GithubHelp home page GithubHelp logo

Comments (18)

w5688414 avatar w5688414 commented on May 29, 2024 1

可以使用较早版本的paddlenlp,比如paddlenlp2.5.2及其以下

from paddlenlp.

lili-changjiang avatar lili-changjiang commented on May 29, 2024

paddlenlp 2.6.0 2.6.1 2.7.0 2.7.1 2.7.2 都试过也不行

from paddlenlp.

lili-changjiang avatar lili-changjiang commented on May 29, 2024

图片1
图片2

from paddlenlp.

zhaoqf-cq avatar zhaoqf-cq commented on May 29, 2024

我把 paddlepaddle-gpu 回退到 2.4.0rc0, cuda 回退到了10.2 然后 构建封闭域训练样本 就得到了label_maps.json, 我就开始准备训练, 一脸懵逼啊...

python train.py \
    --train_path data/train_data.json \
    --dev_path data/dev_data.json \
    --label_maps_path data/label_maps.json \
    --num_epochs 50 \
    --model_name_or_path ernie-3.0-base-zh

文档给出的训练数据和验证集 哪儿来的啊... 我最开始微调训练用过的吗? 我用训练和验证都指定训练时候用到的 label_studio.json,提示我 TypeError: list indices must be integers or slices, not str

我用 训练时候的 train.txt 和 dev.txt 然后提示我 KeyError: 'entity_list'... 好难啊 QAQ

from paddlenlp.

lili-changjiang avatar lili-changjiang commented on May 29, 2024

我把 paddlepaddle-gpu 回退到 2.4.0rc0, cuda 回退到了10.2 然后 构建封闭域训练样本 就得到了label_maps.json, 我就开始准备训练, 一脸懵逼啊...

python train.py \
    --train_path data/train_data.json \
    --dev_path data/dev_data.json \
    --label_maps_path data/label_maps.json \
    --num_epochs 50 \
    --model_name_or_path ernie-3.0-base-zh

文档给出的训练数据和验证集 哪儿来的啊... 我最开始微调训练用过的吗? 我用训练和验证都指定训练时候用到的 label_studio.json,提示我 TypeError: list indices must be integers or slices, not str

我用 训练时候的 train.txt 和 dev.txt 然后提示我 KeyError: 'entity_list'... 好难啊 QAQ

谢谢提供思路,我试试去

from paddlenlp.

lili-changjiang avatar lili-changjiang commented on May 29, 2024

我把 paddlepaddle-gpu 回退到 2.4.0rc0, cuda 回退到了10.2 然后 构建封闭域训练样本 就得到了label_maps.json, 我就开始准备训练, 一脸懵逼啊...

python train.py \
    --train_path data/train_data.json \
    --dev_path data/dev_data.json \
    --label_maps_path data/label_maps.json \
    --num_epochs 50 \
    --model_name_or_path ernie-3.0-base-zh

文档给出的训练数据和验证集 哪儿来的啊... 我最开始微调训练用过的吗? 我用训练和验证都指定训练时候用到的 label_studio.json,提示我 TypeError: list indices must be integers or slices, not str

我用 训练时候的 train.txt 和 dev.txt 然后提示我 KeyError: 'entity_list'... 好难啊 QAQ

还是不行,python data_convert.py --label_studio_file ../data/label_studio.json --splits 0.7 0.3 0 --task_type ext --layout_analysis True 我执行这个还是报错ImportError: cannot import name 'DataCollatorForClosedDomainIE' from 'paddlenlp.data.data_collator',你用的paddlenlp是哪个版本

from paddlenlp.

zhaoqf-cq avatar zhaoqf-cq commented on May 29, 2024

我把 paddlepaddle-gpu 回退到 2.4.0rc0, cuda 回退到了10.2 然后 构建封闭域训练样本 就得到了label_maps.json, 我就开始准备训练, 一脸懵逼啊...

python train.py \
    --train_path data/train_data.json \
    --dev_path data/dev_data.json \
    --label_maps_path data/label_maps.json \
    --num_epochs 50 \
    --model_name_or_path ernie-3.0-base-zh

文档给出的训练数据和验证集 哪儿来的啊... 我最开始微调训练用过的吗? 我用训练和验证都指定训练时候用到的 label_studio.json,提示我 TypeError: list indices must be integers or slices, not str
我用 训练时候的 train.txt 和 dev.txt 然后提示我 KeyError: 'entity_list'... 好难啊 QAQ

还是不行,python data_convert.py --label_studio_file ../data/label_studio.json --splits 0.7 0.3 0 --task_type ext --layout_analysis True 我执行这个还是报错ImportError: cannot import name 'DataCollatorForClosedDomainIE' from 'paddlenlp.data.data_collator',你用的paddlenlp是哪个版本

我没有使用NLP... 我直接使用的源码, 在ide 里面运行这个 data_convert.py

from paddlenlp.

lili-changjiang avatar lili-changjiang commented on May 29, 2024

我把 paddlepaddle-gpu 回退到 2.4.0rc0, cuda 回退到了10.2 然后 构建封闭域训练样本 就得到了label_maps.json, 我就开始准备训练, 一脸懵逼啊...

python train.py \
    --train_path data/train_data.json \
    --dev_path data/dev_data.json \
    --label_maps_path data/label_maps.json \
    --num_epochs 50 \
    --model_name_or_path ernie-3.0-base-zh

文档给出的训练数据和验证集 哪儿来的啊... 我最开始微调训练用过的吗? 我用训练和验证都指定训练时候用到的 label_studio.json,提示我 TypeError: list indices must be integers or slices, not str
我用 训练时候的 train.txt 和 dev.txt 然后提示我 KeyError: 'entity_list'... 好难啊 QAQ

还是不行,python data_convert.py --label_studio_file ../data/label_studio.json --splits 0.7 0.3 0 --task_type ext --layout_analysis True 我执行这个还是报错ImportError: cannot import name 'DataCollatorForClosedDomainIE' from 'paddlenlp.data.data_collator',你用的paddlenlp是哪个版本

我没有使用NLP... 我直接使用的源码, 在ide 里面运行这个 data_convert.py

我安装的paddlenlp2.5.2,然后把 add-doc-ie里面那个paddlenlp替换我自己的paddlenlp D:\Anaconda\envs\table\Lib\site-packages\paddlenlp,就问题了

from paddlenlp.

lili-changjiang avatar lili-changjiang commented on May 29, 2024

下面几个版本2.3.x 2.4.x都试过,都有问题

from paddlenlp.

zhaoqf-cq avatar zhaoqf-cq commented on May 29, 2024

是不是我们看到的文档都是错误的....
#8074
这里面说的,暂不支持...

from paddlenlp.

w5688414 avatar w5688414 commented on May 29, 2024

我咨询了一下,DataCollatorForClosedDomainIE 这个暂停开发了,欢迎开发者贡献,或者使用大模型的解决方案。

from paddlenlp.

zhaoqf-cq avatar zhaoqf-cq commented on May 29, 2024

我咨询了一下,DataCollatorForClosedDomainIE 这个暂停开发了,欢迎开发者贡献,或者使用大模型的解决方案。

大模型不方便本地化部署, 离线使用不方便...

from paddlenlp.

w5688414 avatar w5688414 commented on May 29, 2024

用显存比较大的显卡,使用fp16或者量化后的模型进行推理部署

from paddlenlp.

zhaoqf-cq avatar zhaoqf-cq commented on May 29, 2024

用显存比较大的显卡,使用fp16或者量化后的模型进行推理部署

训练我试过24GB的3090,还有就是16GB的4080s, 想在CPU进行推理部署,使用CPU进行推理,服务器级别的CPU,单核能力不强 推理不出来结果... 是那种卡半个小时,然后结果输出还是为空

from paddlenlp.

w5688414 avatar w5688414 commented on May 29, 2024

目前UIE技术已经停止开发了,cpu的话,使用更轻量化的模型(精度会下降),但不建议使用cpu推理,另外欢迎开发者贡献。

from paddlenlp.

zhaoqf-cq avatar zhaoqf-cq commented on May 29, 2024

目前UIE技术已经停止开发了,cpu的话,使用更轻量化的模型(精度会下降),但不建议使用cpu推理,另外欢迎开发者贡献。

好吧... 目前还行并没有其他能微调的更小的模型了吧,希望有大佬来贡献一下....

from paddlenlp.

w5688414 avatar w5688414 commented on May 29, 2024

uie-small可以,目前由于人力原因,第三方可以贡献一下

from paddlenlp.

ahahaha-zxp avatar ahahaha-zxp commented on May 29, 2024

请提出你的问题

使用的是 paddlepaddle-gpu 2.5.2.post112 paddlenlp 2.5.2 protobuf 3.20.2 报错:ImportError: cannot import name 'DataCollatorForClosedDomainIE' from 'paddlenlp.data.data_collator' 求解

ac98908819853894e7d22170fa67174
可以试试这种,这个代码是 https://github.com/linjieccc/PaddleNLP/tree/add-doc-ie/ 分支拉取的

from paddlenlp.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.