github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-08 00:51:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #12821] gemma3:27b|gemma3:27b-it-qat failed in ollama 0.12.7-rc0 #34256

New Issue
Closed
opened 2026-04-22 17:41:48 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-04-22 17:41:48 -05:00
Owner
Copy Link

Originally created by @jay875037671 on GitHub (Oct 29, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/12821

What is the issue?

those 2 models workd well in 0.12.6(but 0.12.6 qwen3-vl failed though lol)
crash on: compiler=cgo(clang) ggml.c:1921: GGML_ASSERT(ggml_can_repeat(b, a)) failed

logfile:

[GIN] 2025/10/29 - 16:22:50 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/10/29 - 16:22:50 | 200 | 8.5009ms | 127.0.0.1 | GET "/api/tags" [GIN] 2025/10/29 - 16:22:59 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/10/29 - 16:22:59 | 200 | 72.9136ms | 127.0.0.1 | POST "/api/show" ggml_backend_cuda_device_get_memory device GPU-2fb823f9-0c5c-9291-49c5-2685c7450e9a utilizing NVML memory reporting free: 20149010432 total: 34190917632 time=2025-10-29T16:22:59.298+08:00 level=INFO source=cpu_windows.go:139 msg=packages count=1 time=2025-10-29T16:22:59.298+08:00 level=INFO source=cpu_windows.go:186 msg="" package=0 cores=8 efficiency=0 threads=16 time=2025-10-29T16:22:59.322+08:00 level=INFO source=sched.go:559 msg="updated VRAM based on existing loaded models" gpu=GPU-2fb823f9-0c5c-9291-49c5-2685c7450e9a library=CUDA total="31.8 GiB" available="18.8 GiB" time=2025-10-29T16:22:59.399+08:00 level=INFO source=server.go:215 msg="enabling flash attention" time=2025-10-29T16:22:59.422+08:00 level=INFO source=server.go:385 msg="starting runner" cmd="C:\Users\jayli\AppData\Local\Programs\Ollama\ollama.exe runner --ollama-engine --model G:\ollama\models\blobs\sha256-e796792eba26c4d3b04b0ac5adb01a453dd9ec2dfd83b6c59cbf6fe5f30b0f68 --port 7761" time=2025-10-29T16:22:59.435+08:00 level=INFO source=server.go:638 msg="loading model" "model layers"=63 requested=-1 time=2025-10-29T16:22:59.435+08:00 level=INFO source=server.go:643 msg="system memory" total="63.6 GiB" free="46.1 GiB" free_swap="95.0 GiB" time=2025-10-29T16:22:59.435+08:00 level=INFO source=server.go:650 msg="gpu memory" id=GPU-2fb823f9-0c5c-9291-49c5-2685c7450e9a library=CUDA available="18.3 GiB" free="18.8 GiB" minimum="457.0 MiB" overhead="0 B" time=2025-10-29T16:22:59.478+08:00 level=INFO source=runner.go:1337 msg="starting ollama engine" time=2025-10-29T16:22:59.498+08:00 level=INFO source=runner.go:1372 msg="Server listening on 127.0.0.1:7761" time=2025-10-29T16:22:59.499+08:00 level=INFO source=runner.go:1210 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:true KvSize:4096 KvCacheType: NumThreads:8 GPULayers:63[ID:GPU-2fb823f9-0c5c-9291-49c5-2685c7450e9a Layers:63(0..62)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2025-10-29T16:22:59.540+08:00 level=INFO source=ggml.go:135 msg="" architecture=gemma3 file_type=Q4_K_M name="" description="" num_tensors=1247 num_key_values=37 load_backend: loaded CPU backend from C:\Users\jayli\AppData\Local\Programs\Ollama\lib\ollama\ggml-cpu-icelake.dll ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no ggml_cuda_init: found 1 CUDA devices: Device 0: NVIDIA GeForce RTX 5090 D, compute capability 12.0, VMM: yes, ID: GPU-2fb823f9-0c5c-9291-49c5-2685c7450e9a load_backend: loaded CUDA backend from C:\Users\jayli\AppData\Local\Programs\Ollama\lib\ollama\cuda_v13\ggml-cuda.dll time=2025-10-29T16:22:59.632+08:00 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.AVX512=1 CPU.0.AVX512_VBMI=1 CPU.0.AVX512_VNNI=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=750,800,860,870,890,900,1000,1030,1100,1200,1210 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(clang) ggml.c:1921: GGML_ASSERT(ggml_can_repeat(b, a)) failed time=2025-10-29T16:22:59.839+08:00 level=INFO source=sched.go:446 msg="Load failed" model=G:\ollama\models\blobs\sha256-e796792eba26c4d3b04b0ac5adb01a453dd9ec2dfd83b6c59cbf6fe5f30b0f68 error="do load request: Post "http://127.0.0.1:7761/load": read tcp 127.0.0.1:7769->127.0.0.1:7761: wsarecv: An existing connection was forcibly closed by the remote host." [GIN] 2025/10/29 - 16:22:59 | 500 | 708.8267ms | 127.0.0.1 | POST "/api/generate" time=2025-10-29T16:22:59.853+08:00 level=ERROR source=server.go:273 msg="llama runner terminated" error="exit status 0xc0000409"

Relevant log output

OS

Windows

GPU

Nvidia

CPU

AMD

Ollama version

0.12.7-rc0

Originally created by @jay875037671 on GitHub (Oct 29, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/12821 ### What is the issue? those 2 models workd well in 0.12.6(but 0.12.6 qwen3-vl failed though lol) crash on: compiler=cgo(clang) ggml.c:1921: GGML_ASSERT(ggml_can_repeat(b, a)) failed logfile: [GIN] 2025/10/29 - 16:22:50 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/10/29 - 16:22:50 | 200 | 8.5009ms | 127.0.0.1 | GET "/api/tags" [GIN] 2025/10/29 - 16:22:59 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/10/29 - 16:22:59 | 200 | 72.9136ms | 127.0.0.1 | POST "/api/show" ggml_backend_cuda_device_get_memory device GPU-2fb823f9-0c5c-9291-49c5-2685c7450e9a utilizing NVML memory reporting free: 20149010432 total: 34190917632 time=2025-10-29T16:22:59.298+08:00 level=INFO source=cpu_windows.go:139 msg=packages count=1 time=2025-10-29T16:22:59.298+08:00 level=INFO source=cpu_windows.go:186 msg="" package=0 cores=8 efficiency=0 threads=16 time=2025-10-29T16:22:59.322+08:00 level=INFO source=sched.go:559 msg="updated VRAM based on existing loaded models" gpu=GPU-2fb823f9-0c5c-9291-49c5-2685c7450e9a library=CUDA total="31.8 GiB" available="18.8 GiB" time=2025-10-29T16:22:59.399+08:00 level=INFO source=server.go:215 msg="enabling flash attention" time=2025-10-29T16:22:59.422+08:00 level=INFO source=server.go:385 msg="starting runner" cmd="C:\\Users\\jayli\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --model G:\\ollama\\models\\blobs\\sha256-e796792eba26c4d3b04b0ac5adb01a453dd9ec2dfd83b6c59cbf6fe5f30b0f68 --port 7761" time=2025-10-29T16:22:59.435+08:00 level=INFO source=server.go:638 msg="loading model" "model layers"=63 requested=-1 time=2025-10-29T16:22:59.435+08:00 level=INFO source=server.go:643 msg="system memory" total="63.6 GiB" free="46.1 GiB" free_swap="95.0 GiB" time=2025-10-29T16:22:59.435+08:00 level=INFO source=server.go:650 msg="gpu memory" id=GPU-2fb823f9-0c5c-9291-49c5-2685c7450e9a library=CUDA available="18.3 GiB" free="18.8 GiB" minimum="457.0 MiB" overhead="0 B" time=2025-10-29T16:22:59.478+08:00 level=INFO source=runner.go:1337 msg="starting ollama engine" time=2025-10-29T16:22:59.498+08:00 level=INFO source=runner.go:1372 msg="Server listening on 127.0.0.1:7761" time=2025-10-29T16:22:59.499+08:00 level=INFO source=runner.go:1210 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:true KvSize:4096 KvCacheType: NumThreads:8 GPULayers:63[ID:GPU-2fb823f9-0c5c-9291-49c5-2685c7450e9a Layers:63(0..62)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2025-10-29T16:22:59.540+08:00 level=INFO source=ggml.go:135 msg="" architecture=gemma3 file_type=Q4_K_M name="" description="" num_tensors=1247 num_key_values=37 load_backend: loaded CPU backend from C:\Users\jayli\AppData\Local\Programs\Ollama\lib\ollama\ggml-cpu-icelake.dll ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no ggml_cuda_init: found 1 CUDA devices: Device 0: NVIDIA GeForce RTX 5090 D, compute capability 12.0, VMM: yes, ID: GPU-2fb823f9-0c5c-9291-49c5-2685c7450e9a load_backend: loaded CUDA backend from C:\Users\jayli\AppData\Local\Programs\Ollama\lib\ollama\cuda_v13\ggml-cuda.dll time=2025-10-29T16:22:59.632+08:00 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.AVX512=1 CPU.0.AVX512_VBMI=1 CPU.0.AVX512_VNNI=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=750,800,860,870,890,900,1000,1030,1100,1200,1210 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(clang) ggml.c:1921: GGML_ASSERT(ggml_can_repeat(b, a)) failed time=2025-10-29T16:22:59.839+08:00 level=INFO source=sched.go:446 msg="Load failed" model=G:\ollama\models\blobs\sha256-e796792eba26c4d3b04b0ac5adb01a453dd9ec2dfd83b6c59cbf6fe5f30b0f68 error="do load request: Post \"http://127.0.0.1:7761/load\": read tcp 127.0.0.1:7769->127.0.0.1:7761: wsarecv: An existing connection was forcibly closed by the remote host." [GIN] 2025/10/29 - 16:22:59 | 500 | 708.8267ms | 127.0.0.1 | POST "/api/generate" time=2025-10-29T16:22:59.853+08:00 level=ERROR source=server.go:273 msg="llama runner terminated" error="exit status 0xc0000409" ### Relevant log output ```shell ``` ### OS Windows ### GPU Nvidia ### CPU AMD ### Ollama version 0.12.7-rc0
GiteaMirror added the bug label 2026-04-22 17:41:48 -05:00
GiteaMirror commented 2026-04-22 17:41:50 -05:00
Author
Owner
Copy Link

@theuncivilizedarchive commented on GitHub (Oct 29, 2025):

failure within the ggml library, specifically GGML_ASSERT(ggml_can_repeat(b, a)) failed in file ggml.c at line 1921, downgrade Ollama to a stable version or use CPU fallback

<!-- gh-comment-id:3460882679 --> @theuncivilizedarchive commented on GitHub (Oct 29, 2025): failure within the ggml library, specifically GGML_ASSERT(ggml_can_repeat(b, a)) failed in file ggml.c at line 1921, downgrade Ollama to a stable version or use CPU fallback
GiteaMirror commented 2026-04-22 17:41:51 -05:00
Author
Owner
Copy Link

@maglat commented on GitHub (Oct 29, 2025):

Yes, Gemma 3 fail for me in 0.12.7-rc0 as well

<!-- gh-comment-id:3461221184 --> @maglat commented on GitHub (Oct 29, 2025): Yes, Gemma 3 fail for me in 0.12.7-rc0 as well
GiteaMirror commented 2026-04-22 17:41:51 -05:00
Author
Owner
Copy Link

@Burnarz commented on GitHub (Oct 29, 2025):

I'm also using gemma3:27b-it-qat, but got no errors with official release of 0.12.7

<!-- gh-comment-id:3464829867 --> @Burnarz commented on GitHub (Oct 29, 2025): I'm also using gemma3:27b-it-qat, but got no errors with official release of 0.12.7
GiteaMirror commented 2026-04-22 17:41:53 -05:00
Author
Owner
Copy Link

@jessegross commented on GitHub (Oct 29, 2025):

Fixed by #12834

<!-- gh-comment-id:3464860979 --> @jessegross commented on GitHub (Oct 29, 2025): Fixed by #12834
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#34256
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?