mirror of
https://github.com/kvcache-ai/ktransformers.git
synced 2026-11-05 01:07:28 +00:00
[GH-ISSUE #212] 500G内存只用了18G #129
Labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
kvcache-ai/ktransformers#129
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @lymanzhao on GitHub (Feb 13, 2025).
Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/212
500G内存
志强CPU,单卡运行,速度还可以8 token每秒,使用了2.5量化模型。
但只有18g内存用到,猜测需要开启内存载入,同样情况我在llama cpp上遇到过,解决了,但似乎对ktransformers无效。
安装软件:ktransformers-0.2.0+cu124torch24avx2-cp311-cp311-linux_x86_64.whl
命令:
python -m ktransformers.local_chat --model_path "/mnt/data/deepseek/DeepSeek-R1/" --gguf_path "/mnt/data/deepseek/DeepSeek-R1-GGUF/DeepSeek-R1-UD-Q2_K_XL/" --force_think true --cpu_infer 60
@murongweibo commented on GitHub (Feb 13, 2025):
你这个用什么gpu?我的报了错
@shuchong commented on GitHub (Feb 13, 2025):
这个报错是不是和卡有关系,我使用v100也有这个报错
@lymanzhao commented on GitHub (Feb 13, 2025):
我是4090,你这个看着是torch的问题,检查是不是版本问题呢
@Xxianna commented on GitHub (Feb 13, 2025):
同,加载deepseek r1 q4km后只有大概15g内存和15g显存占用
但是加载过程产生了巨大的cache
@futureltj commented on GitHub (Feb 13, 2025):
同样的情况,推理速度很慢。有解吗?
@shiontao commented on GitHub (Feb 13, 2025):
一样的,应该是架构太老了的原因
@murongweibo commented on GitHub (Feb 13, 2025):
应该是卡的框架原因,我这里用的谷歌colab里面的分的T4,我看别人说V100也不行
@ipfgao commented on GitHub (Feb 13, 2025):
这样的话,不用32g内存也有希望运行deepseek r1 q4km了?
@lymanzhao commented on GitHub (Feb 13, 2025):
你满足这点速度吗,这是硬盘缓存限制速度
@lymanzhao commented on GitHub (Feb 14, 2025):
top可以看到真实使用情况
@allendred commented on GitHub (Feb 22, 2025):
vmvare虚拟机单cpu6530 cpu 占用30% 内存共1t只占用7g
这样速度
prompt eval count: 9 token(s)
prompt eval duration: 1.7904412746429443s
prompt eval rate: 5.026693769553994 tokens/s
eval count: 734 token(s)
eval duration: 158.81343626976013s
eval rate: 4.621775192580238 tokens/s
@Xxianna commented on GitHub (Feb 22, 2025):
现在好了,768g内存用了740g。好像是两个numa上各保留了权重镜像。速度是快了😋
@allendred commented on GitHub (Feb 22, 2025):
求教怎么解决的
@Xxianna commented on GitHub (Mar 1, 2025):
安装libnuma-dev再重新编译