SuperAPI AI Directory
Partner with us中文AI newsFavorites

All AI tools

35,707 results · Page 950 of 992

LLM-eval-survey

AI developer platforms

LLM-eval-survey is the official GitHub repository of the large language model evaluation survey paper "A Survey on Evaluation of Large Language Models," used to compile and organize papers and resources related to LLM evaluation. The repository organizes papers according to the chapter structure of the survey, covering dimensions such as What to evaluate and Where to evaluate, and continuously collects evaluation literature in directions such as natural language processing, robustness, ethics, bias and trustworthiness, social sciences, natural sciences and engineering, medical applications, and agent applications. The repository also provides links to related projects, such as PromptBench and LLM-eval for large language model robustness evaluation.

Read more

llm_multiagent_debate

AI agents

llm_multiagent_debate is a codebase for a multi-agent debate framework that implements the paper "Improving Factuality and Reasoning in Language Models through Multiagent Debate," which was accepted at ICML 2024. Its core idea is to have multiple language model instances communicate in a debate format, thereby improving the factual accuracy and reasoning ability of responses. The repository provides experimental code for four types of tasks: arithmetic, GSM (grade school math), biography, and MMLU, including generation and evaluation scripts, making it convenient for researchers to reproduce the multi-agent debate experimental results from the paper.

Read more

LLMsBook

AI developer platforms

LLMsBook is a practical code summary repository for the book "Large Language Model Practical Guide: Application Practice and Scenario Implementation," hosted on GitHub. The project organizes code by book chapters, covering tokenizer training, PEFT-based Llama fine-tuning, DPO-based preference alignment, DeepSeek Coder-based SQL fine-tuning, Baichuan-based role-playing model fine-tuning, Qwen-based dialogue element extraction model fine-tuning, BGE-based vector fine-tuning, ChatGLM3-based knowledge base answer generation task fine-tuning and Streamlit application building, as well as practical content such as building AutoGPT based on Langchain. The repository also records book errata and welcomes feedback via issues.

Read more

lmm-r1

AI developer platforms

LMM-R1 is a fork project of OpenRLHF, aimed at providing reinforcement learning training infrastructure for large multimodal models (LMMs) to enhance their multimodal reasoning capabilities. The project proposes a two-stage rule-based reinforcement learning framework: the first stage is Foundational Reasoning Enhancement (FRE), which uses pure text data to establish a strong reasoning foundation; the second stage is Multimodal Generalization Training (MGT), which extends reasoning capabilities to the multimodal domain. Through this framework, LMM-R1 can significantly improve the reasoning performance of 3B-scale multimodal models using limited training data. The project supports training algorithms such as PPO, REINFORCE++, and RLOO, supports models such as Qwen2.5-VL, Phi3.5-V, and Phi4-Multimodal, and provides complete training scripts and dataset format specifications.

Read more