PKU-YuanGroup MoE-LLaVA issues

PKU-YuanGroup / MoE-LLaVA

Mixture-of-Experts for Large Vision-Language Models

https://arxiv.org/abs/2401.15947

Apache License 2.0

1.9k stars 121 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

[Usage] The training always stuck after formatting inputs

#41 detectRecog closed 6 months ago
7
Inference without Deepspeed

#40 aaronnat23 opened 6 months ago
1
[Discussion] Implementation of Qwen1.5 for the project

#39 cydiachen closed 6 months ago
19
inference error in llavamistral

#38 saeedkhaki92 opened 6 months ago
2
Support cuda 12

#37 fishfree opened 6 months ago
1
Eval on MMVET

#36 BeachWang opened 7 months ago
1
Wrong depedancies, why deepspeed dependency for inference, better transformers integration

#35 sujitvasanth opened 7 months ago
3
finetune阶段内存占用太高

#34 awzhgw opened 7 months ago
2
Can i use this to detect events in the video???

#33 Shekharmeena28 opened 7 months ago
1
License Questions

#32 kbrostrom opened 7 months ago
1
第二阶段，loss下降到多少比较合理？

#31 awzhgw opened 7 months ago
1
MoE-LLaVA-StableLM for 4-bits and 8-bit

#30 NikiBase opened 7 months ago
1
panic on finetune

#29 awzhgw closed 7 months ago
3
用自己的数据训练MOE-LLAVA，pretrain阶段,loss下降的非常快

#28 awzhgw closed 7 months ago
4
Reproducing the stage1 and stage2 Model problem on L40s

#27 cydiachen closed 6 months ago
14
traning dataset?

#26 luohao123 closed 7 months ago
14
是否支持启动的时候，指定use_flash_attion_2 ??

#25 awzhgw closed 7 months ago
2
Whether stage-2 pre-train model(llavaphi-2.7b-finetune) is released?

#24 yucheng-zyc opened 7 months ago
4
Wrong cuda allocation

#23 paulgavrikov opened 7 months ago
5
paper和readme指标不一致

#22 sxu1997 closed 7 months ago
2
languageBindVideo model may be hang ?

#21 awzhgw opened 7 months ago
8
当我使用moe-llava的架构集成了mixtral 7BX8的时候，奇怪的事情发生了

#20 awzhgw closed 7 months ago
4
ERROR: Could not build wheels for flash-attn, which is required to install pyproject.toml-based projects

#19 andysingal opened 7 months ago
2
Error in predict.py

#18 KuofengGao closed 7 months ago
1
Moe finetune error

#17 AdonLee072348 closed 7 months ago
12
Do you replicate the weights of the FFNs from stage 1 or stage 2?

#16 simon-lund closed 7 months ago
4
is support video ?

#15 awzhgw opened 7 months ago
3
Hello, I want to know how much GPU memory is needed to run this model.

#14 dforel closed 7 months ago
2
Openchat, quantisation, multiimage

#13 sujitvasanth closed 7 months ago
5
/deepspeed/comm/comm.py", line 341, in all_to_all_single return cdb.all_to_all_single(output=output, AttributeError: 'NoneType' object has no attribute 'all_to_all_single'

#12 lucasjinreal opened 7 months ago
13
image not processed

#11 leosongwei closed 7 months ago
1
Allow custom storage path of the google/siglip-so400m-patch14-384

#10 leosongwei closed 7 months ago
1
Is llavallama moe supported?

#9 DietDietDiet opened 7 months ago
14
can support mixtral 7BX8 model ?

#8 awzhgw closed 7 months ago
1
Update qwen_generation_utils.py

#7 eltociear opened 7 months ago
0
Very bad at language ability

#6 lucasjinreal closed 7 months ago
1
supports Chinese or multiple images?

#5 BaoyanWang closed 7 months ago
3
Images for training

#4 phellonchen closed 7 months ago
2
Method to Replicate Results from Huggingface Spaces

#3 hiroalchem closed 7 months ago
5
Can the author elaborate a bit more about how the stage 3 was achieved?

#2 CanyonWind closed 7 months ago
2
{'loss': 0.0, 'learning_rate': 1.6877637130801689e-07, 'epoch': 0.0}

#1 whalefa1I closed 7 months ago
14