lyhue1991/torchkeras issues and pull requests

#104 - notebook 示例中使用 save_pretrained 保存 model/checkpoints 的方法有问题

Issue - State: open - Opened by lalalabox 15 days ago

#103 - 一个小问题

Issue - State: open - Opened by Satoshi-Yuen 23 days ago

#102 - 能否添加GAN的支持

Issue - State: open - Opened by zhang090210 about 1 month ago

#101 - how can we add the conversation templates for new LLMs such as Llama3 and Qwen2?

Issue - State: open - Opened by richardsun-voyager about 1 month ago

#100 - module 'torchkeras' has no attribute 'tabular'

Issue - State: open - Opened by jakestale 2 months ago - 1 comment

#99 - FTTransformer 遇到 ModuleNotFoundError: No module named 'omegaconf'

Issue - State: open - Opened by GoldAndRabbit 2 months ago - 1 comment

#98 - model.save_pretrained("waimai_10k_bert") 保存模型报错

Issue - State: open - Opened by haozaiiii 2 months ago - 2 comments

#97 - 不知道是不是版本问题，terminal里老是报warning导致一闪一闪

Issue - State: open - Opened by Charity333333333 2 months ago

#96 - 关于断点重新训练后先前的记录图像会覆盖掉，不能再次基础上进行记录

Issue - State: open - Opened by chenzean 3 months ago

#95 - 请问下 pt 模型文件怎么转化成 transformers 文件结构

Issue - State: open - Opened by CurtisDai 4 months ago

#94 - 调度器学习率的更新是否该在EpochRunner中进行

Issue - State: open - Opened by LeoJhonSong 4 months ago

#93 - list index out of range

Issue - State: open - Opened by luoclab 5 months ago - 1 comment

#91 - ValueError: Attempting to unscale FP16 gradients.

Issue - State: open - Opened by Shubo-Ji 7 months ago - 2 comments

#90 - Update lightgbm.py

Pull Request - State: closed - Opened by zhangyu345293721 7 months ago

#89 - ValueError: Can't find 'adapter_config.json'

Issue - State: open - Opened by xuanyuc 9 months ago

#88 - 模型好像更新了，无法调用

Issue - State: closed - Opened by 14790897 9 months ago - 2 comments

#87 - 能否写一个微调VL-Qwen模型的例子

Issue - State: open - Opened by whysirier 9 months ago

#86 - #

Issue - State: closed - Opened by QimingXuzzuli 10 months ago

#85 - 大佬，能不能增加一下注释，看着有些混乱

Issue - State: open - Opened by liunian1217 10 months ago

#84 - FileNotFoundError: [Errno 2] No such file or directory: 'qwen7b_multirounds/adapter_model.bin' 如何解决

Issue - State: open - Opened by whysirier 10 months ago - 8 comments

#83 - 老铁，能不能把代码的注释加上，代码的参数用途都不晓得，读起来需要根据示例看，太费劲了~~。

Issue - State: open - Opened by Aioim 10 months ago

#82 - Add comments according to Google Code Rules

Pull Request - State: closed - Opened by zhangyu345293721 10 months ago - 1 comment

#81 - 代码中增加一些注释

Pull Request - State: closed - Opened by zhangyu345293721 10 months ago

#80 - master:按照谷歌代码规范增加注释

Pull Request - State: closed - Opened by zhangyu345293721 10 months ago

#79 - python3.7版本使用accelerate报错

Issue - State: open - Opened by zhangyu345293721 10 months ago

#78 - 建议增加参数

Issue - State: open - Opened by jourmore 10 months ago

#77 - 你好，我再vscode中进行使用，好像无法支持vsocde，您有别的方案推荐么，使用wandb，服务器无法连接翻墙

Issue - State: open - Opened by Re-fused 10 months ago

#76 - AttributeError: 'Parameter' object has no attribute 'weight'

Issue - State: closed - Opened by Lxhnnn 10 months ago - 1 comment

#75 - BaiChuan2_13B微调范例，报错“NotImplementedError: Cannot copy out of meta tensor; no data!”

Issue - State: open - Opened by 1022movies 11 months ago - 1 comment

#74 - AttributeError: 'ChatGLM' object has no attribute 'top_p'

Issue - State: open - Opened by Andyszl 11 months ago - 4 comments

#73 - 怎么保存指定帧的参数？

Issue - State: closed - Opened by jourmore 11 months ago

#72 - 国内镜像源还没有3.9.4版本

Issue - State: closed - Opened by luojueling 11 months ago - 1 comment

#71 - YOLOV8_Segment——ultralytics.ipynb 中 torchkeras梦中情炉接口训练报错：NotImplementedError: Module [YOLO] is missing the required "forward" function

Issue - State: open - Opened by new-programer 12 months ago

#70 - 执行ds_train_raw.map(preprocess,batched=True, num_proc=4,remove_columns=ds_train_raw.column_names)时报NameError: name 'cfg' is not defined

Issue - State: open - Opened by Chalous 12 months ago - 1 comment

#69 - chatglm注入单条知识的时候出现错误"Target modules ['query', 'value'] not found in the base model. Please check the target modules and try again.怎么解决"

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

#68 - Qwen-7B多轮对话运行训练的时候报错

Issue - State: closed - Opened by ifyoulovexxz about 1 year ago

#67 - 更看不懂了

Issue - State: closed - Opened by Hzzhang-nlp about 1 year ago

#66 - 我想进行多模态机器翻译，但是我加了图像特征，并且融合后无法进行推理，这种大佬有头绪吗？

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

#65 - 能不能支持Apple Mac GPU？

Issue - State: closed - Opened by Voldemort1986 about 1 year ago

#64 - Qwen-7b微调内容有误

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

#63 - baichuan13b enr 训练出错

Issue - State: closed - Opened by zhouzhou0322 about 1 year ago

#62 - 开源ner数据集分享

Issue - State: closed - Opened by zhouzhou0322 about 1 year ago

#61 - 多卡跑baichuan2报错

Issue - State: closed - Opened by uloveqian2021 about 1 year ago - 1 comment

#60 - 求问：fit_ddp方法报ProcessExitedException: process 0 terminated with signal SIGSEGV错误

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

#59 - 训练模型报错RuntimeError: Tensor on device meta is not on the expected device cuda:0!

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

#58 - 运行ChatGLM2_AdaLoRA——transformers.py 时定义模型报错Target modules ['query', 'value'] not found in the base model.

Issue - State: closed - Opened by wendyDing1102 about 1 year ago - 4 comments

#57 - torchkeras怎么支持每隔n个step保存一次模型

Issue - State: closed - Opened by terminator123 about 1 year ago - 1 comment

#56 - 对断点续训支持得不太好

Issue - State: closed - Opened by hhr114 about 1 year ago - 1 comment

#55 - baichuan13b是否支持多轮对话格式的训练

Issue - State: closed - Opened by YeHaijia about 1 year ago

#54 - 大佬你好 transformer那一章节，后面训练的时候用的是CPU。能否改成GPU呢？或者怎么改一下可以使用GPU进行训练

Issue - State: closed - Opened by willwilling about 1 year ago - 9 comments

#53 - BaiChuan13B_NER——transformers.ipynb训练过程中多卡问题

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

#52 - BaiChuan13B_NER——transformers模型微调后结果变小

Issue - State: closed - Opened by looperEit about 1 year ago

#51 - 注册chatglm魔法命令出错

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

#50 - cannot import name 'LightModel' from 'torchkeras'

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

#49 - where is this dataset "dfner_13k.pkl"

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

#48 - ChatGLM2_AdaLoRA——transformers.ipynb

Issue - State: closed - Opened by zx12671 about 1 year ago - 1 comment

#47 - ModuleNotFoundError: No module named 'transformers.models.mctct.configuration_mctct'

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

#46 - ImportError: cannot import name 'MiniLogCallback' from 'torchkeras.kerascallbacks'

Issue - State: closed - Opened by hhr114 about 1 year ago - 4 comments

#45 - AttributeError: partially initialized module 'torchkeras' has no attribute 'KerasModel' (most likely due to a circular import)

Issue - State: closed - Opened by MaybeRichard about 1 year ago

#44 - ChatGLM+LoRA微调时第一个epoch跑完计算best_score时数组溢出IndexError: outofindex

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

#43 - chatGLM AdaLoRA In[5]报错

Issue - State: closed - Opened by yanyang1024 about 1 year ago - 2 comments

#42 - RuntimeError: Expected is_sm80 to be true, but got false.

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

#41 - 大神，什么时候考虑支持下baichuan得adarola

Issue - State: closed - Opened by dayu1979 about 1 year ago - 1 comment

#40 - 运行chatglm2_AdaLoRA报错

Issue - State: closed - Opened by ddzz0210 about 1 year ago - 2 comments

#39 - 对学习率衰减支持得不好

Issue - State: closed - Opened by hhr114 about 1 year ago - 7 comments

#38 - 保存模型后调用报错 clean_up_tokenization_spaces 显示类型错误，感觉不应该

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

#37 - B站链接失效--ChatGLM2_QLoRA_Kaggle——transformers.ipynb

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

#36 - 是否能支持多卡训练

Issue - State: open - Opened by lacusrinz about 1 year ago - 8 comments

#35 - ModuleNotFoundError: No module named 'torchkeras.chat'

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

#34 - ChatGLM2-6b微调保姆级教程 - Colab支持&更多的注释详解

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

#33 - 执行验证模型时报错

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

#32 - CHATGLM2_ADALORA——TRANSFORMERS.IPYNB tokenizer.build_prompt报错

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

#31 - 关于模型效果的疑惑

Issue - State: closed - Opened by for1314ai about 1 year ago

#30 - AttributeError: 'AcceleratorState' object has no attribute '_mixed_precision'

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

#29 - AttributeError: 'ChatGLMTokenizer' object has no attribute 'build_prompt'

Issue - State: closed - Opened by for1314ai about 1 year ago - 1 comment

#28 - 在运行ChatGLM2_AdaLoRA——transformers.ipynb 执行到22步 ds_train = ds_train_raw.map时报错

Issue - State: closed - Opened by BoFan-tunning about 1 year ago - 3 comments

#27 - 测试chatglm2的adarola的时候报错如下

Issue - State: closed - Opened by dayu1979 about 1 year ago - 2 comments

#26 - 在训练模型时出现RuntimeError: Tensor on device meta is not on the expected device cuda:0!

Issue - State: closed - Opened by YinSonglin1997 about 1 year ago - 2 comments

#25 - model.enable_input_require_grads() 报错

Issue - State: closed - Opened by KMnO4-zx about 1 year ago - 1 comment

#24 - PicklingError: Cannot pickle a prepared model with automatic mixed precision, please unwrap the model with `Accelerator.unwrap_model(model)` before pickling it.

Issue - State: closed - Opened by lucheng07082221 about 1 year ago - 4 comments

#23 - 用ChatGLM2-transformers.ipynb和waimai_10k.csv微调模型不收敛

Issue - State: closed - Opened by CRonaldo1997 about 1 year ago - 2 comments

#22 - UnboundLocalError: local variable 'epoch_log' referenced before assignment

Issue - State: closed - Opened by ChaohuaZhang about 1 year ago - 3 comments

#21 - deepspeed是否支持？

Issue - State: closed - Opened by xxm1668 about 1 year ago - 3 comments

#20 - Expected all tensors to be on the same device, but found at least two devices, cuda:3 and cuda:0! (when checking argument for argument target in method wrapper_CUDA_nll_loss_forward

Issue - State: closed - Opened by amwork2020 about 1 year ago - 1 comment

#19 - 多gpu报错

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

#18 - 多gpu报错

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

#17 - loss计算传入参数问题

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

#16 - torchkeras.LightModel can't be used!

Issue - State: closed - Opened by Srf205 over 1 year ago - 2 comments

#15 - 怎么使用mps来进行训练

Issue - State: closed - Opened by zyq-20000901 over 1 year ago - 1 comment

#14 - ImportError: cannot import name 'Model' from 'torchkeras'

Issue - State: closed - Opened by win32kdie over 1 year ago - 1 comment

#13 - Hello, I am confused about the location of lr_scheduler. Shouldn’t lr_schduler be placed after a round of epoch? Why is it placed after each batch?

Issue - State: closed - Opened by Lavinci almost 2 years ago

#12 - Is this line necessary?

Issue - State: closed - Opened by Lavinci almost 2 years ago - 3 comments

#11 - import torchkeras报错

Issue - State: closed - Opened by T-Atlas almost 2 years ago - 4 comments

#10 - gpu使用

Issue - State: closed - Opened by husky-happy about 2 years ago - 1 comment

#9 - I think the `torchkeras.torchkeras.Model.forward()` function needs to be slightly modified.

Issue - State: closed - Opened by JnchinZ over 2 years ago

#8 - version of pytorch-lightning?

Issue - State: closed - Opened by shenw8023 over 2 years ago

#7 - is it publish in pypi?

Issue - State: closed - Opened by RandBLRB about 3 years ago

#6 - torchkeras 目前能够有在mac m1芯片上安装的教程吗？目前我是用的环境时 miniforge3

Issue - State: closed - Opened by Batman001 about 3 years ago

#5 - The ‘plot_model' is a very useful and powerful function in Keras, Could be realized some day possibly?

Issue - State: closed - Opened by atlantistin over 3 years ago

#4 - Add EarltStopping and beautify the progressbar to make it more like keras.

Pull Request - State: closed - Opened by laugh12321 over 3 years ago - 2 comments

GitHub / lyhue1991/torchkeras issues and pull requests