](https://github.com/tensorflow/tensorflow/blob/master/RELEASE.md)
- [Commits](https://github.com/tensorflow/tensorflow/compare/v1.12.0...v1.15.2)

Signed-off-by: dependabot[bot] <support@github.com>

fix error when batch size = 1

remove tensorboardX; use torch.utils.tensorboard

the same as [requirements.txt](https://github.com/NVIDIA/tacotron2/blob/master/requirements.txt#L9)
```
Traceback (most recent call last):
  File "train.py", line 290, in <module> 
    args.warm_start, args.n_gpus, args.rank, args.group_name, hparams)
  File "train.py", line 248, in train
    hparams.distributed_run, rank)
  File "train.py", line 146, in validate 
    logger.log_validation(val_loss, model, y, y_pred, iteration)
  File "/workspace/tacotron2/logger.py", line 34, in log_validation
    iteration)
  File "/opt/conda/lib/python3.6/site-packages/tensorboardX/writer.py", line 584, in add_image
    image(tag, img_tensor, dataformats=dataformats), global_step, walltime)
  File "/opt/conda/lib/python3.6/site-packages/tensorboardX/summary.py", line 271, in image
    image = make_image(tensor, rescale=rescale)
  File "/opt/conda/lib/python3.6/site-packages/tensorboardX/summary.py", line 307, in make_image
    from PIL import Image
ModuleNotFoundError: No module named 'PIL'
```

Fixed link to download waveglow from inference.py

Fixing concatenation error for fp16 distributed training

Clean slate