IBM / ModuleFormerLinks
ModuleFormer is a MoE-based architecture that includes two different types of experts: stick-breaking attention heads and feedforward experts. We released a collection of ModuleFormer-based Language Models (MoLM) ranging in scale from 4 billion to 8 billion parameters.
☆224Updated last month
Alternatives and similar repositories for ModuleFormer
Users that are interested in ModuleFormer are comparing it to the libraries listed below
Sorting:
- Code repository for the c-BTM paper☆107Updated 2 years ago
- Positional Skip-wise Training for Efficient Context Window Extension of LLMs to Extremely Length (ICLR 2024)☆204Updated last year
- TART: A plug-and-play Transformer module for task-agnostic reasoning☆200Updated 2 years ago
- Experiments on speculative sampling with Llama models☆125Updated 2 years ago
- The official repo for "LLoCo: Learning Long Contexts Offline"☆117Updated last year
- Functional Benchmarks and the Reasoning Gap☆89Updated last year
- ☆95Updated 2 years ago
- Pre-training code for Amber 7B LLM☆168Updated last year
- Multipack distributed sampler for fast padding-free training of LLMs☆201Updated last year
- Evaluating LLMs with CommonGen-Lite☆91Updated last year
- Scripts for generating synthetic finetuning data for reducing sycophancy.☆117Updated 2 years ago
- Data preparation code for Amber 7B LLM☆93Updated last year
- This is the repo for the paper Shepherd -- A Critic for Language Model Generation☆217Updated 2 years ago
- Small and Efficient Mathematical Reasoning LLMs☆72Updated last year
- ☆128Updated last year
- Mixing Language Models with Self-Verification and Meta-Verification☆109Updated 10 months ago
- Just a bunch of benchmark logs for different LLMs☆118Updated last year
- Scaling Data-Constrained Language Models☆342Updated 4 months ago
- Manage scalable open LLM inference endpoints in Slurm clusters☆273Updated last year
- Spherical Merge Pytorch/HF format Language Models with minimal feature loss.☆139Updated 2 years ago
- ☆415Updated last year
- Code for the paper "QMoE: Practical Sub-1-Bit Compression of Trillion-Parameter Models".☆277Updated last year
- ☆127Updated last year
- ☆202Updated 10 months ago
- Code for the paper "Rethinking Benchmark and Contamination for Language Models with Rephrased Samples"☆311Updated last year
- Evaluating LLMs with fewer examples☆163Updated last year
- RuLES: a benchmark for evaluating rule-following in language models☆238Updated 8 months ago
- ☆134Updated last year
- ☆149Updated last year
- Learning to Compress Prompts with Gist Tokens - https://arxiv.org/abs/2304.08467☆297Updated 8 months ago