Oktai Tatanov

Curiosity-Driven Exploration

2

What is SOTA in Curiosity-Driven Exploration currently? Will you plan implement that or simple version?

REINFORCE

2

What about REINFORCE algorithm?

What the batch size was used in paper?

7

@jik876 I see batch_size=16 in config, but I want to clarify that batch size was equaled to 16 per GPU, right? And you used 2 V100 for training with this...

beam search

2

According to README: > Currently, we export components (encoder, decoder) to Caffe2 separately and beam search is implemented in C++. In the near future, we will be able to export...

TalkNet2 original code

4

Hello! You can find original code for TalkNet2 here: https://github.com/NVIDIA/NeMo/blob/main/nemo/collections/tts/models/talknet.py Also take a look here: https://github.com/NVIDIA/NeMo/blob/main/tutorials/tts/3_TTS_TalkNet_Training.ipynb Just for simplify re-implementation :)

Default argument value is mutable

6

Hello, @wengong-jin ! My IDE prevents me that default argument value is mutable (```MolTreeNode``` class). I want to make sure you understand that default argument values are really mutable and...

AUCROC in paper

1

Hi, @jeremykawahara, @hamarneh! Can you explain how we calculate AUCROC for diagnosis? AUCROC is for binary classification, how can we interpret this for multi-class classification (because 1 diagnosis per image)?...

Hi, @soumith! You wrote: https://github.com/soumith/ganhacks/blob/master/README.md#11-dont-balance-loss-via-statistics-unless-you-have-a-good-reason-to One has new experiments: https://twitter.com/theshawwn/status/1260476167053869057

Some bugs in English, German, Spanish, Italian normalizers

1

Hi! I found a bug in English normalization. The following code is applied: ```python normalizer = Normalizer( input_case="cased", lang="en", deterministic=True, ) norm_text = normalizer.normalize(text, punct_post_process=True) ``` text=`Here is mail.nasa.gov.` norm_text=`Here...