zabir-nabil / bangla-tts

Bangla text to speech, Multilingual (Bangla, English) real-time ([almost] in a GPU) speech synthesis library

Bonjour, je suis nabil 👋

🐼 私について

Software engineer with over 5 years of experience in machine learning, specializing in deep learning, computer vision, and NLP applications. My current tech-stack is very much Python and GCP-centric. Currently diving into LLMs!

Research interests are explainable multi-task learning, LLMs, biomedical image analysis, speech synthesis, meta-learning, and xAI. At my current role, I do prompt engineering, write ETL pipelines, and fine-tune LLMs. I'm always chiming in on StackOverflow, helping out where I can. Oh, and by the way,

I am also a GCP certified professional ML Engineer - Certification

🌏 (https://zabir.quest)

🔎 Find me

⚡ Technologies

bangla-tts's Issues

different model

Unable to open table file ./model1/checkpoint_13000: Data loss: not an stable (bad magic number): perhaps your file is in a different file format and you need to use a different restore operator?

is possible to run with own model with tactron.

multiple conflicts when installing the dependencies using requirements.txt file

tensorflow requires multiple packages to be from an older version than mentioned in requirement file

Data loss error

The above error is being thrown when we try to load the weight from the uploaded file.

Error

Taco2 Bn tts request

Bhai salaam, could you help create a notebook for training speech synthesis on tacotron 2? Facing shape mismatch errors on official nvidia notebook.

Training Dataset

Which dataset did you use to train the model? Can you please share the link of that dataset?

nice work!!

could you pls help me with some basic instruction to train this tts on my own data set. pls help. thanks