LiGuo12

Question or bug in blip_pretrain.py

1

Between lines 54-75, after "self.text_encoder = BertModel.from_pretrained('bert-base-uncased',config=encoder_config, add_pooling_layer=False)", where "encoder_config" is 'configs/bert_config.json' , the vocal_size is 31090. Next, after "self.text_encoder.resize_token_embeddings(len(self.tokenizer))", the vocal_size changes to 31092. However, "self.text_encoder_m" does not resize...

keypoints labeling problem

5

Hi ryo, We are using your software to complete an annotation on pose estimation. We need to annotate some key points on an image. How can we display multiple key...

Hi, we cannot download the checkpoints, could you please fix it?

sh make.sh issue

3

When I run sh make.sh, it show the following error. My Python version: 3.8.17 PyTorch version: 2.0.1+cu118 CUDA version: 11.8 CUDA_HOME is set correctly Here is the error: running install...

Reproduction Issue – Results Significantly Lower Than Reported

4

Hi, I have attempted to reproduce your model using the code and hyperparameters provided in the repository and paper. However, my results for both stage 1 and stage 2 are...

Hi all, Thank you for your work. I noticed that the provided code does not include evaluation part. To ensure reproducibility of your work, would it be possible to make...