Zhiyuan Zhao

CPU overload

2

When running program CPU is heavily overload which is not gonna happen on MMDETECTION(in my experience)?

Thanks for your wonderful work! Do you guys test on ScienceQA? And how does Mini-GPT4 perform?

[Usage] Token indices sequence length is longer than the specified maximum sequence length for this model (3278 > 2048). Running this sequence through the model will result in indexing errors

3

### Describe the issue Issue: token length too long. Will this affect model performance? Command: ``` bash scripts/v1_5/finetune_lora.sh ``` Log: ``` Token indices sequence length is longer than the specified...

how to split the whole dataset into each gpu? when multi-gpu training I find that each gpu has to go over the whole dataset, while some other repos usually split...

How to perform multi-task training?

2

Hi, thanks for your great work! Could you please tell me how to perform multi-task training? It seems how to do multi-task training is not specified.\ Thank you very much!

Question about intrisic rank in Figure3(a)

10

Hello, thanks for your exellent work! Could you please share the code about how you calculate the intrisic rank in Figure3(a)? I cannot achieve similar results.

How to change the model weights in safetensors?

1

### Feature request For example, I want to change some weight with shape [K,K,C] into [K,K,C/2], how can I achieve this hacking? ### Motivation N/A ### Your contribution N/A