IBM / ModuleFormer
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.
☆216Updated 6 months ago
Related projects ⓘ
Alternatives and complementary repositories for ModuleFormer
- Multipack distributed sampler for fast padding-free training of LLMs☆175Updated 3 months ago
- Positional Skip-wise Training for Efficient Context Window Extension of LLMs to Extremely Length (ICLR 2024)☆200Updated 5 months ago
- The official repo for "LLoCo: Learning Long Contexts Offline"☆110Updated 4 months ago
- Code repository for the c-BTM paper☆105Updated last year
- TART: A plug-and-play Transformer module for task-agnostic reasoning☆190Updated last year
- Archon provides a modular framework for combining different inference-time techniques and LMs with just a JSON config file.☆119Updated 2 weeks ago
- A simple unified framework for evaluating LLMs☆138Updated this week
- Manage scalable open LLM inference endpoints in Slurm clusters☆237Updated 3 months ago
- Benchmarking LLMs with Challenging Tasks from Real Users☆194Updated this week
- Official repository for "Scaling Retrieval-Based Langauge Models with a Trillion-Token Datastore".☆128Updated this week
- Experiments on speculative sampling with Llama models☆117Updated last year
- LongEmbed: Extending Embedding Models for Long Context Retrieval (EMNLP 2024)☆114Updated 6 months ago
- Repo for Rho-1: Token-level Data Selection & Selective Pretraining of LLMs.☆302Updated 6 months ago
- Parameter-Efficient Sparsity Crafting From Dense to Mixture-of-Experts for Instruction Tuning on General Tasks☆129Updated last month
- ☆121Updated 9 months ago
- Learning to Compress Prompts with Gist Tokens - https://arxiv.org/abs/2304.08467☆264Updated last year
- ☆111Updated last month
- batched loras☆336Updated last year
- Just a bunch of benchmark logs for different LLMs☆113Updated 3 months ago
- This is the repo for the paper Shepherd -- A Critic for Language Model Generation☆211Updated last year
- ☆182Updated 3 weeks ago
- Scaling Data-Constrained Language Models☆321Updated last month
- Official implementation for 'Extending LLMs’ Context Window with 100 Samples'☆73Updated 9 months ago
- Self-Alignment with Principle-Following Reward Models☆148Updated 8 months ago
- DSIR large-scale data selection framework for language model training☆227Updated 7 months ago
- Expert Specialized Fine-Tuning☆143Updated last month
- LOFT: A 1 Million+ Token Long-Context Benchmark☆138Updated last week
- BABILong is a benchmark for LLM evaluation using the needle-in-a-haystack approach.☆150Updated 2 months ago
- ☆294Updated 5 months ago
- The official evaluation suite and dynamic data release for MixEval.☆222Updated last week