Pinned Loading

1. wordle_rl wordle_rl Public

   Fork of ByteDance's verl to train a LLM to play Wordle through finetuning and reinforcement learning.

   Python 3

2. AI4DrugDesign AI4DrugDesign Public

   Automated drug discovery pipeline that takes a target protein and finds relevant compounds using ChEMBL bioactivity search, de novo molecule generation, AutoDock Vina docking, and ADME filtering.

   Python

3. GPT2_Replication GPT2_Replication Public

   Recreation of GPT-2 from scratch with PyTorch

   Jupyter Notebook

4. SketchVLM_Demo SketchVLM_Demo Public

   A training-free, model-agnostic framework that enables VLMs to produce non-destructive, editable SVG overlays on the input image to visually explain their answers.

   TypeScript 3 1