cmathx
cmathx
build-interpolate-lm.sh: It generates interpolation n-gram models from multi-domain original corpus. In the pipeline, I only see n-gram model interpolation from renormModelHdfs which is n-gram model format you defined. How can...
讲道理,大部分都没有想去外国,还是想留在国内,毕竟家人都在国内,还有文化什么的,去了国外还要适应。
I want to train huge data such as 10TB. As reading file in single-thread manner, it's slow. How can kenlm support reading file in multi-thread manner?
I use a machine with 64GB memory and it has only 12GB memory. I run a test with tagger->parse() and it costs long time about 150ms. But when I run...
I'm a bit confuse to num_step. But when training the model, num_step is set to 30. So in lstm model, num_step is 30. Why you set num_step to 1 in...
How long does model(gatednetfvLF-128k-1024-80-0002-300iter-norelu-basic-gatedmoe) takes if use only one GPU?
Hi, kimiyoung. Your transformer-xl idea is interesting. I have some confusion about the git resp. In pytorch and tensorflow branch, 1-billion experiment parameters seems have some difference. First, mem_len and...
This neural network seems art. Waiting for more experiments in NLP tasks.