Saransh Surana

NF4 works a bit differently from uniform quantization since it uses a fixed non-uniform codebook rather than evenly spaced steps. A simple way to implement it is to precompute the...

Thanks for the suggestion. I can work on extending GEPA to optimize field descriptions as well, including recursive handling for Pydantic models. Would you like this to live in the...

Thanks for raising this. I can help expand the steerable model tests to include multiple model initializations, inference runs across different backends, and repeated runs to detect flaky generation. Would...

Update based on review feedback: 1. Moved the parametrized test into the Transformers backend test suite at _tests/models/test_transformers_parametrized.py_, since this test targets the _Transformers_ model implementation specifically. 2. Kept the...

Thanks for raising this. I can look into drafting a shared test suite so each integration just specifies model-specific params (name, API URL, key env var). Would you prefer this...

Thanks for the detailed traceback. This error occurs when the `color` variable has only one unique value. In that case, `seaborn.objects` treats it as continuous, but the scale domain collapses...

I ran a test with 5-minute interval data and can confirm that seasonal_decompose works as expected if you manually specify the correct `period`. In this case, a full day has...

@germanblanco That makes total sense. I hadn’t thought about making the expected seasonality a clear parameter. Thanks for pointing that out. This would be my first contribution here so I’d...

Just to confirm, the idea is to verify that the rotated Clayton copulas (especially 180° for upper-tail dependence) behave as expected and possibly add tests or examples, right? I can...

Got it, so the idea is to start with a helper that takes (u1, u2, rotation, base_copula) and applies the CDF rotation (0°, 90°, 180°, 270°) as defined in the...