Weichen Xu comments

Results 5 comments of


                                            Weichen Xu

I have high acc rate but the meanIU is too low,and the prediction result is bad.

> I use the minidataset in `image-segmentation-keras/test/example_dataset` , my acc is: > > ``` > 2020-03-03 18:11:22.262469: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudnn.so.7 > 512/512 [==============================] - 109s 212ms/step...

How can i get pretrained weights?

Is it correct to keep using adapter_kv_cache during training in litgpt/adapter.py?

@Andrei-Aksionov @carmocca Can you answer my doubts? I would greatly appreciate it

[img2llm_vqa] LAVIS requires transformers>=4.28.0 but CLIP doesn't support transformers>=4.27

I have the same issue, Lavis requires version>4.28, but there are the following errors: AssertionError: BLIP models are not compatible with transformers>=4.27, run pip install transformers==4.25 to downgrade

[img2llm_vqa] LAVIS requires transformers>=4.28.0 but CLIP doesn't support transformers>=4.27

@kopyl Could you please help me solve it? Thank you!