-
Hello,
We are using 2.5 phrase-based customized systems and we would like to re-train one of them with an updated TMX. I've noticed that this is possible for neural:
"when you import a TMX file,…
-
Traceback (most recent call last):
File "./mmt", line 610, in
main()
File "./mmt", line 556, in main
actions[command](args)
File "./mmt", line 143, in main_create
builder.buil…
-
I remember having read something about this: is there a command line option to translate a file without live-adaptation, so that the translation goes faster ?
-
Hi @davidecaroselli @nicolabertoldi
We've been consistently seeing an issue with translation against our ZH MMT3.0 engine. I've tried to gather as much info as I can for you.
The corresponding …
-
Hi,
When trying to resume, I get this error:
```
Traceback (most recent call last):
File "/home/lm-dev8/MMT323/ModernMT/cli/mmt/engine.py", line 632, in _build
method(self, args, skip=skip…
-
I finally make a successful installation. However, the example on README.md
./mmt create en it examples/data/train --train-steps 6000
gives the following error. Any idea?
__________________…
-
Waiting for an indication on a tool to monitor the MMT4 training, I wrote my own. Here is the small piece of Java code, using `gnuplot` to draw the learning curves:
![image](https://user-images.githu…
-
I would enjoy to experiment with an option that would do the following:
- randomize the network, especially the word vector layers.
- train the network WITHOUT updating the word vector layers.
Wo…
-
Hello！
I'm trying the latest version of 3.2.3. But I'm meeting a strange problem.
1. ./mmt create \
--learning-rate-decay-start-at 1000000 \
--learning-rate-decay-steps 50000 \
--…
sdlmw updated
5 years ago
-
Suppose I add a pair fr-FR/fr-CA to a memory.
When I'm translating a sentence from fr-FR to fr-CA, it takes a live-tuning pair being twice the SRC segment, thus fr-FR/fr-FR.
I don't know if the pro…