- 🎓 I study Computer Science, specializing in Intelligent Data Processing Systems
- ✨ In general I`m an enthusiastic and ambitious python developer. For now I specialise in Bot development for telegram/discord with use of LLM ang giving them personality, making them best version for chating.
- 🔑 My key to success is to never stop learning, I am intersted in improving my skill in AI development (Pytorch, TensorFlow, keras, sckit-learn), Language learning (English, Polish, Japanese). I'm mostly fluent in English, Polish, Ukranian
🧑 More about me
- 🔭 I’m currently working on [RAISE](https://github.com/AkiraRy/RAISE)-
🌱 I’m currently learning everything about AI technologies
-
👯 I’m looking to collaborate on AI researching
-
👨💻 All of my projects are available at https://github.com/AkiraRy
-
💬 Ask me about Bot develompent in Discord/Telegram
-
📫 Reach me out at *twitter: @Ak1raRy telegram: @Ak1raRy
I'm in the process of learning pytorch, tensorflow and fastapi.
Have you ever dreamed of having someone with whom you can speak whenever you like? RAISE aims to solve this. Using telegram/discord as a chating interface, whisper for transcribing your voice messages and voicevox for voice generation. We also have a Weaviate vectordb to store your chat history data. Whenever you ask for something we first use similarity search on stored documents to retrieve relevant parts. After that we prompt an LLM with all the info, giving you more related answers.
In future i plan to make a classifier, using which we can classify user query for certain tasks. Based on those tasks we can enable our llm to do a lot of things. Like searching web, telling current weather conditions, calendar appointments and so much more!
- Vector Database for memory retrieval and similarity search Weaviate
- Speech generation using VOICEVOX
- Speech transcription using Whisper
- LLM interference using llama-cpp-python
- And a lot of different features I plan to implement after rewriting the base!