Pinned Loading

1. Tinydeepseek Tinydeepseek Public

   26M parameter language model with Mixture of Experts architecture, Multi-Latent Attention and efficient expert routing. Only activates ~13M parameters per token.

   Jupyter Notebook 1

2. qwen-depth-distill qwen-depth-distill Public

   Prune Qwen2.5-0.5B from 24 to 7 layers, distill it back on 800M tokens, and test what the loss curve can and can't tell you.

   Python

3. LyricGenAI LyricGenAI Public

   lyrics GPT

   Python 1

4. agentic_screening agentic_screening Public

   Hybrid resume screening (vector + cross-encoder + LLM, RRF-fused) with live voice interviews

   Python

5. GPT-Speculative-Inference GPT-Speculative-Inference Public

   Jupyter Notebook 1

6. simple_GRPO simple_GRPO Public

   Forked from lsdefine/simple_GRPO

   A very simple GRPO implement for reproducing r1-like LLM thinking.

   Python