Brian Roach's Projects
General Purpose Audio Effect Removal
Fine tune an LLM using Replicate
ChatGPT-powered QA Github Repository
Contains assets of my personal resume in jsonresume format.
Open Source Free ATS Tool to compare Resumes with Job Descriptions and create a score to rank them.
Voice data <= 10 mins can also be used to train a good VC model!
Chrome extension/ M4L node.js server to simplify the process of downloading samples/ loading into Ableton
Tools to train a generative model on arbitrary audio samples
VSCode extension for Sardine
A python package to extract information from MIDI files
Scribbly is a web application built with Next.js and Prisma that allows users to create and manage their digital journal. With a user-friendly interface powered by Radix UI and Shadcn/UI components, Scribbly makes journaling a breeze. It includes features such as customizable reminders, secure user authentication with Clerk, and a subscription plan
Real-time Neural Timbre Transfer
Create presets for Scyclone: a Real-time Neural Timbre Transfer Plug-in
Foundational Models for State-of-the-Art Speech and Text Translation
808-style MIDI sequencer built around a Teensy LC board
Music segmentation algorithm, based on SF (structural feature)
A Next.js 13 template for building apps with Radix UI and Tailwind CSS.
Beautifully designed components built with Radix UI and Tailwind CSS.
Music theory multitool with a jazz focus
Learn the Shopify + Next.js + Tailwind CSS Stack! SWR, Hydrogen, + more
Midi event transformer for music generation
A wrapper of LLMs that biases its behaviour using prompts and contexts in a transparent manner to the end-users
with 100k context windows on the way, it's now feasible for every dev to have their own smol developer
so-vits-svc fork with realtime support, improved interface and more features.
Create digital souls
songGPT is an experimental open-source project that explores the potential of Language Models, specifically ChatGPT, in generating original and customizable musical compositions. Using Python, the output is automatically converted to MIDI and audio files, enabling users to create unique and diverse music pieces.
Data extraction from documents with ML
Deezer source separation library including pretrained models.