ali-vosoughi/oscar-llava-v1.5-7b-projector

OSCaR LLaVA v1.5 7B Projector

This repository contains the projector artifact staged for the OSCaR public release.

Artifact Type

• Local staging directory: llava-v1.5-7b-pretrain-projector
• Public repo id: ali-vosoughi/oscar-llava-v1.5-7b-projector
• Training data condition: projector pretraining assets used before OSCaR LoRA fine-tuning

Files

• config.json
• mm_projector.bin

Loading

This is a projector-only release. It is intended for the pretraining and fine-tuning workflow documented in the OSCaR code repository.

Example:

bash scripts/train/pretrain_v1_5_13b_projector.sh

Training Configuration

• LLaVA v1.5 stack
• CLIP ViT-L/336 vision tower
• LoRA rank 128
• LoRA alpha 256
• learning rate 2e-4
• 1 epoch
• max sequence length 2048

Related Resources

• Code: https://github.com/nguyennm1024/OSCaR
• Dataset: https://huggingface.co/datasets/ali-vosoughi/oscar-dataset
• Paper: https://arxiv.org/abs/2402.17128