Gant

implement Google Colab version under standard GPU

1

Under sufficient GPU memory, only use two models T2I and ImageCaption

error in torchtext when excuse <python visual-chatgpt.py>

4

I run it on googlecolab, got python=3.9 and pip=23.0.1, when I run `python visual-chatgpt.py`, I got error like this below. It might be environmental pjromblen. `OSError: /usr/local/lib/python3.9/dist-packages/torchtext/lib/libtorchtext.so: undefined symbol: _ZN2at4_ops5zeros4callEN3c108ArrayRefINS2_6SymIntEEENS2_8optionalINS2_10ScalarTy...

how come is this private?

15

I'm confued about the private, I mean when you download the pretrained llm weights on your local machine, and then use your private data to finetune, and the whole process...

Is there any PEFT like lora fintuning way implemented in MiniGPT-4?

Is there ghost_ResNet pre_trained weight for pytorch?

1

I only find the ghostNet pre_trained weight in model folder.