Efficient-Large-Model/VILA1.5-13b weird outputs

Open jhjangjh opened this issue 1 year ago • 3 comments

I am running inference with the Efficient-Large-Model/VILA1.5-13b model.

When using the Efficient-Large-Model/VILA1.5-3b and Efficient-Large-Model/Llama-3-VILA1.5-8B models, the results are generated correctly without any issues.

However, when running inference with the 13B model, I get weird outputs like the ones shown in the image.

Is anyone else experiencing this issue?

Dec 12 '24 11:12 jhjangjh

this usually because you set the wrong conv template. Could you use vicuna_v1 instead?

Dec 19 '24 16:12 Lyken17

I've used both default llava_v0 and vicuna_v1, but the results are the same..

Dec 19 '24 16:12 jhjangjh

hi @jhjangjh , if the issue continues, could you attach your cli / code for reproducement?

Feb 25 '25 09:02 Lyken17