Shouldn't model.tensors have the LM head like pytorch_model.bin does?

#8
by DrRayStorm - opened

For parity it seems so but maybe I'm missing something.

Same here. I tried to import both model.safetensors and pytorch_model.bin, and it seems that model.safetensors doesn't have the model head:

"Some weights of RobertaModel were not initialized from the model checkpoint at homedata/models/llm_encoders/roberta-base and are newly initialized: ['roberta.pooler.dense.bias', 'roberta.pooler.dense.weight']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference."

no problem with pytorch_model.bin file.

Sign up or log in to comment