Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / mymusise/ChatGLM-Tuning issues and pull requests

#273 - 关于保存的adapter_model.bin无实际推理效果的问题

Issue - State: open - Opened by ssgg-code 7 months ago - 2 comments

#270 - 请问这个项目支持chatglm3吗

Issue - State: open - Opened by xhrt 8 months ago

#267 - fix finetune / infer script.

Pull Request - State: closed - Opened by vikotse 10 months ago

#266 - 问题请教

Issue - State: open - Opened by pengcheng-yan 11 months ago

#265 - 对chatglm2使用lora修改之后的提交

Pull Request - State: closed - Opened by duyupeng 11 months ago - 3 comments

#264 - Cheyujie

Pull Request - State: closed - Opened by laocheyujie about 1 year ago

#263 - 这个项目停更了吗

Issue - State: open - Opened by shangzhensen about 1 year ago

#261 - AttributeError: 'ChatGLMModel' object has no attribute 'lm_head'

Issue - State: open - Opened by Daoming009 about 1 year ago - 3 comments

#260 - 请问如何读取checkpoint继续训练？

Issue - State: open - Opened by TingJShen about 1 year ago - 2 comments

#259 - 微调语料格式转换出现乱码

Issue - State: closed - Opened by suboteZjc about 1 year ago - 1 comment

#258 - finetune数据使用data_collator时报错 KeyError：seq_len

Issue - State: closed - Opened by michael0905 about 1 year ago - 2 comments

#256 - 请教一个问题，data_collator中不需要实现attention mask么？

Issue - State: closed - Opened by annw0922 about 1 year ago - 2 comments

#254 - 如何支持多卡跑

Issue - State: open - Opened by a1120235890 about 1 year ago

#253 - 修改max_seq_length好像并没有生效？

Issue - State: closed - Opened by tjulh about 1 year ago - 1 comment

#252 - 请问大佬是否有计划可以支持下qlora？

Issue - State: open - Opened by wuzechuan about 1 year ago - 1 comment

#251 - 怎么做evaluate，计算bleu 和rougue值之类的

Issue - State: open - Opened by ArtificialZeng about 1 year ago

#250 - Lora微调chatglm-6b后chekpoint里面缺乏adapter_config.json文件

Issue - State: open - Opened by wickedvalley about 1 year ago - 1 comment

#249 - 请问大佬什么时候能更新支持chatglm2呢？

Issue - State: open - Opened by pdwfree about 1 year ago - 13 comments
Labels: Feature request

#247 - 有没有用macos来finetune的啊

Issue - State: open - Opened by Json0926 over 1 year ago

#242 - data_collator有问题？

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

#241 - Fix commands in README

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

#240 - 微调的时候可以使用负样本吗？

Issue - State: open - Opened by crasylph over 1 year ago

#239 - tokenize_dataset_rows.py运行报错

Issue - State: closed - Opened by LuJH12 over 1 year ago - 2 comments

#238 - 咨询一下，训练后的模型，怎么在Chatglm-6b上使用？

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

#237 - 咨询一下，支持多显卡么

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

#234 - 期待支持RLHF

Issue - State: open - Opened by yanni-code over 1 year ago

#233 - 怎样强化模型在某些方面的能力？

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

#232 - 如何做batchsize>1的infer

Issue - State: open - Opened by BshoterJ over 1 year ago

#231 - 好像是bitsandbytes的错

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

#230 - 尝试在自己的数据集上使用LORA微调 但是报错

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

#229 - examples/infer_pretrain.ipynb 执行报错

Issue - State: open - Opened by 450586509 over 1 year ago - 1 comment

#227 - 训练后加载模型好像没有效果，这是什么情况？

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

#225 - No model named 'torch._six'

Issue - State: open - Opened by huanan254 over 1 year ago

#219 - 推理Can't find 'adapter_config.json'

Issue - State: open - Opened by jiayi37u over 1 year ago - 5 comments

#217 - [BUG] data pre process bug

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

#215 - finetuning.py是不是不支持int8的权重，只能使用f16权重哈

Issue - State: open - Opened by zlht812 over 1 year ago - 2 comments

#213 - Finetune前后预测结果Answer都是?? ??这样的

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

#212 - 有一个代码上的问题

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

#211 - 建议代码更新一下

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

#209 - 关于加入验证数据的问题

Issue - State: open - Opened by ai169 over 1 year ago - 3 comments

#206 - 完全学习不到数据集的内容

Issue - State: open - Opened by starhui70520 over 1 year ago - 4 comments

#203 - 对话时报错，RuntimeError: self and mat2 must have the same dtype

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

#200 - RuntimeError: expected scalar type Half but found Float

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

#197 - 训练时模型出错

Issue - State: open - Opened by lelegogo26 over 1 year ago - 1 comment

#195 - 请教下，如何进行增量fine-tune

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

#192 - 关于jsonl打开是乱码

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

#190 - expected scalar type Half but found Float while inference

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

#183 - ValueError: 130000 is not in list

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

#182 - 大佬，问一下，多轮对话的数据组织形式是什么？

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

#179 - expected scalar type Half but found Float

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

#176 - 为什么转换成{context:"",target:""}的格式？是哪里定义的吗？

Issue - State: open - Opened by ze00ro over 1 year ago - 3 comments

#175 - 没有看到RLHF的代码

Issue - State: open - Opened by dongdongrj over 1 year ago - 41 comments

#166 - CUDA error: device-side assert triggered是什么情况？

Issue - State: open - Opened by z1968357787 over 1 year ago - 9 comments

#150 - ValueError: 130004 is not in list

Issue - State: open - Opened by itxingqing over 1 year ago - 20 comments

#146 - infer报错

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

#138 - 推理失败 找不到peft_model.py

Issue - State: open - Opened by magnificent1208 over 1 year ago - 3 comments

#131 - 训练速度

Issue - State: open - Opened by huangtao36 over 1 year ago - 9 comments

#122 - 请问chatglm有继续在领域纯文本上进行pretrain的方法吗？

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

#120 - ValueError: weight is on the meta device, we need a `value` to put in on 0.

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

#110 - 抛出异常 No module named 'transformers_modules.'

Issue - State: open - Opened by dayu1979 over 1 year ago - 11 comments

#106 - ValueError: Please specify target_modules in peft_config

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

#101 - 微调结果推理加速

Issue - State: closed - Opened by yxk9810 over 1 year ago - 3 comments

#100 - Fix int8 training

Pull Request - State: open - Opened by sandofree over 1 year ago - 2 comments

#97 - 微调时token长度

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

#95 - [训练问题/错误]求助，如何大模型使用f16时进行LoRA训练

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

#94 - fix prefix attention mask

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

#93 - 请问大家怎么判断模型微调之后是否生效？

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

#92 - 显卡V100 32G，batch_size=1 仍然报OOM

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

#91 - 导出为pytorch格式

Issue - State: closed - Opened by MrToy over 1 year ago

#90 - 训练好后，推理速度很慢

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

#89 - LOSS下降太快

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

#87 - 什么样的显卡能满足训练需求？

Issue - State: closed - Opened by netwolf712 over 1 year ago - 4 comments

#86 - 增加每个checkpoint的config文件的存储，防止加载错误

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