Skip to main content

GenAI Projects

Projects on Large Language Models​

Train your own LLMs from Scratch like GPT-3.5​

  • Implement script for training and evaluating Large Language Models (LLMs) from scratch, such as GPT-3.5.
  • Employ best practices for training LLMs, including infrastructure setup.
  • Implement parallel and distributed computing strategies, including 3D parallelism and similar approaches.

Develop your own ChatGPT from scratch​

  • Implement all stages in the development of a Large Language Model (LLM) similar to ChatGPT, including:
    • Pretraining
    • Supervised Finetuning
    • Reinforcement Learning from Human Feedback (RLHF)
  • Learn and apply the best industry practices for creating a dialogue- optimized LLM like ChatGPT.

Develop your own LLM Application using Prompt Engineering​

  • Master prompt engineering, including its various techniques and methods.
  • Build a Large Language Model (LLM) application, such as a chatbot, using private data and leveraging both ChatGPT API and open- source LLMs.

Create a production-ready X on your private data​

  • Use LlamaIndex to build a production-ready Retrieval-Augmented Generation (RAG) system on your private data.
  • Acquire hands-on experience with advanced components of LlamaIndex, including:
    • Agents
    • Router Query Engine, SubQuestionQuery Engine
    • Open AI Assistants
    • Finetuning with Retrieval Augmentation (RADIT)
    • Finetuning Embeddings

Building End-to-End RAG Apps​

  • Design and develop a RAG-based QA chatbot with a user-friendly interface.
  • Build a full-stack application using LangChain and Streamlit, ensuring seamless data flow between the frontend and backend.
  • Master the deployment of RAG applications for real-world usage.

Build Conversational Apps and Agents​

  • Construct engaging conversational bots using LLMs like ChatGPT, tailoring them for specific use cases.
  • Develop sophisticated AI tools and agents by integrating LLMs with LangChain's flexible capabilities.
  • Design applications for LLM management and interaction within LangChain environments.

Advanced RAG System Development​

  • Implement reranking techniques to significantly enhance search accuracy and relevance within RAG systems.
  • Apply cutting-edge RAG methods from recent research, optimizing performance for specific domains.
  • Adapt RAG systems to handle diverse data types (text, tables, images, etc.) broadening their applicability.

Finetuning LLMs using Soft Prompting, Adaptor techniques using PEFT​

  • Use PEFT (Progressive Layer Dropping-based Efficient Finetuning) for accelerated model finetuning on a single GPU.
  • Finetune Large Language Models (LLMs) for specific downstream use cases using techniques like LoRA, QLoRA, and soft prompting.
  • Build an Instruction Following Large Language Model using PEFT.

Projects on Stable Diffusion Models​

Finetune your own Stable Diffusion models on custom dataset​

  • Gain proficiency in the development and fine-tuning of Stable Diffusion models, understanding its intricacies and applications.
  • Implement industrial practices for the creation and customization of Stable Diffusion models specific to your dataset.

Build your own personalized Text to Image models using DreamBooth​

  • Develop the skill to build personalized Text to Image models using DreamBooth, mastering the conversion of textual descriptions into distinct and tailored visual representations.
  • Implement DreamBooth on your own image dataset, enabling the creation of highly personalized and context-specific text-to-image models that reflect the nuances of your provided data.

Finetune Diffusion models using ControlNets and Instruct Pix2Pix models​

  • Engage in hands-on fine-tuning of diffusion models, particularly employing ControlNets and InstructPix2Pix models for improved performance and adaptability.
  • Integrate and optimize the use of ControlNets and InstructPix2Pix models within the stable diffusion model framework, perfecting the skill of refining and customizing these models for specific tasks and datasets.