Xiaodong Gu
Xiaodong Gu
I encountered the same problem! I reverted theano to Oct 20. The first time that the theano was re-installed, it runs OK. but when I re-run the program, it crashes...
The first error seems to be the root cause. Probably because your machine has a small memory to store temporary code vectors. You can try to reduce the chunk size,...
Using beam search can avoid this. We will upload our latest version which uses beam search.
The SWDA dataset has low quality with repetitions which greedy decoding may be sensitive to. However, our final results should be better than what you showed. Did you finish all...
No, they are not
Check whether the `self.encoder_config` is correctly reloaded. Another possible reason: did you set `self.mlm_mode = 'mdn'` during pretraining?
you can use eclipse JDT to parse the AST. The AST contains class names of each method invocation. You can find JDT through maven. There are also python libraries for...
https://github.com/mast-group/codemining-core/blob/master/src/main/java/codemining/java/tokenizers/JavaTokenizer.java
现有代码只支持线性的序列化，不支持嵌套，你可以把你的方括号也看成一个字符，这样就可以当成一个序列了。
@li-car-fei 可以参考 https://github.com/guxd/DialogBERT/blob/master/prepare_data.py 中的binarize函数，把对话(a list of sequences)转成earray数组。