Jigyasu Krishnan

Results 57 comments of Jigyasu Krishnan

Hi@p-wysocki, I was looking at the #259 and I am trying to understand what's needed to be done here. The gist is that I have to verify this LLM and...

@GStechschulte @tomicapretto I would like to work on this. From the comments above, can I conclude that we need to put a progrss bar around `compute_likelihood_params`?

> @jgyasu thanks for the contribution! I want to check that I understand the logic. You're sub-setting the InferenceData object _before_ doing any computation. Thus, everything that happens later, like...

> Two suggestions > > * Add the option to pass a `random_seed` for reproducibility > * Raise an error if `num_draws is not None and inplace is True`. If...

> @jgyasu thanks for the additional changes. I asked another change, and I also realized we should add a test for this case. Just let me know if you need...

> @jgyasu thanks for the additional changes. I asked another change, and I also realized we should add a test for this case. Just let me know if you need...

> @jgyasu thanks! > > You don't need to create a new class. The one you shared is a parent class used by classes for specific model families. You can...

Another open issue related to `TimeLLM`: #8327