Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / OpenBMB/MiniCPM-V issues and pull requests
#589 - [BUG] <title> bug when converting hf to gguf for minicpm-v-2.6
Issue -
State: open - Opened by chongkuiqi 1 day ago
#588 - [BUG] <title>what are the system requirements for finetuning "openbmb/MiniCPM-Llama3-V-2_5"
Issue -
State: open - Opened by vjaideep08 1 day ago
#587 - [BUG] <在funetune/dataset.py中报告image start token != image end tokens的错误>
Issue -
State: open - Opened by KeepFaithMe 3 days ago
- 1 comment
#586 - How to use the forward call?
Issue -
State: open - Opened by wj210 3 days ago
#585 - 全量微调vision端不稳定
Issue -
State: open - Opened by 128Ghe980 3 days ago
- 6 comments
#584 - How not to save the files in global_step1 during training
Issue -
State: open - Opened by orderer0001 4 days ago
#583 - 想学习【MLLM多模态】快加学习交流群
Issue -
State: closed - Opened by km1994 5 days ago
- 1 comment
Labels: question
#582 - how to modify the system prompt
Issue -
State: open - Opened by orderer0001 5 days ago
- 2 comments
#581 - [BUG] <title>dataset.py文件conversation_to_ids方法input_ids调用shape属性错误
Issue -
State: closed - Opened by KeepFaithMe 7 days ago
#580 - [BUG] <title> Line 432 in MiniCPM-V/finetune/dataset.py, image_placeholder referenced before assignment
Issue -
State: open - Opened by 1ittlesnow 7 days ago
#579 - fix finetune minicpm error
Pull Request -
State: open - Opened by jackyjinjing 7 days ago
#578 - [BUG] finetune minicpm error
Issue -
State: open - Opened by jackyjinjing 7 days ago
- 2 comments
#577 - Finetuning issue
Issue -
State: open - Opened by ragesh2000 7 days ago
- 1 comment
#576 - [BUG] <title> problem when awq self-quantization
Issue -
State: open - Opened by chongkuiqi 7 days ago
- 6 comments
#575 - [BUG] Screenshot of code. OCR does not result in correct code
Issue -
State: open - Opened by woutervancgi 8 days ago
#574 - 是否可以透漏in-context learning是怎么训练的吗
Issue -
State: open - Opened by xushilin1 8 days ago
- 1 comment
#573 - OCR Det
Issue -
State: open - Opened by Hiwyl 8 days ago
- 1 comment
#572 - 请问哪里能找到openai api server部署和调用示例
Issue -
State: closed - Opened by yangxin60-tal 8 days ago
- 1 comment
#571 - symbol(s) not found for architecture arm64
Issue -
State: closed - Opened by ZhonghaoLu 9 days ago
- 1 comment
#570 - [question] Is the resource utilization correct when fine tuning?
Issue -
State: open - Opened by styughjvbn 9 days ago
#569 - which way is the right way to use special token
Issue -
State: open - Opened by 128Ghe980 10 days ago
#568 - 💡 [REQUEST] - <title> how to watch GradCAM by using MiniCPM-V in VQA task
Issue -
State: open - Opened by V2LLAIN 10 days ago
Labels: question
#567 - 2D Positional Embedding Interpolation
Issue -
State: open - Opened by tbsxxxH 11 days ago
#566 - minicpm v2.5 vs v2.6
Issue -
State: open - Opened by wucx888 12 days ago
- 1 comment
#565 - [BUG] <title>whether it is possible to deploy minicpmv2.5 on Android based on llama.cpp (using Qualcomm GPU for inference)?
Issue -
State: open - Opened by My-captain 12 days ago
- 1 comment
#564 - 💡 [REQUEST] - <title> 请问下如何并行推理多个问题呢,demo是单个图片单个问题,可以多个图片,每个图片一个问题按batch推理吗
Issue -
State: open - Opened by dearsar 13 days ago
Labels: question
#563 - [BUG] MiniCPM-Llama3-V-2_5-int4微调出现:RuntimeError: only Tensors of floating point dtype can require gradients
Issue -
State: open - Opened by KeepFaithMe 14 days ago
- 2 comments
#562 - [QUESTION] <smaller language model>
Issue -
State: closed - Opened by ztianlin 15 days ago
- 1 comment
#561 - [BUG] <title>error:unknown argument: --video
Issue -
State: open - Opened by pily1 15 days ago
- 1 comment
#560 - 💡 [REQUEST] - <title>minicpmv2.5的量化支持,autoGPTQ,autoAWQ
Issue -
State: open - Opened by ZHANG-SH97 16 days ago
- 1 comment
Labels: question
#559 - [BUG] <title> merge_and_unload完之后和原来模型参数完全一致,没有变化
Issue -
State: open - Opened by Gary-code 17 days ago
- 1 comment
#558 - 💡 [REQUEST] - Run model v2.6 on Apple Silicon with mps-Pytorch Backend
Issue -
State: open - Opened by RobKnop 17 days ago
Labels: question
#557 - Inference after lora finetuning
Issue -
State: closed - Opened by tristan279 17 days ago
- 1 comment
#556 - Improve exception messages for better readability and error context
Pull Request -
State: open - Opened by mandlinsarah 18 days ago
#555 - [vllm] -
Issue -
State: open - Opened by WoutDeRijck 18 days ago
- 1 comment
Labels: question
#554 - [BUG] <title>是否考虑支持华为910B 300I这种部署 ? 支持NPM推理及训练
Issue -
State: open - Opened by ipv6ok 18 days ago
#553 - [BUG] <title>请问该代码是不支持minicpm-v的模型吗
Issue -
State: closed - Opened by baymax591 18 days ago
- 7 comments
#552 - 利用video_chatgpt finetune minicpm报错
Issue -
State: open - Opened by zhaoyangwei123 19 days ago
- 8 comments
#551 - 💡 [REQUEST] - 为什么没有模型的定义代码?
Issue -
State: closed - Opened by jiaowei-it 19 days ago
- 1 comment
Labels: question
#550 - Merge issue with MiniCPM-Llama-v2.5 lora adapters with base Llama-3-8b text model
Issue -
State: closed - Opened by SrikanthChellappa 19 days ago
- 25 comments
#549 - [BUG] <title>模型正常,使用原始模型可以正常推理,微调过后出现 RuntimeError: probability tensor contains either inf, nan or element < 0
Issue -
State: closed - Opened by Hiwyl 19 days ago
- 1 comment
#548 - [BUG] <title>自己准备的数据集,在加载的时候,出错:image start token != image end tokens,data fetch error
Issue -
State: closed - Opened by XHB-ZMM 20 days ago
- 6 comments
#547 - [BUG] <title>端侧Demo无法下载体验
Issue -
State: closed - Opened by pangpang-xuan 20 days ago
#546 - 微调MiniCPM-v2.6以获得更好的单图片日文OCR和推理能力
Issue -
State: open - Opened by ZijianCHEN-infodeliver 21 days ago
#545 - [BUG] <title>最大输入长度--model_max_length参数传递无效
Issue -
State: open - Opened by GenuineWWD 21 days ago
- 2 comments
#544 - MiniCPM-Llama3-V 2.5 vs 2.6
Issue -
State: closed - Opened by kotaxyz 21 days ago
- 2 comments
#543 - Modify eval_mm for MiniCPM-V 2.6
Pull Request -
State: closed - Opened by lihytotoro 22 days ago
#542 - Android端部署
Issue -
State: closed - Opened by xiahc 22 days ago
- 1 comment
#541 - [BUG] <title> 我在 vllm v0.5.5 版本运行报错
Issue -
State: open - Opened by fastaipod 22 days ago
- 4 comments
#540 - How to use vllm openai api server for video reasoning ?
Issue -
State: open - Opened by flclain 22 days ago
#538 - [BUG] <title>RuntimeError: Expected index [1728, 4096] to be smaller than self [2048, 4096] apart from dimension 0 and to be smal ler size than src [864, 4096]
Issue -
State: closed - Opened by laoShuaiGe 23 days ago
- 1 comment
#537 - 有关lora finetune显存问题
Issue -
State: closed - Opened by zhaoyangwei123 23 days ago
- 1 comment
#536 - [BUG] <title>minicpm-v-2.6和LLaVA-OneVision的区别是什么?
Issue -
State: closed - Opened by AbbottKilig 23 days ago
- 1 comment
#534 - [BUG] <title>
Issue -
State: closed - Opened by 2013358072 23 days ago
- 1 comment
#533 - 💡 [REQUEST]微调进度100%后报错 Watchdog caught collective operation timeout: WorkNCCL
Issue -
State: closed - Opened by xiaobaixue 23 days ago
- 3 comments
Labels: question
#531 - minicpm-v-2.6在OpenCompass评测结果65.2是怎么得到的?
Issue -
State: closed - Opened by ninth99 24 days ago
- 1 comment
#530 - [BUG] <title>
Issue -
State: closed - Opened by Gary-code 24 days ago
- 1 comment
#529 - [BUG] <title>MiniCPM-V-2_6 评估指标没有对齐
Issue -
State: closed - Opened by luyao-cv 24 days ago
- 1 comment
#528 - 我在做垂直领域应用,想先让模型通过预训练学一些文档材料,再进行图文对话微调,这个怎么操作?
Issue -
State: closed - Opened by zpge 24 days ago
- 1 comment
#526 - 请问是否可以将vpm和resampler 这2个模型转为tensorrt来加速推理呢?
Issue -
State: closed - Opened by wu1143100799 24 days ago
- 1 comment
#525 - MiniCPM-V 2.6 int4 24G显存不够用,model.cuda()会OOM
Issue -
State: closed - Opened by MaxKinny 24 days ago
- 4 comments
#523 - [BUG] <title>multi-image微调
Issue -
State: closed - Opened by WorldHellooo 24 days ago
- 1 comment
#522 - about multi-image finetune
Issue -
State: closed - Opened by univa-JASON 24 days ago
- 4 comments
#521 - Finetuning feature added for setting `vision_lr` and `resampler_lr`
Pull Request -
State: open - Opened by 2U1 25 days ago
- 1 comment
#519 - 微调问题
Issue -
State: closed - Opened by 0704highlight 25 days ago
- 1 comment
#518 - [BUG] <title>minicpm-v2.6无法输出准确坐标
Issue -
State: closed - Opened by heart-du 26 days ago
- 2 comments
#513 - Issue with finetuning
Issue -
State: closed - Opened by ragesh2000 26 days ago
- 2 comments
#510 - 端模型跑vlm的延时
Issue -
State: closed - Opened by dingtine 27 days ago
- 1 comment
#509 - [BUG] <title>Error while running `Chat with video`
Issue -
State: closed - Opened by ShuyUSTC 27 days ago
- 3 comments
#508 - I want to set up system prompt during training.
Issue -
State: closed - Opened by orderer0001 28 days ago
- 1 comment
#504 - [BUG] <title>The bug of VLLM for minicpmv
Issue -
State: closed - Opened by Mysnake about 1 month ago
- 1 comment
#503 - 💡 [REQUEST] - 如何复现 OCRBench 精度,是否有专门的提示词?
Issue -
State: closed - Opened by GarrickLin about 1 month ago
- 6 comments
Labels: question
#501 - 为了方便技术交流,拉了一个多模态大模型技术交流群,有需要的大家可以加入
Issue -
State: closed - Opened by feihuamantian about 1 month ago
- 1 comment
#492 - [BUG] 请问有在 TextVQA, VizWiz, VQAv2, OK-VQA 上的 few shots 的评估脚本吗?
Issue -
State: closed - Opened by davidlight2018 about 1 month ago
- 1 comment
#486 - [BUG] Data fetch error - typo
Issue -
State: open - Opened by anthisyme about 1 month ago
- 5 comments
#484 - Update streamlit implementation for MiniCPM-V 2.6
Pull Request -
State: closed - Opened by JamePeng about 1 month ago
#483 - [BUG] number of image start tokens and image end tokens mismatch
Issue -
State: closed - Opened by Aguin about 1 month ago
- 4 comments
#482 - [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
Issue -
State: open - Opened by Fridayfairy about 1 month ago
- 4 comments
#481 - [BUG] 训练速度问题
Issue -
State: closed - Opened by JianbangZ about 1 month ago
- 1 comment
#469 - [BUG] <title>vllm的视频推理示例代码明显错误,且会在vllm/model_executor/models/minicpmv.py报错
Issue -
State: closed - Opened by younger-diao about 1 month ago
- 5 comments
#466 - [BUG] web_demo_2.6.py RuntimeError: probability tensor contains either `inf`, `nan` or element < 0
Issue -
State: closed - Opened by xlxxcc about 1 month ago
- 5 comments
#463 - [BUG]ollama运行2.6版本,聊天报错:llama_get_logits_ith: invalid logits id 10, reason: no logits
Issue -
State: open - Opened by zhb-code about 1 month ago
- 7 comments
#453 - [BUG] <title>ollama官方文档进行的编译部署,无法与模型对话
Issue -
State: open - Opened by Evelynn-n about 1 month ago
- 6 comments
#433 - MAC本地ollama部署2.6版本,‘cblas.h’文件未找到
Issue -
State: closed - Opened by zhb-code about 1 month ago
- 16 comments
#418 - [BUG] Error: llama runner process has terminated: signal: aborted (core dumped)
Issue -
State: closed - Opened by anlogo about 1 month ago
- 5 comments
#408 - [BUG] <title>'MiniCPMVTokenizerFast' object has no attribute 'image_processor'
Issue -
State: closed - Opened by Labmem009 about 2 months ago
- 8 comments
#402 - 修改了微信入口
Pull Request -
State: closed - Opened by LDLINGLINGLING about 2 months ago
- 1 comment
#383 - Fine tuning of MiniCPM-Llama3-V-2_5-int4
Pull Request -
State: open - Opened by TejMakode1523 about 2 months ago
- 1 comment
#361 - [BUG] <title>MiniCPM-V-2的vLLM推理不能使用
Issue -
State: closed - Opened by y-x-x123 2 months ago
- 12 comments
#345 - RuntimeError: Expected index [1728, 4096] to be smaller than self [2048, 4096] apart from dimension 0 and to be smal ler size than src [864, 4096]
Issue -
State: closed - Opened by aoji0606 2 months ago
- 3 comments
#335 - CPMV2.0使用zero3全量微调依然报错
Issue -
State: closed - Opened by daihuidai 2 months ago
- 5 comments
#334 - 增加了量化脚本,swift和xinference的推理readme,在readme中增加了常用模块和新模块的快速导航
Pull Request -
State: closed - Opened by LDLINGLINGLING 2 months ago
- 3 comments
#331 - [BUG] <在使用最新代码微调时遇到问题RuntimeError: Sizes of tensors must match except in dimension 1.>
Issue -
State: closed - Opened by Single430 2 months ago
- 5 comments
Labels: Finetune
#316 - [BUG] <为什么在微调代码中的dataset.py 中读取conversation时要remove bos>
Issue -
State: closed - Opened by Single430 3 months ago
- 3 comments
#300 - [Data info] MiniCPM-llama3-V 2.5
Issue -
State: closed - Opened by emanuelevivoli 3 months ago
- 1 comment
#293 - Update inference_on_multiple_gpus.md
Pull Request -
State: closed - Opened by glenn124f 3 months ago
#268 - [BUG] <title>MiniCPM-V2.0不支持zero3全量微调
Issue -
State: closed - Opened by daihuidai 3 months ago
- 17 comments
#266 - MiniCPM-Llama3-V 2.5 int4 版本支持微调吗?
Issue -
State: closed - Opened by myBigbug 3 months ago
- 7 comments
#196 - Project author team stay tuned: I found out that the llama3-V project is stealing a lot of academic work from MiniCPM-Llama3-V 2.5
Issue -
State: closed - Opened by pzc163 4 months ago
- 53 comments
Labels: SPECIAL ATTENTION
#72 - 流式输出
Issue -
State: closed - Opened by sssssshf 5 months ago
- 2 comments
Labels: duplicate