heyblackC / BetterMixture-Top1-SolutionLinks
天池算法比赛《BetterMixture - 大模型数据混合挑战赛》的第一名top1解决方案
☆33Updated last year
Alternatives and similar repositories for BetterMixture-Top1-Solution
Users that are interested in BetterMixture-Top1-Solution are comparing it to the libraries listed below
Sorting:
- 大语言模型应用:RAG、NL2SQL、聊天机器人、预训练、MOE混合专家模型、微调训练、强化学习、天池数据竞赛☆74Updated 11 months ago
- ☆119Updated last year
- 通用简单工具项目☆22Updated last year
- 2023全球智能汽车AI挑战赛——赛道一:AI大模型检索问答, 75+ baseline☆60Updated 2 years ago
- 怎么训练一个LLM分词器☆154Updated 2 years ago
- 使用 Qwen2ForSequenceClassification 简单实现文本分类任务。☆89Updated last year
- 本项目用于大模型数学解题能力方面的数据集合成,模型训练及评测,相关文章记录。☆99Updated last year
- 使用单个24G显卡,从0开始训练LLM☆56Updated 6 months ago
- 集成Qwen与DeepSeek等先进大语言模型,支持纯LLM+分类层模式及LLM+LoRA+分类层模式,使用transformers模块化设计和训练便于根据需要调整或替换组件。☆17Updated 4 months ago
- ☆13Updated 9 months ago
- Qwen1.5-SFT(阿里, Ali), Qwen_Qwen1.5-2B-Chat/Qwen_Qwen1.5-7B-Chat微 调(transformers)/LORA(peft)/推理☆69Updated last year
- 阿里天池: 2023全球智能汽车AI挑战赛——赛道一:AI大模型检索问答 baseline 80+☆119Updated 2 years ago
- 基于DPO算法微调语言大模型,简单好上手。☆48Updated last year
- 通义千问的DPO训练☆61Updated last year
- ☆81Updated 5 months ago
- LLM+RAG for QA☆22Updated last year
- 1st Solution For Conversational Multi-Doc QA Workshop & International Challenge @ WSDM'24 - Xiaohongshu.Inc☆162Updated 5 months ago
- 专注于中文领域大语言模型,落地到某个行业某个领域,成为一个行业大模型、公司级别或行业级别领域大模型。☆126Updated 10 months ago
- ☆115Updated last year
- Clustering and Ranking: Diversity-preserved Instruction Selection through Expert-aligned Quality Estimation☆90Updated last year
- AFAC2024金融智能创新大赛☆64Updated last year
- deepspeed+trainer简单高效实现多卡微调大模型☆130Updated 2 years ago
- 大语言模型指令调优工具(支持 FlashAttention)☆178Updated 2 years ago
- 大语言模型训练和服务调研☆37Updated 2 years ago
- kaggle 2024 Eedi 第10名 金牌方案☆44Updated last year
- 用于AIOPS24挑战赛的Demo☆64Updated last year
- 一套代码指令微调大模型☆38Updated 2 years ago
- ☆28Updated last year
- 在verl上做reward的定制开发☆140Updated 7 months ago
- Code for a New Loss for Mitigating the Bias of Learning Difficulties in Generative Language Models☆66Updated 10 months ago