pkunlp-icler/FastV
[ECCV 2024] Code for paper: An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models
Python
Issues
- 2
What change did you make at generation time to make "use_cache=False" valid?
#17 opened by Evil-Genius-HY - 0
Why is sys-length equal to 36?
#16 opened by zzzfffsss - 1
FastV Inference and Evaluation
#15 opened by yueeryty - 1
Compatibility with KV-Cache
#13 opened by LukeForeverYoung - 1
Average Attention or Attention on last token?
#14 opened by feiyu12138 - 1
Why llava-hf replaces llava?
#12 opened by zzzfffsss - 3
FileNotFoundError: Couldn't find a dataset script at /cpfs01/shared/public/cl/datasets/OCR-VQA/OCR-VQA.py or any data file in the same directory
#11 opened by zzzfffsss - 3
Qwen没有明显提速
#10 opened by chaos0625 - 3
开源的只是llva请问什么时候可以开源qwen的Demo呢
#4 opened by Cristhine - 1
请问这种方式在训练中适用吗 能不能加速训练呢
#7 opened by weidom - 8
请问当我要评估aokvqa的时候 数据集文件的格式是要什么样的,文件的存放位置是什么样的
#6 opened by Lyn-Lucy - 1
- 4
- 1
paper中的优化耗时问题
#2 opened by JimmyHHua