8 Pull requests merged by 5 people

• Add detailed description in speech mode (en version)

  #763 merged Jan 20, 2025

• update news

  #741 merged Jan 17, 2025

• docs: update README.md

  #735 merged Jan 16, 2025

• add model init in Multimodal Live Streaming code

  #733 merged Jan 16, 2025

• add requirements for webdemo

  #731 merged Jan 16, 2025

• remove unused doc

  #729 merged Jan 16, 2025

• Best Practice with LLaMA-Factory

  #711 merged Jan 16, 2025

• Update requirements.txt

  #713 merged Jan 15, 2025

3 Pull requests opened by 3 people

• Add Login Functionality

  #751 opened Jan 18, 2025

• Update audio part of Llamafactory docs

  #756 opened Jan 19, 2025

• Modify eval_mm for MiniCPM-o 2.6

  #771 opened Jan 21, 2025

43 Issues closed by 13 people

• [BUG] The first 3 online demo links aren't working, I speak English but it responds in Chinese.

  #757 closed Jan 21, 2025

• [BUG] AttributeError: 'MiniCPMOTokenizerFast' object has no attribute 'image_processor'

  #767 closed Jan 21, 2025

• minicpm-V2.6提取图片中的文本信息怎么输出置信度呢

  #686 closed Jan 21, 2025

• [web] <title>根据Readme启动demo后，demo无法在浏览器中调起麦克风和摄像头

  #716 closed Jan 20, 2025

• 💡 [REQUEST] - iPad Application Demo

  #759 closed Jan 20, 2025

• vllm api server-请问如何实现vllm api server推理时的图文穿插

  #758 closed Jan 20, 2025

• [BUG] MiniCPM-2.6o Model Exhibits Echo Behavior When Evaluated with VLMKit

  #755 closed Jan 19, 2025

• [ollama] - 请问这个数据有误吗？

  #691 closed Jan 18, 2025

• [BUG] <title>TypeError: CPMTrainer.training_step() takes 3 positional arguments but 4 were given

  #698 closed Jan 18, 2025

• [BUG] <title>llama minicpmv-cli 有内存泄露问题？

  #703 closed Jan 18, 2025

• [BUG] <title> errored out

  #611 closed Jan 18, 2025

• 是否能通过多GPU实现推理加速？

  #610 closed Jan 18, 2025

• 关于MiniCPM-V-2转gguf

  #598 closed Jan 18, 2025

• [BUG] <title> 提交一个训练代码逻辑错误，代码位于MiniCPM-V/finetune/dataset.py 151行

  #631 closed Jan 18, 2025

• [BUG] <title>NotImplementedError: Cannot copy out of meta tensor; no data! Please use torch.nn.Module.to_empty() instead of torch.nn.Module.to() when moving module from meta to a different device.

  #635 closed Jan 18, 2025

• [vllm] - 一张A10，vllm api方式启动，报显存不足

  #640 closed Jan 18, 2025

• Highlighting detected features in an Image with a boundary box

  #641 closed Jan 18, 2025

• [BUG] <title> Inference error. Replacing the LLM part with Llama-3.1 70B quantized causing error ( RuntimeError: shape mismatch: value tensor of shape [1024] cannot be broadcast to indexing result of shape [1025] )

  #643 closed Jan 18, 2025

• [BUG] <title> 请问在用deepspeed zero3 训练的过程因为minicpm navit的逻辑会导致不同rank上的image feature size 不同这样会hang住，这个是怎么解决的？zero2是没有问题的

  #644 closed Jan 18, 2025

• [BUG] <title> MiniCPM-V 2.6 在mathvista-minitest中测试性能远不如report

  #647 closed Jan 18, 2025

• [llamacpp] - Is it possible to provide the ability to perform video inference by the server mode？

  #485 closed Jan 18, 2025

• [BUG] <title> 按照官方教程编译的ollama，在run myMinicpmv2.6:latest时候，模型直接报错退出

  #482 closed Jan 18, 2025

• [BUG] <title>多卡启动api

  #476 closed Jan 18, 2025

• 💡 [REQUEST] - <title> How to Create the Multimodal embedding for text, image and videos using this model.

  #506 closed Jan 18, 2025

• 利用video_chatgpt finetune minicpm报错

  #552 closed Jan 18, 2025

• [BUG] <title>github项目无法拉取，是一个空项目

  #740 closed Jan 18, 2025

• [BUG] <title>使用github教程，通LLama factory无法微调minicpmo模型

  #737 closed Jan 18, 2025

• [BUG] WebDemo中使用Few Shot功能，如果不传图片无法正常使用

  #658 closed Jan 18, 2025

• [BUG] <title>Local WebUI Demo启动前端和后端服务都是8088端口，这个项目web版到底怎么运行能用起来？还要改什么配置？

  #723 closed Jan 18, 2025

• [BUG] MiniCPM-o 2.6 int4版本的下载链接是404

  #718 closed Jan 16, 2025

• [BUG] LLama-Factory微调MiniCPM-o，报错RuntimeError: a view of a leaf Variable that requires grad is being used in an in-place operation.

  #730 closed Jan 16, 2025

• [BUG] <title>Failure results

  #724 closed Jan 16, 2025

• [BUG] <通过代码或者直接在魔搭上推理minicpm-o-2_6的聊天机器人，上传图片或视频会出现Error, please retry>

  #726 closed Jan 16, 2025

• [BUG] <title>ollama官方文档进行的编译部署，无法与模型对话

  #453 closed Jan 15, 2025

• [ollama] - <title>能不能讓官方ollama支援使用這個模型啊，不然一直看得到用不到

  #454 closed Jan 15, 2025

• 💡 [REQUEST] - <title>Support for Videos

  #456 closed Jan 15, 2025

• [llamacpp] - <title>Llama.CPP HTTP server.

  #457 closed Jan 15, 2025

• [BUG]ollama运行2.6版本，聊天报错：llama_get_logits_ith: invalid logits id 10, reason: no logits

  #463 closed Jan 15, 2025

• Any support for llama.cpp on android?

  #524 closed Jan 15, 2025

• MiniCPM-V2.6是否可以部署到移动端设备？

  #593 closed Jan 15, 2025

• 💡 [REQUEST] - <title>MiniCPM-V 2.6 is supported in PaddleMIX by Paddle Team!

  #695 closed Jan 15, 2025

• [Question] how can use MiniCPM-o 2.6 on iphone

  #714 closed Jan 15, 2025

• Need for help!

  #715 closed Jan 15, 2025

35 Issues opened by 31 people

• [BUG] Error happened during generation: upper bound and larger bound inconsistent with step sign

  #774 opened Jan 21, 2025

• [BUG] <title> Creation of Product Categorization model by fine tuning MiniCPM-V-2_6 model on e-commerce product data

  #773 opened Jan 21, 2025

• MiniCPM-o 2.6有做过文字框或者目标框定位吗？

  #772 opened Jan 21, 2025

• torch.distributed.elastic.multiprocessing.errors.ChildFailedError（exitcode: -9）反复在同一个进度出现

  #770 opened Jan 21, 2025

• [BUG] Running hangs with ds zero3 and batch inference

  #769 opened Jan 21, 2025

• ollama创建openbmb/MiniCPM-o-2_6-gguf报错！

  #768 opened Jan 20, 2025

• 如何减少每张图所占的token数量

  #766 opened Jan 20, 2025

• [BUG] vllm 跑openbmb/MiniCPM-o-2_6-int4

  #765 opened Jan 20, 2025

• Safety

  #764 opened Jan 20, 2025

• [vllm]zmq.error.ZMQError: Operation not supported

  #762 opened Jan 20, 2025

• MiniCPM-o 2.6 gguf can't run on mac

  #761 opened Jan 19, 2025

• [BUG] ChatBot tab: upstream connect error or disconnect/reset before headers...

  #760 opened Jan 19, 2025

• [BUG] flash_attn is needed but not specified in requirements

  #754 opened Jan 19, 2025

• 如何理解【像素可达180万】？

  #753 opened Jan 19, 2025

• Speech SFT

  #752 opened Jan 18, 2025

• MiniCPM-o-2_6-gguf:F16

  #750 opened Jan 18, 2025

• 请问web_demo中聊天机器人为什么要 iframe src="https://minicpm-omni-webdemo-iframe.modelbest.cn"而不使用本地算力呢？

  #749 opened Jan 18, 2025

• [BUG] int4 量化模型，一直在不停胡言乱语

  #748 opened Jan 17, 2025

• 💡 [REQUEST] - 支持Audio的微调方案

  #745 opened Jan 17, 2025

• 树莓派能跑MiniCPM-o 2.6吗？

  #744 opened Jan 17, 2025

• 我看了下模型大小有十多个G，这样也能在手机上运行么？

  #743 opened Jan 17, 2025

• [BUG] <title>vllm部署调用报错：TypeError: Unknown image model type: minicpmo

  #742 opened Jan 17, 2025

• [BUG] android demo 发送消息后闪退

  #739 opened Jan 17, 2025

• realtime video call mode, the audio has a loud noise

  #738 opened Jan 17, 2025

• 💡 [REQUEST] - Add a Multimodal Streaming Gradio Demo

  #736 opened Jan 16, 2025

• [BUG] OpenBMB/llama.cpp build error and segmentation fault on Mac M3

  #734 opened Jan 16, 2025

• 记录一下解决minicpm-o-2.6的running的bug

  #732 opened Jan 16, 2025

• [BUG] 使用MiniCPM-o 2.6 int4模型时，chat.py报错

  #728 opened Jan 16, 2025

• [BUG] MiniCPM-o-2_6-gguf内存和磁盘占用满后一直卡住

  #727 opened Jan 16, 2025

• [BUG] web_demos/minicpm-o_2.6/model_server.py加载MiniCPM-o-2.6-Int4错误

  #725 opened Jan 15, 2025

• [BUG] is_torch_greater_or_equal_than_1_13 under transformers/src/transformers /pytorch_utils.py has been removed in version 4.48.0

  #722 opened Jan 15, 2025

• [BUG] minicpm-v:8b-2.6-q4_K_M Some pictures reported errors

  #721 opened Jan 15, 2025

• ollama issue!

  #720 opened Jan 15, 2025

• [BUG] `.to` is not supported for `4-bit` or `8-bit` bitsandbytes models. Please use the model as it is, since the model has already been set to the correct devices and casted to the correct `dtype`.

  #719 opened Jan 15, 2025

• [BUG] Failed to load MiniCPM-o-2_6-gguf using LM Studio

  #717 opened Jan 15, 2025

10 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• [BUG] <title>极端宽高比的图片无法识别，minicpm-v:8b-2.6-q2_K

  #712 commented on Jan 15, 2025 • 0 new comments

• [ollama] - <title>使用ollama跑minicpm-v时，内存不能释放，会一直占用内存直到死机

  #705 commented on Jan 15, 2025 • 0 new comments

• [llama] <.so版本不一致>

  #697 commented on Jan 15, 2025 • 0 new comments

• [llamacpp] - <title> 为什么llamacpp执行量化模型还要指定一个 f16的 mmproj-model-f16.gguf ?

  #657 commented on Jan 15, 2025 • 0 new comments

• [BUG] <title>llama.cpp CLIP cannot encode some images after building graph CLIP无法编码特定图片

  #650 commented on Jan 15, 2025 • 0 new comments

• [llamacpp] - <title> why gpu and cpu has very different output? and gpu looks worse.

  #474 commented on Jan 15, 2025 • 0 new comments

• [BUG] <title> finetune/dataset.py 有bug

  #652 commented on Jan 16, 2025 • 0 new comments

• [vllm] - <title> vllm 部署了API接口，示例里边有单张图片的推理，但是多张图片的推理，有参考示例吗？

  #674 commented on Jan 18, 2025 • 0 new comments

• 💡 [REQUEST] - Run model v2.6 on Apple Silicon with mps-Pytorch Backend

  #558 commented on Jan 20, 2025 • 0 new comments

• [BUG] <title> lora脚本微调错误

  #651 commented on Jan 20, 2025 • 0 new comments