zeroentropy/zerank-1-small · Missing weights in architecture

When initialising the model using the instructions provided in model card, I get:

Some weights of Qwen3ForSequenceClassification were not initialized from the model checkpoint at zeroentropy/zerank-1-small and are newly initialized: ['score.weight']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.

This tells that some weights from the original Qwen model were not used in your model : maybe it's not a bug and they're not needed, but in that case, it's generally still advised to fill weights with dummy values, if not too large, or to change the original architecture if it's not suited.