GenerateU

GenerateU copied to clipboard

Some weights of T5ForConditionalGeneration were not initialized from the model checkpoint at google/flan-t5-base and are newly initialized: ['temp']

1

Thank you very much for your outstanding work. However, when I was training vg_swinT.yaml, I encountered the following issue: Some weights of T5ForConditionalGeneration were not initialized from the model checkpoint...

ClarkQian

Thank you for your wonderful work, I am curious about the details of your training with frozen MLLM experiment, can you release the relevant code?

znsc30

what about the inference code? How to evaluate open-vocabulary detection performance? can you show more details?

sssssshf

['similarity= (100.0 * text_features @ self.lvis_embed.to(self.device)).softmax(dim=-1)'](https://github.com/FoundationVision/GenerateU/blob/main/projects/DDETRS/ddetrs/ddetrs_vl_uni.py#L315) Why do we need to perform matrix multiplication on **text_features** and **self.lvis_embed**? What does self.lvis_embed refer to? I look forward to the author's answer.

nonoevil

When I use pip install -r requirements.txt, the following error occurs. Is there any problem with the requirements.txt? 

nonoevil