OSError: It looks like the config file at 'video_chatgpt-7B.bin' is not a valid JSON file.

[yavuzc78]

python scripts/apply_delta.py --base-model-path video_chatgpt-7B.bin --target-model-path LLaVA-Lightning-7B-v1-1 --delta-path liuhaotian/LLaVA-Lightning-7B-delta-v1-1

why Tis issue? OSError: It looks like the config file at 'video_chatgpt-7B.bin' is not a valid JSON file.

[mmaaz60]

Hi @yavuzc78,

Thank you for your interest in our work. The --base-model-path should point out to the LLaMA checkpoints, which can be obtained following the instructions here.

python scripts/apply_delta.py \ 
        --base-model-path <path to LLaMA 7B weights> \
        --target-model-path LLaVA-Lightning-7B-v1-1 \
        --delta-path liuhaotian/LLaVA-Lightning-7B-delta-v1-1

Alternatively you can download the ready LLaVA-Lightening-7B weights from mmaaz60/LLaVA-Lightening-7B-v1-1.

[charchit7]

Hey @mmaaz60 @hanoonaR after downloading the ready LLaVA-Lightening-7B weights from mmaaz60/LLaVA-Lightening-7B-v1-1.

where do we need to put that? path is not mentioned in the readme. If you could please tell me. That would really help.

[mmaaz60]

Hi @charchit7

You can store it anywhere and pass it as --model-name argument while running the demo.

python video_chatgpt/demo/video_demo.py 
        --model-name <path to the LLaVA-Lightening-7B-v1-1 weights prepared in step 3> \
        --projection_path <path to the downloaded video-chatgpt weights>

[charchit7]

Got it. Thank you @mmaaz60 for a prompt reply :)

[charchit7]

Hey @mmaaz60 it works and the whole code base is beautifully written. Thank you!