YiyangZhou/POVID
[Arxiv] Aligning Modalities in Vision Large Language Models via Preference Fine-tuning
PythonApache-2.0
Issues
- 0
Loading adapter weights from /root/autodl-tmp/POVID/checkpoint/output/POVID_stage_one_LoRa_bs2 led to unexpected keys not found in the model:
#13 opened by phkhub - 4
- 0
Ref model is not used!
#10 opened by shipengai - 1
您实现的dpo和原始dpo不一样。
#11 opened by shipengai - 4
AssertionError: The parameter 447 has already been reduced. Gradient computed twice for this partition. Multiple gradient reduction is currently not supported
#6 opened by zkysss11235 - 3
Visualize Attention Map
#4 opened by franciscoliu - 1
About the evaluation code
#7 opened by takomc - 0
About Chair evaluation
#9 opened by itsqyh - 0
- 2
About CHAIR evaluation
#5 opened by Kizna1ver - 1
Result for Pope not reproduced [Issue]
#2 opened by rohan598 - 4
dependencies in `pyproject.toml`
#1 opened by elmiraloo - 1
the first loss is not exactly DPO loss
#3 opened by elmiraloo