Home
Sign Up
Sign In
V2EX = way to explore
V2EX 是一个关于分享和探索的地方
Sign Up Now
For Existing Member
Sign In
V2EX
›
分享发现
没人关注 DeepSeek R1-0528 吗?
cat9life
·
May 29, 2025
· 2452 views
This topic created in 333 days ago, the information mentioned may be changed or developed.
https://www.zhihu.com/question/1911132833226916938/answer/1911228976870949080
DeepSeek
R1-0528
语言处理
5 replies
•
2025-06-01 14:07:55 +08:00
1
szboy
May 29, 2025
体验了,很厉害:
https://www.zhihu.com/question/1911132833226916938/answer/1911389127636674383
2
cskzhi
May 31, 2025
奇怪了,我这边部署的是原版蒸馏 8B F16 18G 那个版本,中英文分别问了一个 python 脚本,怎么回答都在说梦话?我描述的应该没问题吧,同样的问题给之前的 r1 蒸馏 32B 模型就没问题
3
cskzhi
May 31, 2025
@
cskzhi
更正: 是 15G 的 BF16 版
4
linuslv
Jun 1, 2025
@
cskzhi
#2 有幻觉很正常吧
5
cskzhi
Jun 1, 2025
@
linuslv
幻觉得厉害主要是,之前用 r1 32B 蒸馏挺正常的
About
·
Help
·
Advertise
·
Blog
·
API
·
FAQ
·
Solana
·
3210 Online
Highest 6679
·
Select Language
创意工作者们的社区
World is powered by solitude
VERSION: 3.9.8.5 · 32ms ·
UTC 14:03
·
PVG 22:03
·
LAX 07:03
·
JFK 10:03
♥ Do have faith in what you're doing.
❯