luofuli / DualRL

A Dual Reinforcement Learning Framework for Unsupervised Text Style Transfer (IJCAI 2019)

https://export.arxiv.org/pdf/1905.10060

luofuli/DualRL Issues

GYAFC dataset
Updated a year ago4
Classifier checkpoint
Updated 2 years ago
How to calculate the accuracy according to multiple references?
Closed 2 years ago
unable to converge
Closed 2 years ago
CalledProcessError: Command returned non-zero exit status 1.
Updated 2 years ago
'DualRL' is not in list
Closed 3 years ago
"ZeroDivisionError: division by zero"
Updated 2 years ago
How to do inference on a single sentence
Updated 3 years ago
Random decode and Greedy decode
Updated 3 years ago
Why BLEU score is always zero?
Updated 3 years ago3
the outputs of all models
Updated 3 years ago1
Restoring of OpenNMT checkpoints
Updated 4 years ago
Which tokenizer is used for preprocessing the data?
Closed 4 years ago2
Question about Baseline Codes
Updated 4 years ago3
N/A
Closed 4 years ago
Style accuracy mismatch 85.6%(Table 1, from the paper) Vs 88.40% (using https://github.com/luofuli/DualRL#step-1-pre-train-classifier and https://github.com/luofuli/DualRL/tree/master/outputs/yelp/DualRL)
Closed 4 years ago1
The original outputs of DualRL
Closed 4 years ago3
How to evaluate on dev set and use early stopping
Closed 4 years ago1
template help
Updated 4 years ago1
pseudo-parallel data for GYAFC
Updated 4 years ago2
关于bleu
Closed 5 years ago2
how to tune hyper-parameters to balance between sentiment accuracy and BLEU score
Updated 5 years ago3
Why did you not use beam search for decoding?
Closed 5 years ago1
Training Duration
Closed 5 years ago1
nmt/model.py
Closed 5 years ago1
nmt/model.py line115 maximum_
Closed 5 years ago
outputs of models on the GYAFC corpus
Closed 5 years ago4
Question about the pseudo-parallel data in the DualRL training stage
Closed 5 years ago5
OOM error
Closed 5 years ago
Wrong Reference Data
Closed 5 years ago1
Can I use the original sentence to initialize the dual_training?
Updated 5 years ago8
How to set the max_output_sequence_length
Closed 5 years ago2
Segmentation fault while training DualRL
Closed 5 years ago1
No code yet.
Closed 5 years ago1