Comments (1)
The default hyper-parameters are the ones we used in the paper and we found that bigger (train) batch sizes yield better results. 512 was basically the biggest batch size we could run.
Decreasing the batch size will probably lead to worse performance.
You basically have two options:
- train on TPUs
- experiment with the
gradient_accumulation_steps
option discussed in the README (it should allow you to split a logical batch into a number of actual batches).
from tapas.
Related Issues (20)
- Inference task for HybridQA dataset
- Inference time HOT 2
- Finding Cell selection confidence
- Training tables with more than 512 cells HOT 1
- how to train the model on hybridqa and ott-qa
- tapas installation is not working HOT 2
- does it support chinese
- Populate float_answer for Tapas Weak supervision for aggregation (WTQ). TypeError: Parameter to CopyFrom() must be instance of same class: expected language.tapas.Question got str. HOT 2
- The code for TableFormer
- Wrong calculation in table HOT 1
- no issue
- Does anyone know of a Tapas tokenizer that is written in Java, C, or Rust?
- create baseline results
- WQT to SQA format conversion script
- TensorFlow No Matching Distribution Error When Installing 'tapas-table-parsing'
- Sudden Increase in Loss value while finetuning
- Generating Pre-Training Data for TAPAS
- Installation Guide Unavailable on Colab Notebook
- Error: Getting Requirements to Build Wheel Failed HOT 1
- ValueError: Too many rows
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 tapas.