Akshaya Shanbhogue
Akshaya Shanbhogue
Specifically, the existing code is treats some variables as variance while the variables are referred to as sigma, which is the standard notation of standard deviation. Fixing the variable names...
I was following the [reparameterization code](https://github.com/MilaNLProc/contextualized-topic-models/blob/master/contextualized_topic_models/networks/decoding_network.py#L117). During [invocation](https://github.com/MilaNLProc/contextualized-topic-models/blob/master/contextualized_topic_models/networks/decoding_network.py#L128C61-L128C80), the variable [`posterior_log_sigma`](https://github.com/MilaNLProc/contextualized-topic-models/blob/master/contextualized_topic_models/networks/decoding_network.py#L124C23-L124C42) is incorrectly named. It's supposed to be `posterior_log_variance`. Similarly, the [inf net](https://github.com/MilaNLProc/contextualized-topic-models/blob/master/contextualized_topic_models/networks/inference_network.py#L52-L53) needs to be renamed to variance instead...