github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #11866] Starting qwen3-coder:30b-a3b-q4_K_M failed #7877

New Issue
Closed
opened 2026-04-12 20:02:04 -05:00 by GiteaMirror · 5 comments
GiteaMirror commented 2026-04-12 20:02:04 -05:00
Owner
Copy Link

Originally created by @EveningQi on GitHub (Aug 12, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11866

What is the issue?

Starting qwen3-coder:30b-a3b-q4_K_M failed, with a prompt of insufficient VRAM. The graphics card is NVIDIA 4090D with 24G of VRAM. The occupancy rate is less than 5%, so why is there a prompt of insufficient VRAM?

Image Image Image

Relevant log output

alloc_tensor_range: failed to initialize tensor output.weight
llama_model_load: error loading model: unable to allocate CUDA0 buffer
llama_model_load_from_file_impl: failed to load model
Aug 12 17:59:48 ps ollama[1343661]: time=2025-08-12T17:59:48.932+08:00 level=WARN source=sched.go:685 msg="gpu VRAM usage didn't recover within timeout" seconds=5.538058371 runner.size="18.9 GiB" runner.vram="18.9 GiB" runner.parallel=1 runner.pid=1349054 runner.model=/usr/share/ollama/.ollama/models/blobs/sha256-e9183b5c18a0cf736578c1e3d1cbd4b7e98e3ad3be6176b68c20f156d54a07ac

OS

Linux

GPU

Nvidia

CPU

AMD

Ollama version

0.11.4

Originally created by @EveningQi on GitHub (Aug 12, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11866 ### What is the issue? Starting qwen3-coder:30b-a3b-q4_K_M failed, with a prompt of insufficient VRAM. The graphics card is NVIDIA 4090D with 24G of VRAM. The occupancy rate is less than 5%, so why is there a prompt of insufficient VRAM? <img width="1906" height="767" alt="Image" src="https://github.com/user-attachments/assets/9efbec8f-fe5e-4a78-a88c-3fccc0a1dbb0" /> <img width="1058" height="566" alt="Image" src="https://github.com/user-attachments/assets/586c68ea-f55d-4ab2-a8fc-87ed24f9fedd" /> <img width="922" height="566" alt="Image" src="https://github.com/user-attachments/assets/e76f1984-4d90-49d9-8cfb-f6ee8bb07f4d" /> ### Relevant log output ```shell alloc_tensor_range: failed to initialize tensor output.weight llama_model_load: error loading model: unable to allocate CUDA0 buffer llama_model_load_from_file_impl: failed to load model Aug 12 17:59:48 ps ollama[1343661]: time=2025-08-12T17:59:48.932+08:00 level=WARN source=sched.go:685 msg="gpu VRAM usage didn't recover within timeout" seconds=5.538058371 runner.size="18.9 GiB" runner.vram="18.9 GiB" runner.parallel=1 runner.pid=1349054 runner.model=/usr/share/ollama/.ollama/models/blobs/sha256-e9183b5c18a0cf736578c1e3d1cbd4b7e98e3ad3be6176b68c20f156d54a07ac ``` ### OS Linux ### GPU Nvidia ### CPU AMD ### Ollama version 0.11.4
GiteaMirror added the bug label 2026-04-12 20:02:04 -05:00
GiteaMirror commented 2026-04-12 20:02:05 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Aug 12, 2025):

Full text server log will help in debugging.

<!-- gh-comment-id:3179383117 --> @rick-github commented on GitHub (Aug 12, 2025): Full text server log will help in debugging.
GiteaMirror commented 2026-04-12 20:02:05 -05:00
Author
Owner
Copy Link

@EveningQi commented on GitHub (Aug 13, 2025):

全文服务器日志将有助于调试。
This is the latest log output from running qwen3:30b-a3b-q4_K_M. Could a master take a look at what the cause is

Image

log.txt

<!-- gh-comment-id:3182051598 --> @EveningQi commented on GitHub (Aug 13, 2025): > 全文服务器日志将有助于调试。 This is the latest log output from running qwen3:30b-a3b-q4_K_M. Could a master take a look at what the cause is <img width="993" height="519" alt="Image" src="https://github.com/user-attachments/assets/0fde889e-6196-4dca-8b31-988d0028c299" /> [log.txt](https://github.com/user-attachments/files/21746559/log.txt)
GiteaMirror commented 2026-04-12 20:02:06 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Aug 13, 2025):

Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/libggml-cuda.so
Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/cuda_v12/libggml-cuda.so

The CUDA library is being loaded from two different locations. #11211. Delete /usr/lib/ollama/cuda_v*.

<!-- gh-comment-id:3184251844 --> @rick-github commented on GitHub (Aug 13, 2025): ``` Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/libggml-cuda.so Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/cuda_v12/libggml-cuda.so ``` The CUDA library is being loaded from two different locations. #11211. Delete `/usr/lib/ollama/cuda_v*`.
GiteaMirror commented 2026-04-12 20:02:06 -05:00
Author
Owner
Copy Link

@EveningQi commented on GitHub (Aug 13, 2025):

Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/libggml-cuda.so
Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/cuda_v12/libggml-cuda.so

CUDA 库正在从两个不同的位置加载。 #11211。删除/usr/lib/ollama/cuda_v*。

Thank you, thank you. I'll delete it tomorrow and try to run it again

<!-- gh-comment-id:3184343261 --> @EveningQi commented on GitHub (Aug 13, 2025): > ``` > Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/libggml-cuda.so > Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/cuda_v12/libggml-cuda.so > ``` > > CUDA 库正在从两个不同的位置加载。 [#11211](https://github.com/ollama/ollama/issues/11211)。删除`/usr/lib/ollama/cuda_v*`。 Thank you, thank you. I'll delete it tomorrow and try to run it again
GiteaMirror commented 2026-04-12 20:02:06 -05:00
Author
Owner
Copy Link

@EveningQi commented on GitHub (Aug 14, 2025):

Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/libggml-cuda.so
Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/cuda_v12/libggml-cuda.so

CUDA 库正在从两个不同的位置加载。 #11211。删除/usr/lib/ollama/cuda_v*。
Thank you, thank you. The model can now be successfully run

<!-- gh-comment-id:3186383169 --> @EveningQi commented on GitHub (Aug 14, 2025): > ``` > Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/libggml-cuda.so > Aug 13 11:14:18 ps ollama[1437584]: load_backend: loaded CUDA backend from /usr/lib/ollama/cuda_v12/libggml-cuda.so > ``` > > CUDA 库正在从两个不同的位置加载。 [#11211](https://github.com/ollama/ollama/issues/11211)。删除`/usr/lib/ollama/cuda_v*`。 Thank you, thank you. The model can now be successfully run
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#7877
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?