-
Ben Gyori has now support for expressions in distribution parameters so we can generate multi-level models. We need to properly sort the expressions so they are evaluated in order.
-
Notice tracel-ai from burn framework, this software must substitute to high performance predictions, like robotics, predict from data lake. Some molecular pretrained models use RoBERTa as base model, …
-
Named Entity Recognition with Spacy works great when the language is defined for a given run, since we can specify a single language-specific model. Most interesting applications are multi-lingual, ho…
-
# Brief description
`ionics_fits` provides LaserFlop derived models that are suitable for single-ion crystals. However, the dynamics differen for-multi-ion crystals. Currently multi-ion crystal suppo…
-
### Jan version
v0.5.6
### Describe the Bug
When I select just the 4060 GPU, models load onto the 2070.
When I select just the 2070, models load onto the 4060.
### Steps to Reproduce
_No respons…
-
Hello,
One of the main use cases when deploying real-world models is to train several "expert models" on different subtasks of the main inference task, then using some linear combination of them to…
-
Hi,
I think the type of forecaster @KishManani mentions can be created with two approaches:
1. Using a multi-output regressor (multi-target) from [sklearn.multioutput](https://scikit…
-
### What would you like to see?
These days some of the organizations split their model files (I assume for the easier download), eg. Qwen2.5 is split into 3.5GB parts. I could not import this kind of…
-
## Problem summary
When defining a discriminated type that extends *another* discriminated type, the generated constructors for the downstream types don't set the right values, which results in malfo…
-
Discounting functionality is not actually needed, as we allow time series anyway. So, we can remove code without reducing functionality. (Users should be able to calculate discounting externally. On t…