kongds / E5-V

E5-V: Universal Embeddings with Multimodal Large Language Models
https://arxiv.org/abs/2407.12580
163 stars 7 forks source link

Have you tried conducting experiments based on llava-v1.5-7b? #4

Closed Code-kunkun closed 2 months ago

Code-kunkun commented 2 months ago

Have you tried conducting experiments based on llava-v1.5-7b? What were the results like?

kongds commented 2 months ago

We did not fine-tune it in previous experiments, but we tested other models like LLava-v1.6 and phi3 in Appendix C.

Code-kunkun commented 2 months ago

Thanks for your reply!