DAMO-NLP-SG / VCD

[CVPR 2024 Highlight] Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding
Apache License 2.0
177 stars 8 forks source link

Request for Parameter and Prompt Configurations Used in Table 3 #20

Open ShanWang-Shan opened 1 week ago

ShanWang-Shan commented 1 week ago

I am writing to inquire about the parameters and prompts used in Table 3. Could you kindly share the configurations and prompts you applied for open-ended generation across each model?

ShanWang-Shan commented 1 week ago

I would also like to ask for the Hugging Face link for the specific version of Qwen-VL that you used.