MILVLG / imp

a family of highly capabale yet efficient large multimodal models
Apache License 2.0
167 stars 16 forks source link

Steps to reproduce your llama.cpp implementation #22

Open Sowee0 opened 3 months ago

Sowee0 commented 3 months ago

Hello! Reading the project's summarized technical report, it is stated that the imp models were converted to llama.cpp for model comparison. What are the steps to reproduce this conversion? Are there any plans to share the converted gguf models?

Thanks in advance!