local demo couldn't detect anything about detgpt HOT 4 OPEN

I had a similar problem, and it turned out that the LLAMA model was the wrong version. Perhaps you can check if the model version is v1.1.

from detgpt.

@shizhediao @hendrydong @pipilurj Can you answer this for me, thank you very much!

from detgpt.

Hi, thanks you very much for your interest! Could you please check whether the model parameters are correctly loaded? Are there any reported errors?

from detgpt.

loaded detector

Loading VIT
Loading VIT Done
Loading Q-Former
Loading Q-Former Done
Loading LLAMA
Loading checkpoint shards: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████| 2/2 [00:52<00:00, 26.08s/it]
Loading LLAMA Done
Load 4 training prompts
Prompt Example
###Human: Describe this image in detail. ###Assistant:
Load BLIP2-LLM Checkpoint: ./output_models/task_tuned_linear_7b.pth
ckpt_path: ./output_models/task_tuned_linear_7b.pth

Initialization Finished

/data2/xingyc/anaconda3/envs/detgpt2/lib/python3.9/site-packages/gradio/outputs.py:43: UserWarning: Usage of gradio.outputs is deprecated, and will not be supported in the future, please import your components from gradio.components
warnings.warn(
/data2/xingyc/anaconda3/envs/detgpt2/lib/python3.9/site-packages/gradio/components.py:164: UserWarning: Unknown style parameter: full_width
warnings.warn(f"Unknown style parameter: {key}")
/data2/xingyc/anaconda3/envs/detgpt2/lib/python3.9/site-packages/gradio/components.py:164: UserWarning: Unknown style parameter: full_height
warnings.warn(f"Unknown style parameter: {key}")
Running on local URL: http://127.0.0.1:7860
system prompt: You must strictly answer the question step by step:
Step-1. describe the given image in detail.
Step-2. find all the objects related to user input, and concisely explain why these objects meet the requirement.
Step-3. list out all related objects existing in the image strictly as follows: <Therefore the answer is: [object_names]>.
Complete all 3 steps as detailed as possible.
You must finish the answer with complete sentences.
I want to have a cold beverage.

From run grounding, oringinal response message

No match found.

/data2/xingyc/anaconda3/envs/detgpt2/lib/python3.9/site-packages/transformers/modeling_utils.py:862: FutureWarning: The device argument is deprecated and will be removed in v5 of Transformers.
warnings.warn(
/data2/xingyc/anaconda3/envs/detgpt2/lib/python3.9/site-packages/torch/utils/checkpoint.py:25: UserWarning: None of the inputs have requires_grad=True. Gradients will be None
warnings.warn("None of the inputs have requires_grad=True. Gradients will be None")

Detector predicted phrases [] @pipilurj

from detgpt.