<![CDATA[Mugi_jiang的博客]]>

<![CDATA[Mugi_jiang的博客]]>https://blog.youkuaiyun.com/Mugi_jiangzh-cnhttps://blog.youkuaiyun.com/<![CDATA[国内经典多模态大模型工作2——MiniCPM-V系列（MiniCPM-Llama3-V 2.5、MiniCPM-V-2.6解读）（持续更新）]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/142857851https://blog.youkuaiyun.com/Mugi_jiang/article/details/142857851Mugi_jiangSat, 12 Oct 2024 11:02:42 +0800<![CDATA[国内经典多模态大模型工作1——Qwen-VL系列（Qwen-VL、Qwen2-VL解读）]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/142789703https://blog.youkuaiyun.com/Mugi_jiang/article/details/142789703Mugi_jiangWed, 09 Oct 2024 20:16:10 +0800<![CDATA[【MLLM阅读笔记-10】Groma精读: Localized Visual Tokenization for Grounding Multimodal Large Language Models]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/141719143https://blog.youkuaiyun.com/Mugi_jiang/article/details/141719143Mugi_jiangMon, 02 Sep 2024 16:29:33 +0800<![CDATA[【MLLM阅读笔记-11】EAGLE精读，英伟达探索MLLM的视觉编码器混合：Exploring The Design Space for Multimodal LLMs with Mixture o]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/141815285https://blog.youkuaiyun.com/Mugi_jiang/article/details/141815285Mugi_jiangMon, 02 Sep 2024 16:25:01 +0800<![CDATA[【MLLM阅读笔记-9】KOSMOS-2精读，多模态大模型目标检测：Grounding Multimodal Large Language Models to the World]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/141393793https://blog.youkuaiyun.com/Mugi_jiang/article/details/141393793Mugi_jiangFri, 30 Aug 2024 17:24:40 +0800<![CDATA[Kosmos-2的部署和测试教程（huggingface版）]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/141722910https://blog.youkuaiyun.com/Mugi_jiang/article/details/141722910Mugi_jiangFri, 30 Aug 2024 17:20:20 +0800<![CDATA[目标检测多模态大模型实践：貌似是全网唯一Shikra的部署和测试教程，内含各种踩坑以及demo代码]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/141318345https://blog.youkuaiyun.com/Mugi_jiang/article/details/141318345Mugi_jiangThu, 22 Aug 2024 16:21:13 +0800<![CDATA[【MLLM阅读笔记-8】VisionLLM v2精读： An End-to-End Generalist Multimodal Large Language Model]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/141324787https://blog.youkuaiyun.com/Mugi_jiang/article/details/141324787Mugi_jiangMon, 19 Aug 2024 18:50:46 +0800<![CDATA[【多模态大模型paper阅读笔记-7】Shikra精读，多模态大模型目标检测：Unleashing Multimodal LLM‘s Referential Dialogue Magic]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/141264581https://blog.youkuaiyun.com/Mugi_jiang/article/details/141264581Mugi_jiangMon, 19 Aug 2024 10:32:57 +0800<![CDATA[LLaVA部署报错：ImportError: cannot import name ‘LlavaLlamaForCausalLM‘ from ‘llava.model‘]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140824011https://blog.youkuaiyun.com/Mugi_jiang/article/details/140824011Mugi_jiangWed, 31 Jul 2024 15:24:12 +0800<![CDATA[（AAAI24）CLIP-Gaze: Towards General Gaze Estimation via Visual-Linguistic Model，借助视觉语言模型的泛化视线估计]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140773572https://blog.youkuaiyun.com/Mugi_jiang/article/details/140773572Mugi_jiangTue, 30 Jul 2024 11:55:28 +0800<![CDATA[LLaVA微调debug：binascii.Error: Invalid base64-encoded string: number of data characters (678697) canno]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140774166https://blog.youkuaiyun.com/Mugi_jiang/article/details/140774166Mugi_jiangMon, 29 Jul 2024 16:22:56 +0800<![CDATA[LLaVA微调debug：ImportError: xxx/flash_attn_2_cuda.cpython-310-x86_64-linux-gnu.so: undefined symbol:]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140773914https://blog.youkuaiyun.com/Mugi_jiang/article/details/140773914Mugi_jiangMon, 29 Jul 2024 16:20:33 +0800<![CDATA[为什么多模态大模型中使用Q-Former的工作变少了？附Q-Former结构简介]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140691780https://blog.youkuaiyun.com/Mugi_jiang/article/details/140691780Mugi_jiangFri, 26 Jul 2024 11:46:51 +0800<![CDATA[【多模态大模型paper阅读笔记-6】Grounding多模态，LLaVA-Grounding: Grounded Visual Chat with Large Multimodal Models]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140670544https://blog.youkuaiyun.com/Mugi_jiang/article/details/140670544Mugi_jiangFri, 26 Jul 2024 10:15:25 +0800<![CDATA[大模型微调框架swift简介]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140621155https://blog.youkuaiyun.com/Mugi_jiang/article/details/140621155Mugi_jiangWed, 24 Jul 2024 18:19:21 +0800<![CDATA[【多模态大模型paper阅读笔记-5】LLaVA：Visual Instruction Tuning，5千字精读，最适合入门多模态大模型的工作]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140664540https://blog.youkuaiyun.com/Mugi_jiang/article/details/140664540Mugi_jiangWed, 24 Jul 2024 18:13:27 +0800<![CDATA[MiniCPMV微调bug：ninja: build stopped: subcommand failed. CalledProcessError: Command ‘[‘ninja‘, ‘-v‘]]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140355030https://blog.youkuaiyun.com/Mugi_jiang/article/details/140355030Mugi_jiangThu, 11 Jul 2024 16:19:26 +0800<![CDATA[pytorch中的两种转置.T和.t()]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140300688https://blog.youkuaiyun.com/Mugi_jiang/article/details/140300688Mugi_jiangWed, 10 Jul 2024 17:43:33 +0800<![CDATA[MiniCPM-V2.5大模型微调：cusparse.h: No such file or directory...libcudart.so.11.0: cannot open shared ob]]>https://blog.youkuaiyun.com/Mugi_jiang/article/details/140320617https://blog.youkuaiyun.com/Mugi_jiang/article/details/140320617Mugi_jiangWed, 10 Jul 2024 13:57:58 +0800