Runner diagnostic (team godzilla, DSN 2026)

A tiny randomly initialised HubertForSequenceClassification (36k parameters, 1 layer). It is not a real model: it only checks whether the evaluation runner processes clips at all. It runs 6,659 clips in about 3 minutes on a single CPU thread. No training data was used.

Downloads last month
11
Safetensors
Model size
36.3k params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support