Comments (3)
atulkum
commented on September 7, 2024
The reason is we want to keep "beam size == batch size". at the start every row in a beam has a <s> token.
And then we pick the output from the beam having highest score.
from pointer_summarizer.
tim5go
commented on September 7, 2024
I also observed if beam_size = 4 and batch_size = 32, it will throw dimension mismatch error here, is it the reason you mentioned?
It is quite weird I have to set them identical.
from pointer_summarizer.
atulkum
commented on September 7, 2024
To make the code simpler we run 1 beam per batch. You can code multiple beam per batch too, just need to process every beam separately.
from pointer_summarizer.
Related Issues (20)
- Python3 support? HOT 13
- During the training and verification process, when "step = 0", the "coverage" is initialized differently. During training, the coverage is an all-zero tensor, but this is not the case during prediction. HOT 1
- url correction HOT 1
- What is the version of tensorflow？ HOT 1
- Test time custom decoding!!
- Training saturates early? HOT 3
- Vector encode input extend vocab HOT 3
- what's function of the eval.py when i check the train.py ,it does't call the eval.py , save the model directly? HOT 1
- question about eval HOT 1
- eval.py decode.py HOT 2
- when i train it with coverage ,the loss is nan when i get 250k iter? HOT 1
- how to use valid dataset to select a bestmodel to test? HOT 1
- How to train with Coverage? HOT 1
- 'Encoder' object has no attribute 'tx_proj' HOT 3
- What is the specific implementation of pointer network HOT 1
- Can the code here be trained with multiple GPUs HOT 1
- Discrepancy with implementation and the paper
- Retraining model cause optimizer duplicate parameter error HOT 3
- How to choose the best training model
- Duplicated computation with LSTM?
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 pointer_summarizer.