Skip to content
View ynyetname's full-sized avatar
🌴
On vacation
🌴
On vacation
6 followers · 18 following

Highlights

  • Pro

Block or report ynyetname

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
ynyetname/README.md

😎 About Me:

🎯 I'm currently working on

Reproducing and extending research papers in DPO/RLHF and RAG — currently a two-stage SFT→DPO pipeline for humor generation, and a GitHub-repo explainer with RAGAS evaluation.

🀝 I'm looking to collaborate on

Open-source LLM fine-tuning and evaluation projects β€” especially anything involving preference optimization, retrieval-augmented generation, or reproducing ML papers on constrained hardware.

πŸ’¬ Ask me about

DPO/LoRA fine-tuning on consumer GPUs, reproducing research papers, or anything RAG-related.

⚑ Fun fact

I once found that a published paper's own recommended hyperparameters break the fix it proposes β€” a 6GB laptop GPU was enough to catch it.

🌐 Socials:

Instagram LinkedIn Reddit X

πŸ’» Tech Stack:

C++ JavaScript LaTeX Markdown Python AWS Azure nVIDIA FastAPI JWT OpenCV Streamlit TailwindCSS Apache Apache Airflow MongoDB MySQL MicrosoftSQLServer Postgres Keras Matplotlib mlflow NumPy Pandas Plotly PyTorch scikit-learn Scipy TensorFlow GitHub Actions Git GitHub Docker Kubernetes Postman

πŸ“Š GitHub Stats:



πŸ† GitHub Trophies

πŸ” Top Contributed Repo

Pinned Loading

  1. qweneq2latx qweneq2latx Public

    Fine-tuned Qwen2-VL-7B with 4-bit LoRA to turn equation images into LaTeX, training the whole model on a free Colab T4 with Unsloth and TRL.

    Jupyter Notebook 1

  2. robonomous robonomous Public

    Real-time multi-object tracking with YOLOv8 detection and BoT-SORT + CLIP ReID β€” evaluated on MOT17 (43.3% HOTA).

    Python 1

  3. dataforge dataforge Public

    Built a modular Python pipeline for document understanding, from entity extraction and knowledge graph construction to embedding-based retrieval for RAG applications.

    Python

  4. reddit_mcp_server reddit_mcp_server Public

    Built an MCP server that gives LLM agents access to Reddit posts, comments, and subreddit search through PRAW.

    Python

  5. humordpo humordpo Public

    Built and validated a DPO+LoRA pipeline for humor-caption ranking on just 6GB VRAM, and found a Ξ² trade-off and judge-validation gap that the original paper overlooked.

    Python 1

  6. codebarag codebarag Public