Popular GitHub Repositories for Prefix

Connectionist Temporal Classification (CTC) decoding algorithms: best path, beam search, lexicon search, prefix search, and token passing. Implemented in Python.

[NAACL 2022 Findings] Good Visual Guidance Makes A Better Extractor: Hierarchical Visual Prefix for Multimodal Entity and Relation Extraction

[AAAI 2024] Investigating the Effectiveness of Task-Agnostic Prefix Prompt for Instruction Following

An efficient implementation of RNN-T Prefix Beam Search in C++/CUDA.

Supplementary material for "Understanding Parameter-Efficient Finetuning of Large Language Models: From Prefix Tuning to Adapters"

Accelerate LLM preference tuning via prefix sharing with a single line of code

SMASHED is a toolkit designed to apply transformations to samples in datasets, such as fields extraction, tokenization, prompting, batching, and more. Supports datasets from Huggingface, torchdata iterables, or simple lists of dictionaries.

[EMNLP 2023 (Findings)] Schema-adaptable Knowledge Graph Construction

The official repo of paper "Self-Control of LLM Behaviors by Compressing Suffix Gradient into Prefix Controller"

用于论文《Marconi：混合大型语言模型时代的词缀缓存》[MLSys '25] 的制品

Example code for prefix-tuning GPT/GPT-NeoX models and for inference with trained prefixes

[ICON 2020] TensorFlow Code for "End-to-End Automatic Speech Recognition System for Gujarati"

Repository for our paper "Frustratingly Easy Jailbreak of Large Language Models via Output Prefix Attacks". https://www.researchsquare.com/article/rs-4385503/latest

?Code for paper Parallelizable Complex Neural Dynamics Models for Temperature Estimation with Hardware Acceleration.

Single-pass Parallel Prefix Scan with Decoupled Look-back in Mojo

Comparison of different adaptation methods on PEFT for fine-tuning downstream tasks or benchmarks.

File Organizer Tool – Cross-platform app that organizes files into folders by filename prefixes. Offers GUI & CLI, 10-language support, prefix removal, undo option, and real-time preview. Built with Python & Tkinter using only the standard library.

docker-compose for vllm, support sticky-sessions using traefik to enable the prefix-caching feature of vllm

Comparing QLoRA, Prompt & Prefix Tuning on Mistral-7B for medical instruction-following

Implementation of the following parameter-efficient fine-tuning methods on GPT-2 for summarization: Soft prompt tuning by optimizing prefix embeddings, LoRA, and fine-tuning only the last classifier layer while keeping the rest of the model frozen