sdan / selfextendLinks
an implementation of Self-Extend, to expand the context window via grouped attention
☆119Updated 2 years ago
Alternatives and similar repositories for selfextend
Users that are interested in selfextend are comparing it to the libraries listed below
Sorting:
- Just a bunch of benchmark logs for different LLMs☆119Updated last year
- inference code for mixtral-8x7b-32kseqlen☆105Updated 2 years ago
- An easy-to-understand framework for LLM samplers that rewind and revise generated tokens☆150Updated 3 weeks ago
- Generate Synthetic Data Using OpenAI, MistralAI or AnthropicAI☆222Updated last year
- Full finetuning of large language models without large memory requirements☆94Updated 4 months ago
- Notus is a collection of fine-tuned LLMs using SFT, DPO, SFT+DPO, and/or any other RLHF techniques, while always keeping a data-first app…☆169Updated 2 years ago
- QLoRA: Efficient Finetuning of Quantized LLMs☆79Updated last year
- Low-Rank adapter extraction for fine-tuned transformers models☆180Updated last year
- ☆165Updated 5 months ago
- ☆119Updated last year
- ☆45Updated 2 years ago
- Mixing Language Models with Self-Verification and Meta-Verification☆112Updated last year
- Experiments on speculative sampling with Llama models☆127Updated 2 years ago
- ☆137Updated last year
- Efficient vector database for hundred millions of embeddings.☆211Updated last year
- Fully fine-tune large models like Mistral, Llama-2-13B, or Qwen-14B completely for free☆232Updated last year
- Cerule - A Tiny Mighty Vision Model☆68Updated 2 months ago
- Comprehensive analysis of difference in performance of QLora, Lora, and Full Finetunes.☆83Updated 2 years ago
- Optimizing Causal LMs through GRPO with weighted reward functions and automated hyperparameter tuning using Optuna☆59Updated 3 months ago
- ☆86Updated 2 years ago
- Data preparation code for Amber 7B LLM☆94Updated last year
- Merge Transformers language models by use of gradient parameters.☆213Updated last year
- Evaluating LLMs with CommonGen-Lite☆93Updated last year
- ☆68Updated last year
- ModuleFormer is a MoE-based architecture that includes two different types of experts: stick-breaking attention heads and feedforward exp…☆226Updated 4 months ago
- Reimplementation of the task generation part from the Alpaca paper☆119Updated 2 years ago
- Simple replication of [ColBERT-v1](https://arxiv.org/abs/2004.12832).☆82Updated last year
- Small and Efficient Mathematical Reasoning LLMs☆73Updated 2 years ago
- This is our own implementation of 'Layer Selective Rank Reduction'☆240Updated last year
- Load multiple LoRA modules simultaneously and automatically switch the appropriate combination of LoRA modules to generate the best answe…☆157Updated last year