-
Since the recent updates, it usually requires a 10% more peak memory usage at the point when the program just finished the optimization history loading and starting the remaining train work. Could thi…
-
I used the docker image to train a model, but when I tried to export it I got this error message:
```
Ignoring @/caffe2/caffe2/contrib/aten:aten_op as it is not a valid file.
Traceback (most rece…
-
I run the train.py below
```
python3 train.py $DATA_DIR \
--lr 0.5 --clip-norm 0.1 --dropout 0 --max-tokens 8000 \
--arch fconv_iwslt_de_en \
--save-dir $MODEL_DIR \
…
-
Hello MMT team,
First, wish you all a wonderful and great New Year and more success in 2018.
I came across [this paper](http://www.aclweb.org/anthology/P16-1009) and I thought I should share it …
-
Trying to run fairseq using the following command results in error
$ python train.py data-bin/iwslt14.tokenized.de-en --lr 0.25 --clip-norm 0.1 --dropout 0.2 --max-tokens 4000 - -arch fco…
-
```
from torchtext import data, datasets
EN = data.Field()
FR = data.Field()
train, val, test = datasets.IWSLT.splits(exts=('.en', '.fr'), fields=(EN, FR))
```
gives the following error:
``…
-
When I run train.py, there is an error. What is the problem?The error message is as follows:
| epoch 001: 0%| | 0/820…
-
I'm using pytorch without cuda in MAC OS in python 3.6.
If I change the position of FR and EN, the code works fine.
```python
from torchtext import data
from torchtext import datasets
import…
-
```
File '/home/wen/1.research/zh-en/iwslt/with_transformer/models/transformer.py', line 181, in forward
output = self.dropout(self.w_2(self.relu(self.w_1(x))))
File '/home/wen/anaconda2/lib…
-
The most frequent question on IWSLT for the Neural Moneky was if we have some numbers of how big model can we fit into the GPU with comparison to Nematus.