Kumo Tabular medium regressor

NVIDIA's Kumo Tabular medium regressor (medium/regressor.pt from nvidia/Kumo-Tabular at revision 3c3e10bbdb590ace29e7026847f92db3c603096d), converted to safetensors for gfi. The weights are unchanged. The network emits 999 quantiles.

Files

model.safetensors: weights, under the parameter names of NVIDIA's PyTorch model.
config.json: architecture_name and config, the model's sizes as named in NVIDIA's structured-data-models (MODEL_KWARGS["medium"]).

License

OpenMDW 1.1, as for the original Kumo Tabular weights. NVIDIA's copyright notice and the license are kept in LICENSE.

Source

NVIDIA structured-data-models

Downloads last month: 20

Safetensors

Model size

62.5M params

Tensor type

F32

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for general-forecast/kumo-tabular-medium-regressor

Base model

nvidia/Kumo-Tabular

Finetuned

(6)

this model