Comments (2)
Looks like this issue has been noted here
from albert.
The issue was fixed by doing the following.
- Click "Show Code" on the code cell where parameters (Bucket, Task, and Albert_Model) are filled in
- If you've already run the script once, you'll need to delete the
download_glue_repo
folder. This can be done by adding the line!rm -rf download_glue_repo
right after the# Download glue data.
comment - Instead of cloning the broken repo, clone the fixed repo instead, which can be found here and was mentioned here. This can be done by changing the
!git clone
line to!git clone https://gist.github.com/fef1601580f269eca73bf26a198595f3.git download_glue_repo
- Rerun everything. This time, the dataset should be downloaded correctly.
from albert.
Related Issues (20)
- train albert on custom dataset
- The output of the features of a given text consists of two tensor. Are they both the deep features?
- torch.nn.modules.module.ModuleAttributeError: 'AlbertEmbeddings' object has no attribute 'bias' HOT 1
- The exact English pretraining data and Chinese pretraining data that are exact same to the BERT paper's pretraining data.
- albert base fine-tuned on squad2.0 gets stuck in loop when predicting on new file HOT 1
- Wrong pieces for control symbols after loading SentencepieceProcessor from official model HOT 2
- fine tune on my own English dataset
- Discrepancy in tokenization results using albert's tokenizer and sentencepiece library
- which word segmentation tool is used for pretraining Chinese ALBERT
- Probable error on line 306 in `create_pretraining_data.py` for albert
- Prediction Fails on default Colab HOT 2
- How to get the test embeddings from output of fine-tuned model (tutorial)
- when training in Race , The eval_accuracy is flat , it only has three numbers which are 0.0, 0.33334, 0.66667, 1.0
- Difference between v1 and v2 for xxlarge
- Wrong evaluate result on Squad2.0
- The results can't be reproduced HOT 2
- Improvement to how the `app` and `pages` files conflict is shown. Especially the last log line `"pages/" - "app/"` made it seem like you should remove the `pages` folder altogether. This was a bug in how the `''` case was displayed. After having a look at this I went further and added exactly which file caused the conflict given that `app` allows you to create `app/(home)/page.js` and such it saves some digging for what the actual conflicting file is. Similarly in `pages` both `pages/dashboard/index.js` and `pages/dashboard.js` are possible.
- Load in Browser Tensorflow
- Why do I find inconsistencies between the output of my ALBERT model converted to ONNX format and tested with ONNX Runtime, compared to the original PyTorch format model?
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
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.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from albert.