[GH-ISSUE #244] Configuring Latest DeepSeek-V3-Chat-multi-gpu-marlin.yaml throws RuntimeError #145

Closed
opened 2026-10-29 14:34:58 +00:00 by bsakaguchi · 1 comment
Owner

Originally created by @ArYuZzz on GitHub (Feb 14, 2025).
Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/244

The default DeepSeek-V3-Chat-multi-gpu-marlin.yaml in the Release V0.2 works, but the latest yaml in the main branch does't work.

RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cuda:1!
Originally created by @ArYuZzz on GitHub (Feb 14, 2025). Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/244 The default `DeepSeek-V3-Chat-multi-gpu-marlin.yaml` in the `Release V0.2` works, but the latest yaml in the `main branch` does't work. ``` RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cuda:1! ```
Author
Owner

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

https://github.com/kvcache-ai/ktransformers/issues/483

<!-- gh-comment-id:2667907981 --> @ArYuZzz commented on GitHub (Feb 19, 2025): https://github.com/kvcache-ai/ktransformers/issues/483
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#145
No description provided.