how do we pass multiple indices as input to generate multiple outputs in word_language model

[kmario23]

The current codebase of word_language_model/generate.py uses a single (randomly sampled) index as input and generates a text based on this.

Now, I'd like to extend this a bit and would like to pass a set of indices (i.e. > 1) as input and be able to generate a set of texts as output. I tried it with a simple loop based approach of iteratively querying the model but it's taking hours to do this task, since it has to be done sequentially.

Any ideas about how to pass in a list of indices as input, particularly in the line: word_language_model/generate.py#L56 ? This can be called as batchified generate function!