{"id":19529839,"url":"https://github.com/internlm/internlm-wqx","last_synced_at":"2025-10-27T09:09:03.228Z","repository":{"id":245082083,"uuid":"810648257","full_name":"InternLM/InternLM-WQX","owner":"InternLM","description":null,"archived":false,"fork":false,"pushed_at":"2024-07-05T12:22:46.000Z","size":26,"stargazers_count":19,"open_issues_count":0,"forks_count":2,"subscribers_count":3,"default_branch":"main","last_synced_at":"2025-04-04T12:21:53.382Z","etag":null,"topics":["internlm-20b","internlm2","llm"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/InternLM.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"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":"2024-06-05T05:27:35.000Z","updated_at":"2025-02-11T16:05:49.000Z","dependencies_parsed_at":"2024-11-11T01:29:48.972Z","dependency_job_id":"3af93ceb-7b0f-4185-963b-38579ced1cdd","html_url":"https://github.com/InternLM/InternLM-WQX","commit_stats":null,"previous_names":["internlm/internlm-wqx"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/InternLM%2FInternLM-WQX","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/InternLM%2FInternLM-WQX/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/InternLM%2FInternLM-WQX/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/InternLM%2FInternLM-WQX/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/InternLM","download_url":"https://codeload.github.com/InternLM/InternLM-WQX/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":250986162,"owners_count":21518452,"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":["internlm-20b","internlm2","llm"],"created_at":"2024-11-11T01:27:50.736Z","updated_at":"2025-10-27T09:08:58.197Z","avatar_url":"https://github.com/InternLM.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cdiv align=\"center\"\u003e\n\n\u003cimg src=\"https://raw.githubusercontent.com/InternLM/InternLM/main/assets/logo.svg\" width=\"200\"/\u003e\n  \u003cdiv\u003e \u003c/div\u003e\n  \u003cdiv align=\"center\"\u003e\n    \u003cb\u003e\u003cfont size=\"5\"\u003eInternLM2-WQX\u003c/font\u003e\u003c/b\u003e\n    \u003csup\u003e\n      \u003ca href=\"https://internlm.intern-ai.org.cn/\"\u003e\n        \u003ci\u003e\u003cfont size=\"4\"\u003eHOT\u003c/font\u003e\u003c/i\u003e\n      \u003c/a\u003e\n    \u003c/sup\u003e\n    \u003cdiv\u003e \u003c/div\u003e\n  \u003c/div\u003e\n\n[![license](https://raw.githubusercontent.com/InternLM/InternLM/main/assets/license.svg)](./LICENSE)\n\n\n\nInternLM2-WQX-20B \u003ca href=\"https://huggingface.co/internlm/internlm2-wqx-20b\"\u003e🤗\u003c/a\u003e \u003ca href=\"https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm2-wqx-20b/summary\"\u003e\u003cimg src=\"./assets/modelscope_logo.png\" width=\"20px\"\u003e\u003c/a\u003e ｜ InternLM2-WQX-VL-20B \u003ca href=\"https://huggingface.co/internlm/internlm2-wqx-vl-20b\"\u003e🤗\u003c/a\u003e \u003ca href=\"https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm2-wqx-vl-20b/summary\"\u003e\u003cimg src=\"./assets/modelscope_logo.png\" width=\"20px\"\u003e\u003c/a\u003e\n\u003c/div\u003e\n\n# Introduction\n\nInternLM2-WQX与InternLM2-WQX-VL是InternLM团队于2024年高考前夕最新推出的文曲星系列模型。\n\n高考覆盖各类学科及题型，同时因其开考前的“绝密性”，被视作中国最具权威的考试之一，成为评估考生综合能力的“试金石”。这一面向人类设计的高难度综合性测试，目前普遍被研究者用于考察大模型的智能水平。InternLM2-WQX系列模型在2024年高考评测集[GAOKAO-Eval](https://github.com/open-compass/GAOKAO-Eval)上取得了优异的成绩，综合表现与GPT-4o相当，且超越了国内外一系列开源大模型，体现了InternLM2-WQX系列模型优秀的性能。\n\n我们即将更新关于文曲星系列模型数据准备的相关说明，敬请期待。\n\n\n# Model Zoo\n\n\n| Model                       | HuggingFace                          | ModelScope                           | Release Date |\n| --------------------------- | ----------------------------------------- | ---------------------------------------- | ------------ |\n| **InternLM2-WQX-20B**          | [🤗internlm2-wqx-20b](https://huggingface.co/internlm/internlm2-wqx-20b) | [\u003cimg src=\"./assets/modelscope_logo.png\" width=\"20px\" /\u003e internlm2-wqx-20b](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm2-wqx-20b/summary) | 2024-06-04   |\n| **InternLM2-WQX-VL-20B**          | [🤗internlm2-wqx-vl-20b](https://huggingface.co/internlm/internlm2-wqx-vl-20b) | [\u003cimg src=\"./assets/modelscope_logo.png\" width=\"20px\" /\u003e internlm2-wqx-vl-20b](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm2-wqx-vl-20b/summary) | 2024-06-04   |\n\n\n## MD5 Check\n\n### LLM权重文件的md5值\n```\nmd5sum ./*\n5209adfd6ef7d1724848ff0372362568  ./model-00001-of-00004.safetensors\ne37ee2eafecfed543d10dca75998204e  ./model-00002-of-00004.safetensors\nea3da8035b0c2a31c369dd463adf9b52  ./model-00003-of-00004.safetensors\nf1ff218f801c69fd4c12c534b64e1b60  ./model-00004-of-00004.safetensors\n```\n\n### MLLM权重文件的md5值\n```\nmd5sum ./*\n158657dbae9bc369d67cf4bfbdfaaf71  ./pytorch_model-00001-of-00005.bin\nc21db8ac1315c10df768f6c3ae3f2825  ./pytorch_model-00002-of-00005.bin\nebc4b0b70e8e9f1adc0b728558d650fb  ./pytorch_model-00003-of-00005.bin\neaa393a66dc632d0a6f0f7d815c439bb  ./pytorch_model-00004-of-00005.bin\n7e6e3237d99a7e8bd7ca9ba10747bfdb  ./pytorch_model-00005-of-00005.bin\n\n./clip_l_560_pro7b/*\n97b05f40ee9826eda467489eed65f85c  ./clip_l_560_pro7b/pytorch_model.bin\n```\n\n# Quick Start\n\n### 快速调用**InternLM2-WQX-20B**语言模型\n\n使用transformers 后端进行推理\n\n```python\nimport torch\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\n\ndevice = \"cuda\"\n\ntokenizer = AutoTokenizer.from_pretrained(\"internlm/internlm2-wqx-20b\", trust_remote_code=True)\nmodel = AutoModelForCausalLM.from_pretrained(\n    \"internlm/internlm2-wqx-20b\",\n    torch_dtype=torch.bfloat16,\n    trust_remote_code=True\n).to(device).eval()\n\nquery = \"已知圆柱和圆锥的底面半径相等，侧面积相等，且它们的高均为$ \\\\sqrt { 3 }$，则圆锥的体积为（ ）．\\nA. $ 2 \\\\sqrt { 3 } \\\\pi$\\nB. $ 3 \\\\sqrt { 3 } \\\\pi$\\nC. $ 6 \\\\sqrt { 3 } \\\\pi$\\nD. $ 9 \\\\sqrt { 3 } \\\\pi$\"\n\ninputs = tokenizer(query, return_tensors=\"pt\")\n\ninputs = inputs[\"input_ids\"].to(device)\n\ngen_kwargs = {\"max_length\": 1024, \"do_sample\": False}\n\noutputs = model.generate(inputs, **gen_kwargs)\noutputs = outputs[0].cpu().tolist()[len(inputs[0]) :]\n\nresponse = tokenizer.decode(outputs, skip_special_tokens=True)\nprint(response)\n```\n\n使用vllm 后端进行推理：\n\n```python\nfrom vllm import LLM, SamplingParams\n\nmodel_name = \"internlm/internlm2-wqx-20b\"\nprompts = [\"已知圆柱和圆锥的底面半径相等，侧面积相等，且它们的高均为$ \\\\sqrt { 3 }$，则圆锥的体积为（ ）．\\nA. $ 2 \\\\sqrt { 3 } \\\\pi$\\nB. $ 3 \\\\sqrt { 3 } \\\\pi$\\nC. $ 6 \\\\sqrt { 3 } \\\\pi$\\nD. $ 9 \\\\sqrt { 3 } \\\\pi$\"]\nsampling_params = SamplingParams(temperature=0.0, max_tokens=1024)\n\nllm = LLM(\n    model=model_name,\n    trust_remote_code=True,\n    enforce_eager=True,\n)\n\noutputs = llm.generate(prompts, sampling_params)\n\nfor output in outputs:\n    prompt = output.prompt\n    generated_text = output.outputs[0].text\n    print(f\"Prompt: {prompt!r}, \\nGenerated text: {generated_text!r}\")\n```\n\n### **InternLM2-WQX-20B**语言模型的 Web UI\n\n使用transformers后端进行推理：\n\n```\npython web_ui_wqx.py -m internlm/internlm2-wqx-20b\n```\n\n### 快速调用**InternLM2-WQX-VL-20B**视觉语言模型\n\n使用transformers后端进行推理:\n\n```python\nfrom PIL import Image\nfrom io import BytesIO\nimport requests\nfrom transformers import AutoModel, AutoTokenizer, AutoModelForCausalLM\nimport torch\nfrom infer_wqx_vl import process_query_and_image, HD_transform\n\nmodel_path = \"internlm/internlm2-wqx-vl-20b\"\ntokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)\nmodel = AutoModel.from_pretrained(model_path, torch_dtype=torch.bfloat16, trust_remote_code=True).cuda().eval()\nmodel.cuda().half()\nmodel.tokenizer = tokenizer\n\nimage_url = \"https://ks-1302698447.cos.ap-shanghai.myqcloud.com/img/phymerge.png\"\nquery = \"体育课上两位同学在室内羽毛球场进行羽毛球比赛，羽毛球在空中上升的运动轨迹如图中虚线所示，考虑空气阻力，羽毛球加速度方向示意图可能正确的是（\\u3000\\u3000） \\nA:\u003cIMAGE 0\u003e  \\nB: \u003cIMAGE 1\u003e  \\nC:\u003cIMAGE 2\u003e  \\nD:\u003cIMAGE 3\u003e \"\n\nresponse = requests.get(image_url)\nimage = Image.open(BytesIO(response.content))\nembeds, im_mask = process_query_and_image(query, image, model, HD_transform)\n\noutputs = model.generate(inputs_embeds=embeds, im_mask=im_mask,\n                            temperature=0.0, max_new_tokens=256, num_beams=1,\n                            do_sample=False, repetition_penalty=1.0)\noutput_token = outputs[0]\noutput_text = model.tokenizer.decode(output_token, add_special_tokens=False)\nprint(output_text)\n#  \u003cs\u003e 斜向下\n# 答案是：C\u003c/s\u003e\n```\n针对这个选项里面有图片的考题，我们将图片进行了合并并标记上`\u003cIMAGE {id}\u003e`来让语言模型能理解多图考题。 当前示例展示的是已经拼接好的图片，详细的图像预处理请参考[GAOKAO-Eval](https://github.com/open-compass/GAOKAO-Eval)中的多模态处理工具。\n\n### **InternLM2-WQX-VL-20B**语言模型的 Web UI\n\n使用transformers后端进行推理：\n\n```\npython web_ui_wqx_vl.py -m internlm/internlm2-wqx-vl-20b\n```\n\n# Citation\n\n```bibtex\n@misc{2024internlm2wqx,\n    title={https://github.com/InternLM/InternLM-WQX},\n    author={InternLM Team},\n    howpublished = {\\url{https://github.com/InternLM/InternLM-WQX}},\n    year={2024}\n}\n```","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Finternlm%2Finternlm-wqx","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Finternlm%2Finternlm-wqx","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Finternlm%2Finternlm-wqx/lists"}