radarfudan Goto Github PK

h3

Language Modeling with the H3 State Space Model

hasee

hippo-code

hyena

JAX/Flax implementation of the Hyena Hierarchy

in-context-learning

in-context-operator-networks

ICON for in-context operator learning

interest

Temporal re-weighting improve the long-term memory learning.

keops

KErnel OPerationS, on CPUs and GPUs, with autodiff and without memory overflows

leastsquares

CS 240H final project

lhj

Can we use lightning as data loader and jax as the models?

lightning-hydra-template

PyTorch Lightning + Hydra. A very user-friendly template for ML experimentation. ⚡🔥⚡

lit-gpt

Hackable implementation of state-of-the-art open-source LLMs based on nanoGPT. Supports flash attention, 4-bit and 8-bit quantization, LoRA and LLaMA-Adapter fine-tuning, pre-training. Apache 2.0-licensed.

lm-evaluation-harness

A framework for few-shot evaluation of autoregressive language models.

lm-rmt_bala

Recurrent Memory Transformer

lp_solver

project for Linear programming

mamba

mamba-minimal

Simple, minimal implementation of the Mamba SSM in one file of PyTorch.

mamba-minimal-jax

minitorch

new_repo

othello_mamba

Evaluating the Mamba architecture on the Othello game

picogpt

An unnecessarily tiny implementation of GPT-2 in NumPy.

profiling-cuda-in-torch

pyprobml

Python code for "Probabilistic Machine learning" book by Kevin Murphy

radarfudan

radarfudan.github.io

rwkv-cuda

The CUDA version of the RWKV language model ( https://github.com/BlinkDL/RWKV-LM )

rwkv-lm

RWKV is an RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best of RNN and transformer - great performance, fast inference, saves VRAM, fast training, "infinite" ctx_len, and free sentence embedding.

rwkv-v2-rnn-pile

RWKV-v2-RNN trained on the Pile. See https://github.com/BlinkDL/RWKV-LM for details.

s5