-
when i test the base model it gives me 32.5 in blue score, when i do use vanilla-knn it always 0
to build datastore i used this configuration
`CUDA_VISIBLE_DEVICES=1 python $PROJECT_PATH/knnbox-sc…
-
您好,请问您使用v100训练100k大概花了多长时间啊?有没有多卡训练的版本使其能够在3090上快一点训练呢?恕我才疏学浅,还不能解决这个问题,如果您能提供回复,不胜感激!salute!
1、训练命令是:bash ./train_scripts/iwslt_en_de.sh 6 de en
2、显卡占用内存情况显示如下:
[0] NVIDIA GeForce RTX 3090 | 62°C,…
-
Hi,
thank you again for your clean code. Here I have a question about the decoder_nll loss.
According to your code, you calculate the decoder_nll loss in this way:
```
x_start_mean = model.m…
-
### Discussed in https://github.com/facebookresearch/hydra/discussions/2651
Originally posted by **Hannibal046** April 30, 2023
Hi,
Thanks so much for this great projects! I am relatively ne…
-
Hello!
I'm facing this issue while running LNA-ED training command:
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 fairseq-hydra-train --config-dir ${IWSLT_ROOT}/config/ --config-name lna_ed.yaml
Wou…
-
Hi,
I could successfully train a model on IWSLT De-En. Now I have some questions about inference:
1. `decoder_attention_mask` is not in the args. Should I just comment it out? https://github.com…
-
## 🐛 Bug
Hey! I tried following the instructions [here](https://github.com/hlt-mt/FBK-fairseq/blob/master/fbk_works/ALIGNATT_SIMULST_AGENT_INTERSPEECH2023.md) to run the AlignATT agent on the `en->…
-
Hi
Trying to run the dummy waitk example with Python 3.10 gives the following error:
```
Traceback (most recent call last):
File "/home/shared/bhaddow/anaconda3/envs/simul-st-rl/bin/simulev…
-
## ❓ Questions and Help
I want to test finetuning nllb models(3.3B) , I followed the doc in [Finetuning NLLB models](https://github.com/facebookresearch/fairseq/tree/nllb/examples/nllb/modeling)
wit…
-
I tried to use how2 recipe from egs2/how2/asr1. However, when I execute run.sh, an error happened
local/data.sh: error executing wget http://islpc21.is.cs.cmu.edu/ramons/iwslt2019.tar.gz
It seems th…