HaozheZhao / MIC

MMICL, a state-of-the-art VLM with the in context learning ability from ICL, PKU

HaozheZhao/MIC Issues

推理所需的显卡
Updated 5 months ago1
multiple gpus loading
Updated 5 months ago3
Environment Issues
Updated 7 months ago
sp_token=32110
Updated 8 months ago
Bad inference output
Updated 9 months ago
multi-image inference
Updated 9 months ago
Online Demo service is not available
Updated 9 months ago1
Do you commit the latest code?
Updated 9 months ago
Composition of MIC_5M
Closed 10 months ago1
原始数据组织
Closed a year ago11
finetune problem
Updated 10 months ago8
训练数据处理:mic_sampled
Updated 10 months ago3
Winoground Eval
Closed 10 months ago3
复现问题
Closed 10 months ago3
How to achieve multi-turn conversation
Closed 10 months ago1
the generation code of train dataset.
Closed 10 months ago4
How to finetune MMICL on my own datasets
Closed 10 months ago6
How to train the model using MIC_sampled?
Updated 10 months ago1
Batch inference support?
Closed a year ago2
MME和MMBench上的评测代码
Closed a year ago2
failed in datasets.load_dataset
Updated a year ago1
Code for VQAv2 evaluation
Closed a year ago11
Usage of the replace_token
Closed a year ago2
wiil you release the checkpoint?
Closed a year ago10
Bugs during inference
Closed a year ago1
Will you release the training data?
Closed a year ago1
训练数据处理过程
Closed a year ago1
论文在哪里可以找到，搜了一圈没找到
Closed a year ago2
Training Error : AttributeError: 'InstructBlipConfig' object has no attribute 'hidden_size'
Closed a year ago5
想请教一下 infer 文件里面的 config 变量具体是什么值呢？
Closed a year ago1