jsedoc

Yes, indeed, I've finally gotten to the caching problem. I am currently just copying this to a tensor in a circular buffer style manner. Is there a better way?

@jaseweston for HRED there's already changes made in Julian's fork of ParlAI (https://github.com/julianser/ParlAI).

@andreamad8 can you post the responses that you got? ... and how you got them. We can't seem to match the ones reported.

Thanks!!! In the paper, it says that a response was chosen from 10 responses in top-k. This is always the problem with sampling that reproducibility becomes an issue. Especially when...

the dict file can be found [here](https://drive.google.com/file/d/12MM8NBYmjDnac3XxZq8zUkoCqF15Y5ob/view?usp=sharing). For the dataset please email the authors (" Building End-To-End Dialogue Systems Using Generative Hierarchical Neural Network Models" by Serban et al (currently...

Please ask the original authors for the dataset.

the dict file can be found [here](https://drive.google.com/file/d/12MM8NBYmjDnac3XxZq8zUkoCqF15Y5ob/view?usp=sharing). The eval source file is available [here](https://raw.githubusercontent.com/chateval/application/master/chateval/assets/prompts/ncm.txt) from chateval.org. Please note that this eval set is from the paper Neural Conversation Model (Vinyals...

For the dataset please email the authors (" Building End-To-End Dialogue Systems Using Generative Hierarchical Neural Network Models" by Serban et al (currently on the MovieTriples dataset). This is dataset...

Tagging @kelvinguu for some help.