Comments (5)
Problém teďkon je v tom, že training loop potřebuje všechny *kodéry, aby vytvořil feed_dicty. Co kdybychom převedly zodpovědnost za vytváření enkodérských feed_dictů na decoder? Vlastně by to kopírovalo strom, ve kterým jsou ty objekty pospojovaný v konfiguraci. Každej objekt by měl zodpovědnost za vytváření feed_dictů svých synů (nebo otců jestli koukáš odspoda nebo odshora)
from neuralmonkey.
Ten training loop je vůbec problematická entita, nebylo by lepší s ním nejdřív něco udělat, než se pustíme tady do toho?
from neuralmonkey.
Právě na tom dělám, to bylo zdrojem toho nápadu, kterej jsem sem připsal.
from neuralmonkey.
Já bych to nepřeváděl na dekodér (ale je to spíš otázka vkusu). Důvod je, že můžeme mít víc dekodérů, který budou sdílet některý z enkodérů a asi bysme se chtěli vyhnout tomu, aby se tam feedovalo dvakrát. Potom by se muselo kontrolovat, jestli už není nafeedováno. Přišlo by mi hezčí všechno, co je "feedable" dát do setu a pak to nakrmit.
from neuralmonkey.
Vyřešeno s #151.
from neuralmonkey.
Related Issues (20)
- max_length must be undefined when using SequenceLabeler HOT 1
- learning_utils.join_exection_results does not support OutputSeries == dict
- from_dataset does not exist HOT 2
- dataset.from_files HOT 1
- Confusing Exception message in dataset.py
- Add an output buffering for neuralmonkey-run HOT 2
- Unify dropout usage
- GreedyRunner should not fetch training-related tensors by default during inference HOT 4
- Running as server is broken HOT 1
- neuralmonkey-run with RNN model does not work without reference HOT 1
- Dataset series should support max_len (max_size) flag.
- Neural Monkey does not throw exception when main.initial_variables contains nonexistent path.
- Neural Monkey should throw Exception when tf.Saver.restore fails
- How to train a transformer model with multi-source encoders ? HOT 3
- Exception: Unexpected fields: runners_batch_size HOT 1
- Requirement of editing the post-edit.ini and translation.ini for APE and MT respectively HOT 1
- The Model Configuration in Machine Translation task HOT 1
- some questions about multi-source based transformer model HOT 2
- Did you mean file './dataset.load_dataset_from_files'? HOT 1
- Interested in your paper HOT 1
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 neuralmonkey.