[GH-ISSUE #494] 这能在4090跑起来r1 ? 有没有高手真的能跑的 #349

Closed
opened 2026-10-29 14:38:19 +00:00 by bsakaguchi · 4 comments
Owner

Originally created by @652994331 on GitHub (Feb 19, 2025).
Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/494

想问问事deepseek-r1 什么量级的, 确定可以跑吗, 感谢🙏, 如果可以想去搞个机器试试

Originally created by @652994331 on GitHub (Feb 19, 2025). Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/494 想问问事deepseek-r1 什么量级的, 确定可以跑吗, 感谢🙏, 如果可以想去搞个机器试试
Author
Owner

@yansiyu550 commented on GitHub (Feb 19, 2025):

1T DRAM,14GB VRAM,单卡4090就能跑,是Q4量化版本的,不是满血版,CPU要支持AMX

<!-- gh-comment-id:2668095788 --> @yansiyu550 commented on GitHub (Feb 19, 2025): 1T DRAM,14GB VRAM,单卡4090就能跑,是Q4量化版本的,不是满血版,CPU要支持AMX
Author
Owner

@paranoiagu commented on GitHub (Feb 19, 2025):

一点要支持 AMX 吗?

<!-- gh-comment-id:2670043518 --> @paranoiagu commented on GitHub (Feb 19, 2025): 一点要支持 AMX 吗?
Author
Owner

@VinayaLee commented on GitHub (Feb 20, 2025):

一点要支持 AMX 吗?

v0.3.0需要,之前的版本不需要

<!-- gh-comment-id:2670153557 --> @VinayaLee commented on GitHub (Feb 20, 2025): > 一点要支持 AMX 吗? v0.3.0需要,之前的版本不需要
Author
Owner

@moneyisallyouneed commented on GitHub (Feb 23, 2025):

单卡4090+256G DDR5 跑Q2_K_XL,在v0.2.1能到12.7t/s

<!-- gh-comment-id:2676748357 --> @moneyisallyouneed commented on GitHub (Feb 23, 2025): 单卡4090+256G DDR5 跑Q2_K_XL,在v0.2.1能到12.7t/s
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
kvcache-ai/ktransformers#349
No description provided.