Dayiheng Liu
Losin94
AI & ML interests
None yet
Recent Activity
authored
a paper
6 days ago
How Abilities in Large Language Models are Affected by Supervised
Fine-tuning Data Composition
authored
a paper
6 days ago
ProphetNet: Predicting Future N-gram for Sequence-to-Sequence
Pre-training
authored
a paper
6 days ago
OccuQuest: Mitigating Occupational Bias for Inclusive Large Language
Models
Organizations
Losin94's activity
How to transform the existing 1.8B into Qwen1.5-MoE-A2.7B?
2
#1 opened 9 months ago
by
wnma3mz
貌似很拉跨，一个7B的模型3090显存都不够载入，要是不安装它推荐的加速包，速度慢的像狗。
15
#12 opened over 1 year ago
by
boxter007