lucidrains / CALM-pytorchLinks
Implementation of CALM from the paper "LLM Augmented LLMs: Expanding Capabilities through Composition", out of Google Deepmind
☆178Updated last year
Alternatives and similar repositories for CALM-pytorch
Users that are interested in CALM-pytorch are comparing it to the libraries listed below
Sorting:
- Parameter-Efficient Sparsity Crafting From Dense to Mixture-of-Experts for Instruction Tuning on General Tasks (EMNLP'24)☆148Updated last year
- Scaling Data-Constrained Language Models☆342Updated 3 months ago
- ☆136Updated 10 months ago
- ☆270Updated last year
- LLM-Merging: Building LLMs Efficiently through Merging☆203Updated last year
- ☆127Updated last year
- This is the official repository for Inheritune.☆113Updated 7 months ago
- ☆192Updated 5 months ago
- Language models scale reliably with over-training and on downstream tasks☆100Updated last year
- Recurrent Memory Transformer☆150Updated 2 years ago
- Self-Alignment with Principle-Following Reward Models☆166Updated 2 weeks ago
- Implementation of the Llama architecture with RLHF + Q-learning☆168Updated 8 months ago
- ☆192Updated last year
- Implementation of 🥥 Coconut, Chain of Continuous Thought, in Pytorch☆180Updated 3 months ago
- ☆100Updated last year
- Positional Skip-wise Training for Efficient Context Window Extension of LLMs to Extremely Length (ICLR 2024)☆206Updated last year
- The official repo for "LLoCo: Learning Long Contexts Offline"☆116Updated last year
- X-LoRA: Mixture of LoRA Experts☆244Updated last year
- ☆155Updated last year
- ☆202Updated 10 months ago
- Multipack distributed sampler for fast padding-free training of LLMs☆201Updated last year
- Code to reproduce "Transformers Can Do Arithmetic with the Right Embeddings", McLeish et al (NeurIPS 2024)☆193Updated last year
- ☆196Updated last month
- Pre-training code for Amber 7B LLM☆168Updated last year
- Code for NeurIPS'24 paper 'Grokked Transformers are Implicit Reasoners: A Mechanistic Journey to the Edge of Generalization'☆230Updated 2 months ago
- ModuleFormer is a MoE-based architecture that includes two different types of experts: stick-breaking attention heads and feedforward exp…☆224Updated 2 weeks ago
- The official implementation of the paper "What Matters in Transformers? Not All Attention is Needed".☆176Updated 6 months ago
- Code for Zero-Shot Tokenizer Transfer☆137Updated 8 months ago
- Replicating O1 inference-time scaling laws☆90Updated 10 months ago
- Official repository for the paper "SwitchHead: Accelerating Transformers with Mixture-of-Experts Attention"☆99Updated last year