thudm/visualglm-6b issues and pull requests

#91 - fix merge lora bug

Pull Request - State: closed - Opened by 1049451037 over 1 year ago

#90 - 模型微调后，合并LoRA部分的参数到原始的权重报错

Issue - State: open - Opened by chensiqin over 1 year ago - 2 comments

#89 - 有没有batch推理的实现呢？

Issue - State: open - Opened by CaicaiJason over 1 year ago - 4 comments

#88 - centos 环境使用pip安装依赖报错

Issue - State: open - Opened by shihzenq over 1 year ago - 2 comments

#87 - update lora

Pull Request - State: closed - Opened by 1049451037 over 1 year ago

#86 - preserve linear parallel lora

Pull Request - State: closed - Opened by 1049451037 over 1 year ago

#85 - QLoRA微调问题

Issue - State: closed - Opened by lljjson over 1 year ago - 9 comments

#84 - Update README.md

Pull Request - State: open - Opened by eltociear over 1 year ago

#83 - 推理qlora报错

Issue - State: closed - Opened by WangRongsheng over 1 year ago - 2 comments

#82 - finetune image size

Issue - State: open - Opened by hardlipay over 1 year ago - 5 comments

#81 - 读取图片太慢了

Issue - State: open - Opened by freelancerllm over 1 year ago - 13 comments

#80 - 微调需要多大的显存？

Issue - State: open - Opened by dizhenx over 1 year ago - 5 comments

#79 - 在本地服务器部署完成，能够启动，无法推理，报错“超时！请稍等几分钟再重试”

Issue - State: open - Opened by Baorui-up over 1 year ago - 2 comments

#78 - 模型微调出错

Issue - State: closed - Opened by 2132660698 over 1 year ago - 16 comments

#77 - add qlora

Pull Request - State: closed - Opened by 1049451037 over 1 year ago

#76 - 关于在centos后台运行后关闭secureCRT终端再通过浏览器访问出错的问题。

Issue - State: open - Opened by yinjiaoyuan over 1 year ago - 4 comments

#75 - Update requirements.txt

Pull Request - State: open - Opened by dizhenx over 1 year ago - 1 comment

#74 - utoModel.from_pretrained("../visualglm-6b", trust_remote_code=True).half().cuda() error

Issue - State: open - Opened by wavelet2008 over 1 year ago - 4 comments

#73 - 设置加载模型地址的config文件在哪？自动下载模型的地址在哪？感谢

Issue - State: open - Opened by haorunze over 1 year ago - 7 comments

#72 - 本地下载模型部署visualGLM，运行cli_demo.py脚本，文本对话正常，输入图片路径，报错：Could not load library libcublasLt.so.10

Issue - State: open - Opened by mknow over 1 year ago - 3 comments

#71 - ImportError: cannot import name 'COMMON_SAFE_ASCII_CHARACTERS' from 'charset_normalizer.constant' 有遇到吗？

Issue - State: open - Opened by fryeggs over 1 year ago - 1 comment

#70 - 上传照片后报错这个是什么原因

Issue - State: open - Opened by forkyguo over 1 year ago - 2 comments

#69 - python cli_demo.py 16gb电脑内存加载模型直接卡死了

Issue - State: open - Opened by 459936800 over 1 year ago - 1 comment

#68 - QLoRA微调支持

Issue - State: closed - Opened by WangRongsheng over 1 year ago - 1 comment

#67 - “超时！请稍等几分钟再重试。”啥情况？

Issue - State: open - Opened by szdengdi over 1 year ago - 4 comments

#66 - 后续的模型会加入物体检测定位的能力吗？

Issue - State: closed - Opened by RobinHan24 over 1 year ago - 1 comment

#65 - 加载VisualGLM-6b模型提示no attribute

Issue - State: open - Opened by majijia505 over 1 year ago - 4 comments

#64 - 如何让模型下载到指定目录

Issue - State: closed - Opened by PurpleMStone over 1 year ago - 2 comments

#63 - 使用本地下载模型时，打包上传的过程中校验MD5可能会出错的问题

Issue - State: open - Opened by zzxs over 1 year ago - 1 comment

#62 - 直接运行cli_demo.py报错No module named 'sat'，并不会自动下载sat模型

Issue - State: closed - Opened by Senna1960321 over 1 year ago - 7 comments

#61 - 微调完成后推理加载问题

Issue - State: closed - Opened by magicwang1111 over 1 year ago - 9 comments

#59 - 会开源训练代码和数据集么？

Issue - State: open - Opened by ldfandian over 1 year ago - 2 comments

#58 - Update finetune_visualglm.py

Pull Request - State: open - Opened by freelancerllm over 1 year ago

#57 - 我显卡的显存明明是80GB，为什么却提示只有很少？

Issue - State: closed - Opened by WangRongsheng over 1 year ago - 2 comments

#56 - 请教一下默认的image encoder是什么？

Issue - State: open - Opened by ldfandian over 1 year ago - 3 comments

#55 - 关于微调训练的问题

Issue - State: closed - Opened by WangRongsheng over 1 year ago - 10 comments

#54 - 运行finetune报错

Issue - State: open - Opened by GXKIM over 1 year ago - 2 comments

#53 - ERROR: Unknown arg use_final_layernorm

Issue - State: open - Opened by yaofeng over 1 year ago - 2 comments

#52 - Finetune error: RuntimeError: FIND was unable to find an engine to execute this computation

Issue - State: closed - Opened by yt7589 over 1 year ago - 2 comments

#51 - Finetune to others language ?

Issue - State: closed - Opened by batman-do over 1 year ago - 2 comments

#50 - 如果要基于 chatglm-6b 来训练一个多模态语言模型，glm 的 token部分该怎么处理呢？

Issue - State: closed - Opened by shtdbb over 1 year ago - 2 comments

#49 - readme.md remove ">>>"

Pull Request - State: closed - Opened by bobo0810 over 1 year ago

#48 - finetune原文件运行三处报错

Issue - State: closed - Opened by magicwang1111 over 1 year ago - 9 comments

#47 - 通过微调可以使模型拥有基本的中文OCR能力吗?

Issue - State: open - Opened by JustWB13 over 1 year ago - 7 comments

#46 - transformers version capitable with requirements.txt

Issue - State: closed - Opened by nameless0704 over 1 year ago - 1 comment

#45 - 微调需要什么配置？

Issue - State: closed - Opened by dpcross over 1 year ago - 1 comment

#44 - 打字机页面推理速度很慢

Issue - State: open - Opened by dpcross over 1 year ago - 2 comments

#43 - tuning 遇到了 RuntimeError: Error building extension 'fused_adam'

Issue - State: open - Opened by dhhcj over 1 year ago - 11 comments

#42 - transformers version?

Issue - State: closed - Opened by nameless0704 over 1 year ago - 3 comments

#41 - web应用没有显存清理机制

Issue - State: closed - Opened by Crazybean-lwb over 1 year ago - 2 comments

#40 - 生成答案的时候返回的token长度超出导致重启服务后Chatbot生成模块显示空白，其他模块正常

Issue - State: open - Opened by ddddlee1007 over 1 year ago - 1 comment

#39 - 话说能支持下苹果的MPS吗,现在mac m2上运行报错

Issue - State: open - Opened by wacdev over 1 year ago - 6 comments

#38 - fintune错误

Issue - State: closed - Opened by freelancerllm over 1 year ago - 14 comments

#37 - 有训练代码吗？

Issue - State: open - Opened by vanpersie32 over 1 year ago - 5 comments

#36 - 求问使用的 chatGLM 的版本

Issue - State: open - Opened by rover5056 over 1 year ago - 1 comment

#35 - web_demo.py 启动默认端口号是7860，可以指定为其它端口号吗？

Issue - State: closed - Opened by xiongmozhou over 1 year ago - 1 comment

#34 - 使用CPU加载模型推理时出现如下报错

Issue - State: open - Opened by yuunnn-w over 1 year ago - 3 comments

#33 - No space left on device，系统盘没有空间，能修改模型下载位置吗

Issue - State: closed - Opened by xiongmozhou over 1 year ago - 1 comment

#32 - modeling_chatglm.py里self.dtype具体是指？

Issue - State: open - Opened by iamsile over 1 year ago - 4 comments

#31 - Is this model image only or can it be used with text conversation as well?

Issue - State: open - Opened by git4sun over 1 year ago - 1 comment

#30 - "addmm_impl_cpu_" not implemented for 'Half'

Issue - State: open - Opened by tzwo over 1 year ago - 1 comment

#29 - 运行web_demo_hf.py时遇到tmp文件写入被拒的问题

Issue - State: closed - Opened by yangxingrui over 1 year ago - 4 comments

#28 - 求一个quantize好的模型

Issue - State: open - Opened by billzhao9 over 1 year ago - 3 comments
Labels: enhancement

#27 - model use left-padding or right-padding ??

Issue - State: closed - Opened by iamsile over 1 year ago - 2 comments

#26 - 可以直接用VisualGLM-6b进行reward model的训练吗？

Issue - State: closed - Opened by iamsile over 1 year ago - 2 comments

#25 - Is there will have a paper or a technical report?

Issue - State: open - Opened by diaoyechao over 1 year ago - 1 comment

#24 - cuDNN error: CUDNN_STATUS_NOT_INITIALIZED

Issue - State: closed - Opened by aka99 over 1 year ago - 4 comments

#23 - SAT_HOME里已经有下载好了模型了，运行web_demo里还要再下载一次呢？

Issue - State: closed - Opened by dydwgmcnl4241 over 1 year ago - 4 comments

#22 - pip install deepspeed error

Issue - State: open - Opened by luieswww over 1 year ago - 8 comments

#21 - 接受多张图片作为输入

Issue - State: open - Opened by tuteng0915 over 1 year ago - 1 comment

#20 - Update web_demo_hf.py

Pull Request - State: closed - Opened by tuteng0915 over 1 year ago - 2 comments

#19 - 训练细节和训练细节

Issue - State: closed - Opened by guozhiyao over 1 year ago - 1 comment

#18 - Torch not compiled with CUDA enabled

Issue - State: closed - Opened by MuCune over 1 year ago - 5 comments

#17 - ImportError: This modeling file requires the following packages that were not found in your environment: sat. Run `pip install sat`

Issue - State: closed - Opened by yuanjie-ai over 1 year ago - 3 comments

#16 - Create api_hf.py

Pull Request - State: closed - Opened by xszyou over 1 year ago - 1 comment

#15 - Issue with API mode: unexpected keyword argument 'mems'

Issue - State: open - Opened by git4sun over 1 year ago - 3 comments

#14 - 坐等大神给出微调代码

Issue - State: closed - Opened by xxm1668 over 1 year ago - 7 comments

#13 - 训练脚本以及训练细节是否会提供？

Issue - State: open - Opened by freelancerllm over 1 year ago - 1 comment

#12 - 上传图片第一次问答无论如何都是先输出描述

Issue - State: open - Opened by RipperTs over 1 year ago - 2 comments

#11 - huggingface上的visualglm-6b模型文件是不是有问题？

Issue - State: closed - Opened by kingdomad over 1 year ago - 6 comments

#10 - Update README.md

Pull Request - State: closed - Opened by chenshaoju over 1 year ago - 1 comment

#9 - 本地加载模型室提示找不到 model_config.json

Issue - State: closed - Opened by frnksmdlkedjnfr over 1 year ago - 13 comments

#8 - 输入一张数学试卷，输出了一堆随机数

Issue - State: closed - Opened by changjiuxiong over 1 year ago - 2 comments

#7 - 能否列出具体的硬件资源要求？

Issue - State: open - Opened by puppywang over 1 year ago - 2 comments
Labels: enhancement

#6 - windows 安装不了依赖

Issue - State: closed - Opened by happyi over 1 year ago - 4 comments

#5 - linux环境缺少sat包且pip无法安装成功

Issue - State: closed - Opened by ImGoodBai over 1 year ago - 6 comments

#4 - 一个badcase

Issue - State: closed - Opened by lucasjinreal over 1 year ago - 2 comments

#3 - web_demo_hf.py报错：RuntimeError: GET was unable to find an engine to execute this computation

Issue - State: closed - Opened by MingJiaAn over 1 year ago - 3 comments

#2 - 加入储存长期对话记忆的embedding space是否能解决上下文长度问题？

Issue - State: closed - Opened by win10ogod over 1 year ago - 1 comment

#1 - VisualGLM-6B的训练

Issue - State: closed - Opened by WangRongsheng over 1 year ago - 3 comments

GitHub / thudm/visualglm-6b issues and pull requests