Skip to content
View abi2024's full-sized avatar
0 followers · 5 following

Block or report abi2024

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.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

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

Report abuse

Pinned Loading

  1. DeepSeek-V3-Architecture-Implementation DeepSeek-V3-Architecture-Implementation Public

    Converting SmolLM2-135M to DeepSeek-V3 with MLHA and MoE

    Python

  2. UNet-Segmentation UNet-Segmentation Public

    A modular U-Net implementation in PyTorch built from scratch. Features a dynamic Model Factory to benchmark architectural variations (Pooling vs. Strided, Transpose vs. Upsample) and loss functions…

    Python

  3. smollm2-135-implementation smollm2-135-implementation Public

    Complete from-scratch implementation of SmolLM2-135M, reverse-engineered from the pretrained model.

    Jupyter Notebook

  4. FlashMoE-Serve FlashMoE-Serve Public

    High-performance MoE inference engine. Features fused OpenAI Triton kernels, continuous batching, and NF4 quantization. +72% throughput on RTX 3060.

    Python