Add mishkal to requirements for diacritization support

UI improvements: dark header, white text, professional styling

Redesign UI: professional theme, clean header, fix header background

Update app.py: model selector, auto-diacritize, MSA pitch/speed, 200ms trim

Add model checkpoints via Git LFS

Fix Gradio 5 compatibility: move css/js to Blocks()

Update Gradio to latest version

Add HF Space deployment files

ui interface for inference

app for ui

Update .gitignore to include new dataset and model checkpoints; add hifigan-asc and tacotron2_epoch_0096 model files

Add BucketBatchSampler to optimize batching by mel length

Add mel length computation to NewOmaniTTSDataset and refactor finetune_speaker for bucket sampling

Add max_audio_seconds parameter to NewOmaniTTSDataset and filter long audio files

Refactor dataset downloading and extraction process in omani_speaker_finetune.ipynb

Add code to download processed Omani dataset from Google Drive in the notebook

Add preprocessing script and dataset inspection notebook for Omani TTS dataset

Refactor code structure for improved readability and maintainability

Implement code changes to enhance functionality and improve performance

Refactor code structure for improved readability and maintainability

Update requirements and add configuration section in fine-tuning notebook

Update Omani dataset download script to download and extract dataset zip from Google Drive

Update Omani dataset download script to improve error handling and ensure proper permissions

Update Omani dataset download script to handle Google throttling

Update Omani dataset download script to include sharing instructions and improve error handling

Update Omani speaker fine-tuning notebook: install requirements, download dataset, and refactor code structure

Add OmaniTTSDataset for annotated Omani corpus and fine-tuning notebook

cleander audio inference

hifi-gan vocoder for inference

Refactor WaveRNN training notebook for improved clarity and structure

Update hyperparameters, .gitignore, and WaveRNN training notebook

Update default checkpoint name in WaveRNNConfig for consistency

Add WaveRNN dataset class and update hyperparameters for training

Enhance Tacotron2 training notebook with data downloading and merging functionality

Add training visualization images for epochs 1 to 45 and update inference notebook with device handling in model.py

Resolve notebook merge conflicts; keep clean cells and empty outputs

Merge pull request #1 from HazardCodeBolt/cloud-changes

Merge branch 'main' into cloud-changes

Cloud changes

Merge branch 'main' of https://github.com/HazardCodeBolt/tacotron2-wavernn

Refactor audio processing in AudioMelConversions and update inference methods in Tacotron2

Remove commented-out code in model.py

Remove unnecessary whitespace in model.py

Update .gitignore to include tacotron2-wavernn files; modify inference notebook to reset execution count and clean up output cells

Refactor dataset path handling in hyperparams and model; update .gitignore to include environment files

Update dataset path in hyperparams and model; add training test script

Add support for Arabic characters in Tokenizer and update imports

Add audio processing functions and update imports in training notebook

first commit