{"id":13754217,"url":"https://github.com/lightyear-turing/TuringMM-34B-Chat","last_synced_at":"2025-05-09T22:31:29.756Z","repository":{"id":220344020,"uuid":"748033621","full_name":"lightyear-turing/TuringMM-34B-Chat","owner":"lightyear-turing","description":null,"archived":false,"fork":false,"pushed_at":"2024-02-28T09:43:59.000Z","size":561,"stargazers_count":9,"open_issues_count":2,"forks_count":1,"subscribers_count":1,"default_branch":"main","last_synced_at":"2024-11-16T07:33:11.962Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/lightyear-turing.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-01-25T06:05:41.000Z","updated_at":"2024-07-04T09:15:58.000Z","dependencies_parsed_at":"2024-02-04T12:31:03.538Z","dependency_job_id":"3c2e86cc-43b2-4ebc-b63f-0e50d5635a03","html_url":"https://github.com/lightyear-turing/TuringMM-34B-Chat","commit_stats":null,"previous_names":["lightyear-turing/turingmm-34b-chat"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/lightyear-turing%2FTuringMM-34B-Chat","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/lightyear-turing%2FTuringMM-34B-Chat/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/lightyear-turing%2FTuringMM-34B-Chat/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/lightyear-turing%2FTuringMM-34B-Chat/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/lightyear-turing","download_url":"https://codeload.github.com/lightyear-turing/TuringMM-34B-Chat/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":253335684,"owners_count":21892713,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-08-03T09:01:50.354Z","updated_at":"2025-05-09T22:31:29.227Z","avatar_url":"https://github.com/lightyear-turing.png","language":"Python","funding_links":[],"categories":["A01_文本生成_文本对话"],"sub_categories":["大语言对话模型及数据"],"readme":"\u003c!-- ![](images/turing.png) --\u003e\n\n\u003cdiv align=center\u003e\u003cimg src=\"images/turing.png\"/\u003e\u003c/div\u003e\n\n\u003cdiv align=\"center\"\u003e\n    \u003ch1\u003e\n        TuringMM 智娃AI大模型\n    \u003c/h1\u003e\n\u003c/div\u003e\n\n\n\u003cdiv align=\"center\"\u003e\n    \u003ca href=\"https://github.com/HIT-SCIR/Chinese-Mixtral-8x7B/pulls\"\u003e\n        \u003cimage src=\"https://img.shields.io/badge/License-Apache_2.0-green.svg\"\u003e\u003c/image\u003e\n    \u003c/a\u003e\n\u003c/div\u003e\n\n\n\n# 模型介绍\n\n- TuringMM-34B-Chat是一款开源的中英文Chat模型，由北京光年无限科技有限公司基于Yi-34B开源模型、基于14w的精标教育数据进行sft微调以及15W对齐数据进行DPO偏好学习得到的一个微调模型。\n\n- Yi 系列模型以双语语言模型为目标，在 3T 多语言语料库上进行训练，成为全球最强的 LLM 模型之一，在语言理解、常识推理、阅读理解等方面表现出良好的前景；\n\n\u003e 请注意，TuringMM-34B-Chat仍然可能生成包含事实性错误的误导性回复或包含偏见/歧视的有害内容，请谨慎鉴别和使用生成的内容。\n\n\u003cbr\u003e\n\n## 📥 模型下载\n\n\n| 模型名称  | 模型大小  | HuggingFace  | ModelScope  |\n|:--------:|:-------:|:--------------:|:---------------:|\n|     TuringMM-34B-Chat    | 65G  |     [🤗HuggingFace](https://huggingface.co/lightyear-turing/TuringMM-34B-Chat)     |    [🤖ModelScope](https://modelscope.cn/models/lightyearturing/TuringMM-34B-Chat/summary)   |\n\n\n\u003cbr\u003e\n\n## 模型评估结果\n\n### 模型基准测试C-EVAL\n- C-Eval 是一个全面的中文基础模型评估套件。它包含了13948个多项选择题，涵盖了52个不同的学科和四个难度级别\n\n\n|                       |  发布公司 | 平均得分 | hard | steam | 社会科学 | 人文科学 | 其他 |\n|-----------------------|----------|--------|------|-------|---------|---------|-----|\n| **Qwen-72B-Chat**     |  阿里云 | 79.5   | -    | 74.5  | 89.1     | 81.2   | 78.1 |\n| **XVERSE-65B**        |  元象智能 | 72.4   | 50.8 | 65.7  | 85       | 74     | 71.8 |\t\n| **XuanYuan-70B**      |  度小满   | 71.9   | 53.6 | 67.7  | 83.3     | 73.9   | 67.4 |\n| **TuringMM-34B-Chat** |  光年无限 | **80.7** | **60.2** | 73.8 | **89.3** | **82.3** | **83.7** |\n\n\u003cbr\u003e\n\n# 模型推理\n\n推理所需的模型权重、源码、配置已发布在 Hugging Face，下载链接见本文档最开始的表格。我们在此示范多种推理方式。程序会自动从Hugging Face 下载所需资源。\n\n## Python 代码方式\n\n```python\nimport torch\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\nfrom transformers.generation.utils import GenerationConfig\n\ntokenizer = AutoTokenizer.from_pretrained(\"lightyear-turing/TuringMM-34B-Chat\", use_fast=False, trust_remote_code=True)\nmodel = AutoModelForCausalLM.from_pretrained(\"lightyear-turing/TuringMM-34B-Chat\", device_map=\"auto\",\n                                             torch_dtype=torch.bfloat16, trust_remote_code=True)\n\nmodel.generation_config = GenerationConfig.from_pretrained(\"lightyear-turing/TuringMM-34B-Chat\")\n\n# 流式输入示例\nfor res in model.chat_stream(tokenizer, '你好啊！', max_token=200, history=[], do_sample=True):\n    print(i)\n\n# 非流式输出示例\npret, history = model.chat(tokenizer, '你好啊！', max_token=200, history=[], do_sample=True)\nprint(pret)\n```\n\n- 可以使用添加环境变量的方式控制显卡使用 `export CUDA_VISIBLE_DEVICES=0,1`（使用了0、1号显卡）\n- 或者在代码添加 os.environ['CUDA_VISIBLE_DEVICES'] = \"0,1\"\n\n## 页面方式体验\n\n```shell\npython run_chat_web.py --checkpoint_path '/your-model-path'\n```\n\u003cimg src=\"images/web_demo.png\"/\u003e\n\n本页面基于gradio工具开发，是为 Chat 对话场景设计。\n\n\n\u003cbr\u003e\n\n# 示例输出\n\n## 游戏互动\n\n`````                                                                                           \n用户：\t你好啊\nTuringMM\t你好！我是一名机器人，名叫智娃，很高兴见到你。有什么我可以帮助你的吗？\n\n用户：\t你可以和我一起玩吗\nTuringMM\t当然可以！我们可以一起玩游戏，聊天，了解更多关于地球的事情。你想玩什么游戏呢？\n\n用户：\t我们可以一起玩成语接龙吗\nTuringMM\t成语接龙？听起来很有趣！那我们一起开始吧，成语接龙第一字是“羊”，你能接上吗？\n\n用户：\t没想到，你换一个吧\nTuringMM\t没问题，成语接龙第二字是“牛”，这个简单多了吧？那么，你接什么呢？\n`````\n\n## 闲聊\n\n`````                                                                                                          \n用户：\t我不想去学校了\nTuringMM\t看起来你今天心情不太好，是吗？你为什么不想去学校呢？\n\n用户：\t学校有人欺负我\nTuringMM\t我听到你的声音里有些难过和不安，是吗？你感觉很不舒服，是吗？可以告诉我发生了什么事情吗？\n\n用户：\t他们抢我零食\nTuringMM\t这听起来很令人沮丧。你知道，这并不是第一次发生，对吧？你有没有告诉老师或家长呢？他们能帮助你。如果你不想要，他们也不会强迫你分享。你愿意试试跟他们说说话吗？                           \n\n`````\n\u003cbr\u003e\n\n# 企业介绍\n\n图灵机器人，全名北京光年无限科技有限公司，是国内强大的AI交互技术和智能OS公司，使命愿景是“实现人机自由对话、让AI机器人进入亿万家庭”。 目前，北京光年无限专注于消费类/智能学习硬件场景，并自主研发了一系列基于儿童教育智能硬件的AI技术产品，包括儿童语音对话、视觉AI交互、儿童教育AI大模型等。北京光年无限以AI技术研发驱动 ，AI领域的技术发明专利超过500件，主要产品如图灵机器人、Turing OS等在行业有良好的客户口碑。其中，主营业务产品“图灵机器人”累计超过百万级的AI对话机器人开发者，在智慧家庭/教育硬件等AI机器人场景拥有深度积累。\n\n\u003cbr\u003e\n\n# 声明\n\n由于模型参数量限制、训练数据的清洗程度等因素，本项目开源的模型可能存在以下局限性：\n- 对于事实性知识，容易产生错误的回复。\n- 由于未经过无害化微调，可能会产生歧视、危害、违背伦理道德的言论。\n- 在代码、推理上的能力仍有欠缺。\n\n基于以上模型的局限性，我们要求本项目的代码、数据、模型不得用于对社会造成危害的用途，且应当遵循基座模型的商业许可。\n\n---\nlicense: Apache License 2.0\n---","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Flightyear-turing%2FTuringMM-34B-Chat","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Flightyear-turing%2FTuringMM-34B-Chat","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Flightyear-turing%2FTuringMM-34B-Chat/lists"}