alexjc / nanogpt-speedrunLinks
NanoGPT (124M) in 5 minutes
☆11Updated 4 months ago
Alternatives and similar repositories for nanogpt-speedrun
Users that are interested in nanogpt-speedrun are comparing it to the libraries listed below
Sorting:
- An open source replication of the stawberry method that leverages Monte Carlo Search with PPO and or DPO☆29Updated this week
- Simple GRPO scripts and configurations.☆58Updated 4 months ago
- Optimizing Causal LMs through GRPO with weighted reward functions and automated hyperparameter tuning using Optuna☆53Updated 4 months ago
- Latent Large Language Models☆18Updated 10 months ago
- Train a SmolLM-style llm on fineweb-edu in JAX/Flax with an assortment of optimizers.☆17Updated 3 months ago
- Collection of autoregressive model implementation☆85Updated 2 months ago
- GoldFinch and other hybrid transformer components☆45Updated 11 months ago
- Demonstration that finetuning RoPE model on larger sequences than the pre-trained model adapts the model context limit☆63Updated 2 years ago
- Modeling code for a BitNet b1.58 Llama-style model.☆25Updated last year
- implementation of https://arxiv.org/pdf/2312.09299☆21Updated 11 months ago
- Transformer with Mu-Parameterization, implemented in Jax/Flax. Supports FSDP on TPU pods.☆30Updated 3 weeks ago
- Simple repository for training small reasoning models☆33Updated 4 months ago
- Training hybrid models for dummies.☆23Updated 5 months ago
- ☆47Updated 4 months ago
- ☆27Updated 11 months ago
- entropix style sampling + GUI☆26Updated 7 months ago
- Fast approximate inference on a single GPU with sparsity aware offloading☆38Updated last year
- My Implementation of Q-Sparse: All Large Language Models can be Fully Sparsely-Activated☆32Updated 10 months ago
- A public implementation of the ReLoRA pretraining method, built on Lightning-AI's Pytorch Lightning suite.☆33Updated last year
- A fast, local, and secure approach for training LLMs for coding tasks using GRPO with WebAssembly and interpreter feedback.☆30Updated 2 months ago
- ☆36Updated last month
- an open source reproduction of NVIDIA's nGPT (Normalized Transformer with Representation Learning on the Hypersphere)☆101Updated 3 months ago
- Explorations into the proposal from the paper "Grokfast, Accelerated Grokking by Amplifying Slow Gradients"☆101Updated 6 months ago
- ☆49Updated last year
- A byte-level decoder architecture that matches the performance of tokenized Transformers.☆63Updated last year
- NanoGPT-speedrunning for the poor T4 enjoyers☆66Updated 2 months ago
- look how they massacred my boy☆63Updated 8 months ago
- ☆63Updated 9 months ago
- NanoGPT (124M) quality in 2.67B tokens☆28Updated last month
- https://x.com/BlinkDL_AI/status/1884768989743882276☆28Updated last month