github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-22 05:31:43 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #13357] Cannot use ministral-3 #8821

New Issue
Closed
opened 2026-04-12 21:36:17 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-04-12 21:36:17 -05:00
Owner
Copy Link

Originally created by @dipping5115 on GitHub (Dec 6, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/13357

What is the issue?

Cannot use ministral-3, not use graphics card, and constantly displays memory insufficient.

Relevant log output

12月 07 00:36:31 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:31 | 200 |    4.666222ms |       127.0.0.1 | GET      "/api/tags"
12月 07 00:36:33 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:33 | 200 |      13.202µs |       127.0.0.1 | HEAD     "/"
12月 07 00:36:33 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:33 | 200 |      20.936µs |       127.0.0.1 | GET      "/api/ps"
12月 07 00:36:40 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:40 | 200 |       13.45µs |       127.0.0.1 | HEAD     "/"
12月 07 00:36:40 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:40 | 200 |   34.931486ms |       127.0.0.1 | POST     "/api/show"
12月 07 00:36:40 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:40 | 200 |     419.301µs |       127.0.0.1 | POST     "/api/generate"
12月 07 00:36:40 debian ollama[1574]: ggml_backend_cuda_device_get_memory device GPU-8f1b8bb2-6dc1-6a85-f0ff-351151a53251 utilizing NVML memory reporting fr
ee: 799145984 total: 8585740288
12月 07 00:36:41 debian ollama[1574]: time=2025-12-07T00:36:41.065+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama ru
nner --ollama-engine --port 33755"
12月 07 00:36:42 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:42 | 200 |        13.5µs |       127.0.0.1 | HEAD     "/"
12月 07 00:36:42 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:42 | 200 |       7.255µs |       127.0.0.1 | GET      "/api/ps"
12月 07 00:36:57 debian systemd[1]: Stopping ollama.service - Ollama Service...
12月 07 00:36:57 debian systemd[1]: ollama.service: Deactivated successfully.
12月 07 00:36:57 debian systemd[1]: Stopped ollama.service - Ollama Service.
12月 07 00:36:57 debian systemd[1]: ollama.service: Consumed 3h 4min 36.944s CPU time over 1d 6h 1min 35.928s wall clock time, 7.2G memory peak, 9.2M memory
 swap peak.
12月 07 00:36:57 debian systemd[1]: Started ollama.service - Ollama Service.
12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.723+08:00 level=INFO source=routes.go:1544 msg="server config" env="map[CUDA_VISIBLE_DEVICE
S: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:4096
 OLLAMA_DEBUG:INFO OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_L
LM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/usr/share/ollama/.ollama/models OLLAMA_MULTIUSER_CACHE:f
alse OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[* http://localhost https://localhost http://l
ocalhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* htt
ps://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false OLLAMA_VULKAN:false ROCR
_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]"
12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.726+08:00 level=INFO source=images.go:522 msg="total blobs: 50"
12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.727+08:00 level=INFO source=images.go:529 msg="total unused blobs removed: 0"
12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.727+08:00 level=INFO source=routes.go:1597 msg="Listening on [::]:11434 (version 0.13.1)"
12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.728+08:00 level=INFO source=runner.go:67 msg="discovering available GPUs..."
12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.728+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama 
runner --ollama-engine --port 46353"
12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.789+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama 
runner --ollama-engine --port 38257"
12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.945+08:00 level=INFO source=runner.go:106 msg="experimental Vulkan support disabled.  To en
able, set OLLAMA_VULKAN=1"
12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.945+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama 
runner --ollama-engine --port 34761"
12月 07 00:36:58 debian ollama[594433]: time=2025-12-07T00:36:58.004+08:00 level=INFO source=types.go:42 msg="inference compute" id=GPU-8f1b8bb2-6dc1-6a85-f
0ff-351151a53251 filter_id="" library=CUDA compute=8.9 name=CUDA0 description="NVIDIA GeForce RTX 4060 Laptop GPU" libdirs=ollama,cuda_v12 driver=12.4 pci_i
d=0000:01:00.0 type=discrete total="8.0 GiB" available="7.7 GiB"
12月 07 00:36:58 debian ollama[594433]: time=2025-12-07T00:36:58.004+08:00 level=INFO source=routes.go:1638 msg="entering low vram mode" "total vram"="8.0 G
iB" threshold="20.0 GiB"
12月 07 00:37:11 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:11 | 200 |      22.999µs |       127.0.0.1 | HEAD     "/"
12月 07 00:37:11 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:11 | 200 |   40.185698ms |       127.0.0.1 | POST     "/api/show"
12月 07 00:37:11 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:11 | 200 |   36.364435ms |       127.0.0.1 | POST     "/api/show"
12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.378+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama 
runner --ollama-engine --port 42969"
12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.509+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama 
runner --ollama-engine --model /usr/share/ollama/.ollama/models/blobs/sha256-094eb0a75095db5a9f83e51323879750023d7050d008a9b2899bf9f47c4926e5 --port 40657"
12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.510+08:00 level=INFO source=sched.go:443 msg="system memory" total="15.2 GiB" free="3.8 GiB
" free_swap="8.2 MiB"
12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.510+08:00 level=INFO source=sched.go:450 msg="gpu memory" id=GPU-8f1b8bb2-6dc1-6a85-f0ff-35
1151a53251 library=CUDA available="7.3 GiB" free="7.7 GiB" minimum="457.0 MiB" overhead="0 B"
12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.510+08:00 level=INFO source=server.go:702 msg="loading model" "model layers"=27 requested=-
1
12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.518+08:00 level=INFO source=runner.go:1398 msg="starting ollama engine"
12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.518+08:00 level=INFO source=runner.go:1433 msg="Server listening on 127.0.0.1:40657"
12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.522+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:fit LoraPath:[] Par
allel:1 BatchSize:512 FlashAttention:false KvSize:4096 KvCacheType: NumThreads:8 GPULayers:27[ID:GPU-8f1b8bb2-6dc1-6a85-f0ff-351151a53251 Layers:27(0..26)] 
MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.545+08:00 level=INFO source=ggml.go:136 msg="" architecture=mistral3 file_type=Q4_K_M name=
"" description="" num_tensors=458 num_key_values=45
12月 07 00:37:11 debian ollama[594433]: load_backend: loaded CPU backend from /usr/local/lib/ollama/libggml-cpu-alderlake.so
12月 07 00:37:11 debian ollama[594433]: ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
12月 07 00:37:11 debian ollama[594433]: ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
12月 07 00:37:11 debian ollama[594433]: ggml_cuda_init: found 1 CUDA devices:
12月 07 00:37:11 debian ollama[594433]:   Device 0: NVIDIA GeForce RTX 4060 Laptop GPU, compute capability 8.9, VMM: yes, ID: GPU-8f1b8bb2-6dc1-6a85-f0ff-35
1151a53251
12月 07 00:37:11 debian ollama[594433]: load_backend: loaded CUDA backend from /usr/local/lib/ollama/cuda_v12/libggml-cuda.so
12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.582+08:00 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 C
PU.0.AVX_VNNI=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=500,520,600,610,700,750,800,860,890,900,
1200 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(gcc)
12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.067+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:fit LoraPath:[] Par
allel:1 BatchSize:512 FlashAttention:false KvSize:4096 KvCacheType: NumThreads:8 GPULayers:[] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=WARN source=server.go:989 msg="model request too large for system" requeste
d="9.3 GiB" available="3.8 GiB" total="15.2 GiB" free="3.8 GiB" swap="8.2 MiB"
12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=device.go:245 msg="model weights" device=CPU size="3.1 GiB"
12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=device.go:256 msg="kv cache" device=CPU size="416.0 MiB"
12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=device.go:267 msg="compute graph" device=CPU size="9.0 GiB"
12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=device.go:272 msg="total memory" size="12.5 GiB"
12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=sched.go:470 msg="Load failed" model=/usr/share/ollama/.ollama/
models/blobs/sha256-094eb0a75095db5a9f83e51323879750023d7050d008a9b2899bf9f47c4926e5 error="model requires more system memory (9.3 GiB) than is available (3
.8 GiB)"
12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:close LoraPath:[] P
arallel:0 BatchSize:0 FlashAttention:false KvSize:0 KvCacheType: NumThreads:0 GPULayers:[] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
12月 07 00:37:12 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:12 | 500 |  1.017087922s |       127.0.0.1 | POST     "/api/generate"
12月 07 00:37:24 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:24 | 200 |      14.214µs |       127.0.0.1 | HEAD     "/"
12月 07 00:37:24 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:24 | 200 |   71.564641ms |       127.0.0.1 | POST     "/api/show"
12月 07 00:37:24 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:24 | 200 |   63.145927ms |       127.0.0.1 | POST     "/api/show"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.081+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama 
runner --ollama-engine --port 41547"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.250+08:00 level=INFO source=server.go:209 msg="enabling flash attention"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.250+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama 
runner --ollama-engine --model /usr/share/ollama/.ollama/models/blobs/sha256-aeda25e63ebd698fab8638ffb778e68bed908b960d39d0becc650fa981609d25 --port 38143"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.250+08:00 level=INFO source=sched.go:443 msg="system memory" total="15.2 GiB" free="3.8 GiB
" free_swap="8.3 MiB"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.250+08:00 level=INFO source=sched.go:450 msg="gpu memory" id=GPU-8f1b8bb2-6dc1-6a85-f0ff-35
1151a53251 library=CUDA available="7.3 GiB" free="7.7 GiB" minimum="457.0 MiB" overhead="0 B"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.250+08:00 level=INFO source=server.go:702 msg="loading model" "model layers"=35 requested=-
1
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.259+08:00 level=INFO source=runner.go:1398 msg="starting ollama engine"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.260+08:00 level=INFO source=runner.go:1433 msg="Server listening on 127.0.0.1:38143"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.261+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:fit LoraPath:[] Par
allel:1 BatchSize:512 FlashAttention:true KvSize:4096 KvCacheType: NumThreads:8 GPULayers:35[ID:GPU-8f1b8bb2-6dc1-6a85-f0ff-351151a53251 Layers:35(0..34)] M
ultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.301+08:00 level=INFO source=ggml.go:136 msg="" architecture=gemma3 file_type=Q4_K_M name=""
 description="" num_tensors=883 num_key_values=36
12月 07 00:37:25 debian ollama[594433]: load_backend: loaded CPU backend from /usr/local/lib/ollama/libggml-cpu-alderlake.so
12月 07 00:37:25 debian ollama[594433]: ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
12月 07 00:37:25 debian ollama[594433]: ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
12月 07 00:37:25 debian ollama[594433]: ggml_cuda_init: found 1 CUDA devices:
12月 07 00:37:25 debian ollama[594433]:   Device 0: NVIDIA GeForce RTX 4060 Laptop GPU, compute capability 8.9, VMM: yes, ID: GPU-8f1b8bb2-6dc1-6a85-f0ff-35
1151a53251
12月 07 00:37:25 debian ollama[594433]: load_backend: loaded CUDA backend from /usr/local/lib/ollama/cuda_v12/libggml-cuda.so
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.339+08:00 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 C
PU.0.AVX_VNNI=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=500,520,600,610,700,750,800,860,890,900,
1200 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(gcc)
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.632+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:alloc LoraPath:[] P
arallel:1 BatchSize:512 FlashAttention:true KvSize:4096 KvCacheType: NumThreads:8 GPULayers:35[ID:GPU-8f1b8bb2-6dc1-6a85-f0ff-351151a53251 Layers:35(0..34)]
 MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:commit LoraPath:[] 
Parallel:1 BatchSize:512 FlashAttention:true KvSize:4096 KvCacheType: NumThreads:8 GPULayers:35[ID:GPU-8f1b8bb2-6dc1-6a85-f0ff-351151a53251 Layers:35(0..34)
] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=ggml.go:482 msg="offloading 34 repeating layers to GPU"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=ggml.go:489 msg="offloading output layer to GPU"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=ggml.go:494 msg="offloaded 35/35 layers to GPU"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:240 msg="model weights" device=CUDA0 size="3.1 GiB"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:245 msg="model weights" device=CPU size="525.0 MiB"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:251 msg="kv cache" device=CUDA0 size="254.0 MiB"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:262 msg="compute graph" device=CUDA0 size="1.1 GiB"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:267 msg="compute graph" device=CPU size="5.0 MiB"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:272 msg="total memory" size="5.0 GiB"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=sched.go:517 msg="loaded runners" count=1
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=server.go:1294 msg="waiting for llama runner to start respondin
g"
12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=server.go:1328 msg="waiting for server to become available" sta
tus="llm server loading model"
12月 07 00:37:27 debian ollama[594433]: time=2025-12-07T00:37:27.077+08:00 level=INFO source=server.go:1332 msg="llama runner started in 1.83 seconds"
12月 07 00:37:27 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:27 | 200 |  2.127643316s |       127.0.0.1 | POST     "/api/generate"
12月 07 00:37:29 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:29 | 200 |  1.040514045s |       127.0.0.1 | POST     "/api/chat"

OS

Linux

GPU

Nvidia

CPU

Intel

Ollama version

ollama version is 0.13.1

Originally created by @dipping5115 on GitHub (Dec 6, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/13357 ### What is the issue? Cannot use ministral-3, not use graphics card, and constantly displays memory insufficient. ### Relevant log output ```shell 12月 07 00:36:31 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:31 | 200 | 4.666222ms | 127.0.0.1 | GET "/api/tags" 12月 07 00:36:33 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:33 | 200 | 13.202µs | 127.0.0.1 | HEAD "/" 12月 07 00:36:33 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:33 | 200 | 20.936µs | 127.0.0.1 | GET "/api/ps" 12月 07 00:36:40 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:40 | 200 | 13.45µs | 127.0.0.1 | HEAD "/" 12月 07 00:36:40 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:40 | 200 | 34.931486ms | 127.0.0.1 | POST "/api/show" 12月 07 00:36:40 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:40 | 200 | 419.301µs | 127.0.0.1 | POST "/api/generate" 12月 07 00:36:40 debian ollama[1574]: ggml_backend_cuda_device_get_memory device GPU-8f1b8bb2-6dc1-6a85-f0ff-351151a53251 utilizing NVML memory reporting fr ee: 799145984 total: 8585740288 12月 07 00:36:41 debian ollama[1574]: time=2025-12-07T00:36:41.065+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama ru nner --ollama-engine --port 33755" 12月 07 00:36:42 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:42 | 200 | 13.5µs | 127.0.0.1 | HEAD "/" 12月 07 00:36:42 debian ollama[1574]: [GIN] 2025/12/07 - 00:36:42 | 200 | 7.255µs | 127.0.0.1 | GET "/api/ps" 12月 07 00:36:57 debian systemd[1]: Stopping ollama.service - Ollama Service... 12月 07 00:36:57 debian systemd[1]: ollama.service: Deactivated successfully. 12月 07 00:36:57 debian systemd[1]: Stopped ollama.service - Ollama Service. 12月 07 00:36:57 debian systemd[1]: ollama.service: Consumed 3h 4min 36.944s CPU time over 1d 6h 1min 35.928s wall clock time, 7.2G memory peak, 9.2M memory swap peak. 12月 07 00:36:57 debian systemd[1]: Started ollama.service - Ollama Service. 12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.723+08:00 level=INFO source=routes.go:1544 msg="server config" env="map[CUDA_VISIBLE_DEVICE S: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:4096 OLLAMA_DEBUG:INFO OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_L LM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/usr/share/ollama/.ollama/models OLLAMA_MULTIUSER_CACHE:f alse OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[* http://localhost https://localhost http://l ocalhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* htt ps://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false OLLAMA_VULKAN:false ROCR _VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]" 12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.726+08:00 level=INFO source=images.go:522 msg="total blobs: 50" 12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.727+08:00 level=INFO source=images.go:529 msg="total unused blobs removed: 0" 12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.727+08:00 level=INFO source=routes.go:1597 msg="Listening on [::]:11434 (version 0.13.1)" 12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.728+08:00 level=INFO source=runner.go:67 msg="discovering available GPUs..." 12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.728+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 46353" 12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.789+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 38257" 12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.945+08:00 level=INFO source=runner.go:106 msg="experimental Vulkan support disabled. To en able, set OLLAMA_VULKAN=1" 12月 07 00:36:57 debian ollama[594433]: time=2025-12-07T00:36:57.945+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 34761" 12月 07 00:36:58 debian ollama[594433]: time=2025-12-07T00:36:58.004+08:00 level=INFO source=types.go:42 msg="inference compute" id=GPU-8f1b8bb2-6dc1-6a85-f 0ff-351151a53251 filter_id="" library=CUDA compute=8.9 name=CUDA0 description="NVIDIA GeForce RTX 4060 Laptop GPU" libdirs=ollama,cuda_v12 driver=12.4 pci_i d=0000:01:00.0 type=discrete total="8.0 GiB" available="7.7 GiB" 12月 07 00:36:58 debian ollama[594433]: time=2025-12-07T00:36:58.004+08:00 level=INFO source=routes.go:1638 msg="entering low vram mode" "total vram"="8.0 G iB" threshold="20.0 GiB" 12月 07 00:37:11 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:11 | 200 | 22.999µs | 127.0.0.1 | HEAD "/" 12月 07 00:37:11 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:11 | 200 | 40.185698ms | 127.0.0.1 | POST "/api/show" 12月 07 00:37:11 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:11 | 200 | 36.364435ms | 127.0.0.1 | POST "/api/show" 12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.378+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 42969" 12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.509+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --model /usr/share/ollama/.ollama/models/blobs/sha256-094eb0a75095db5a9f83e51323879750023d7050d008a9b2899bf9f47c4926e5 --port 40657" 12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.510+08:00 level=INFO source=sched.go:443 msg="system memory" total="15.2 GiB" free="3.8 GiB " free_swap="8.2 MiB" 12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.510+08:00 level=INFO source=sched.go:450 msg="gpu memory" id=GPU-8f1b8bb2-6dc1-6a85-f0ff-35 1151a53251 library=CUDA available="7.3 GiB" free="7.7 GiB" minimum="457.0 MiB" overhead="0 B" 12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.510+08:00 level=INFO source=server.go:702 msg="loading model" "model layers"=27 requested=- 1 12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.518+08:00 level=INFO source=runner.go:1398 msg="starting ollama engine" 12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.518+08:00 level=INFO source=runner.go:1433 msg="Server listening on 127.0.0.1:40657" 12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.522+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:fit LoraPath:[] Par allel:1 BatchSize:512 FlashAttention:false KvSize:4096 KvCacheType: NumThreads:8 GPULayers:27[ID:GPU-8f1b8bb2-6dc1-6a85-f0ff-351151a53251 Layers:27(0..26)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" 12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.545+08:00 level=INFO source=ggml.go:136 msg="" architecture=mistral3 file_type=Q4_K_M name= "" description="" num_tensors=458 num_key_values=45 12月 07 00:37:11 debian ollama[594433]: load_backend: loaded CPU backend from /usr/local/lib/ollama/libggml-cpu-alderlake.so 12月 07 00:37:11 debian ollama[594433]: ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no 12月 07 00:37:11 debian ollama[594433]: ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no 12月 07 00:37:11 debian ollama[594433]: ggml_cuda_init: found 1 CUDA devices: 12月 07 00:37:11 debian ollama[594433]: Device 0: NVIDIA GeForce RTX 4060 Laptop GPU, compute capability 8.9, VMM: yes, ID: GPU-8f1b8bb2-6dc1-6a85-f0ff-35 1151a53251 12月 07 00:37:11 debian ollama[594433]: load_backend: loaded CUDA backend from /usr/local/lib/ollama/cuda_v12/libggml-cuda.so 12月 07 00:37:11 debian ollama[594433]: time=2025-12-07T00:37:11.582+08:00 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 C PU.0.AVX_VNNI=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=500,520,600,610,700,750,800,860,890,900, 1200 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(gcc) 12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.067+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:fit LoraPath:[] Par allel:1 BatchSize:512 FlashAttention:false KvSize:4096 KvCacheType: NumThreads:8 GPULayers:[] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" 12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=WARN source=server.go:989 msg="model request too large for system" requeste d="9.3 GiB" available="3.8 GiB" total="15.2 GiB" free="3.8 GiB" swap="8.2 MiB" 12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=device.go:245 msg="model weights" device=CPU size="3.1 GiB" 12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=device.go:256 msg="kv cache" device=CPU size="416.0 MiB" 12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=device.go:267 msg="compute graph" device=CPU size="9.0 GiB" 12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=device.go:272 msg="total memory" size="12.5 GiB" 12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=sched.go:470 msg="Load failed" model=/usr/share/ollama/.ollama/ models/blobs/sha256-094eb0a75095db5a9f83e51323879750023d7050d008a9b2899bf9f47c4926e5 error="model requires more system memory (9.3 GiB) than is available (3 .8 GiB)" 12月 07 00:37:12 debian ollama[594433]: time=2025-12-07T00:37:12.298+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:close LoraPath:[] P arallel:0 BatchSize:0 FlashAttention:false KvSize:0 KvCacheType: NumThreads:0 GPULayers:[] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" 12月 07 00:37:12 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:12 | 500 | 1.017087922s | 127.0.0.1 | POST "/api/generate" 12月 07 00:37:24 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:24 | 200 | 14.214µs | 127.0.0.1 | HEAD "/" 12月 07 00:37:24 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:24 | 200 | 71.564641ms | 127.0.0.1 | POST "/api/show" 12月 07 00:37:24 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:24 | 200 | 63.145927ms | 127.0.0.1 | POST "/api/show" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.081+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 41547" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.250+08:00 level=INFO source=server.go:209 msg="enabling flash attention" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.250+08:00 level=INFO source=server.go:392 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --model /usr/share/ollama/.ollama/models/blobs/sha256-aeda25e63ebd698fab8638ffb778e68bed908b960d39d0becc650fa981609d25 --port 38143" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.250+08:00 level=INFO source=sched.go:443 msg="system memory" total="15.2 GiB" free="3.8 GiB " free_swap="8.3 MiB" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.250+08:00 level=INFO source=sched.go:450 msg="gpu memory" id=GPU-8f1b8bb2-6dc1-6a85-f0ff-35 1151a53251 library=CUDA available="7.3 GiB" free="7.7 GiB" minimum="457.0 MiB" overhead="0 B" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.250+08:00 level=INFO source=server.go:702 msg="loading model" "model layers"=35 requested=- 1 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.259+08:00 level=INFO source=runner.go:1398 msg="starting ollama engine" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.260+08:00 level=INFO source=runner.go:1433 msg="Server listening on 127.0.0.1:38143" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.261+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:fit LoraPath:[] Par allel:1 BatchSize:512 FlashAttention:true KvSize:4096 KvCacheType: NumThreads:8 GPULayers:35[ID:GPU-8f1b8bb2-6dc1-6a85-f0ff-351151a53251 Layers:35(0..34)] M ultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.301+08:00 level=INFO source=ggml.go:136 msg="" architecture=gemma3 file_type=Q4_K_M name="" description="" num_tensors=883 num_key_values=36 12月 07 00:37:25 debian ollama[594433]: load_backend: loaded CPU backend from /usr/local/lib/ollama/libggml-cpu-alderlake.so 12月 07 00:37:25 debian ollama[594433]: ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no 12月 07 00:37:25 debian ollama[594433]: ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no 12月 07 00:37:25 debian ollama[594433]: ggml_cuda_init: found 1 CUDA devices: 12月 07 00:37:25 debian ollama[594433]: Device 0: NVIDIA GeForce RTX 4060 Laptop GPU, compute capability 8.9, VMM: yes, ID: GPU-8f1b8bb2-6dc1-6a85-f0ff-35 1151a53251 12月 07 00:37:25 debian ollama[594433]: load_backend: loaded CUDA backend from /usr/local/lib/ollama/cuda_v12/libggml-cuda.so 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.339+08:00 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 C PU.0.AVX_VNNI=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=500,520,600,610,700,750,800,860,890,900, 1200 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(gcc) 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.632+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:alloc LoraPath:[] P arallel:1 BatchSize:512 FlashAttention:true KvSize:4096 KvCacheType: NumThreads:8 GPULayers:35[ID:GPU-8f1b8bb2-6dc1-6a85-f0ff-351151a53251 Layers:35(0..34)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=runner.go:1271 msg=load request="{Operation:commit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:true KvSize:4096 KvCacheType: NumThreads:8 GPULayers:35[ID:GPU-8f1b8bb2-6dc1-6a85-f0ff-351151a53251 Layers:35(0..34) ] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=ggml.go:482 msg="offloading 34 repeating layers to GPU" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=ggml.go:489 msg="offloading output layer to GPU" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=ggml.go:494 msg="offloaded 35/35 layers to GPU" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:240 msg="model weights" device=CUDA0 size="3.1 GiB" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:245 msg="model weights" device=CPU size="525.0 MiB" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:251 msg="kv cache" device=CUDA0 size="254.0 MiB" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:262 msg="compute graph" device=CUDA0 size="1.1 GiB" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:267 msg="compute graph" device=CPU size="5.0 MiB" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=device.go:272 msg="total memory" size="5.0 GiB" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=sched.go:517 msg="loaded runners" count=1 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=server.go:1294 msg="waiting for llama runner to start respondin g" 12月 07 00:37:25 debian ollama[594433]: time=2025-12-07T00:37:25.822+08:00 level=INFO source=server.go:1328 msg="waiting for server to become available" sta tus="llm server loading model" 12月 07 00:37:27 debian ollama[594433]: time=2025-12-07T00:37:27.077+08:00 level=INFO source=server.go:1332 msg="llama runner started in 1.83 seconds" 12月 07 00:37:27 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:27 | 200 | 2.127643316s | 127.0.0.1 | POST "/api/generate" 12月 07 00:37:29 debian ollama[594433]: [GIN] 2025/12/07 - 00:37:29 | 200 | 1.040514045s | 127.0.0.1 | POST "/api/chat" ``` ### OS Linux ### GPU Nvidia ### CPU Intel ### Ollama version ollama version is 0.13.1
GiteaMirror added the bug label 2026-04-12 21:36:17 -05:00
GiteaMirror commented 2026-04-12 21:36:19 -05:00
Author
Owner
Copy Link

@Tyguy047 commented on GitHub (Dec 6, 2025):

It looks like Ollama is trying to load the model to the systems ram (which is out of space) instead of VRAM. Could be a driver issue. I would try installing NVIDIA's proprietary drivers because it could have something to do with Ollama not detecting your GPU as a useable device.

Before you do that you can also try:

ollama run mistral-3 --num-gpu 27

or

export OLLAMA_NUM_GPU=-1
ollama run mistral-3

Also try restarting Ollama

sudo systemctl restart ollama

And make sure CUDA is working with

nvidia-smi

if nvidia-smi isn't found it is likely a driver issue.

<!-- gh-comment-id:3620984722 --> @Tyguy047 commented on GitHub (Dec 6, 2025): It looks like Ollama is trying to load the model to the systems ram (which is out of space) instead of VRAM. Could be a driver issue. I would try installing NVIDIA's proprietary drivers because it could have something to do with Ollama not detecting your GPU as a useable device. Before you do that you can also try: ``` ollama run mistral-3 --num-gpu 27 ``` or ``` export OLLAMA_NUM_GPU=-1 ollama run mistral-3 ``` Also try restarting Ollama ``` sudo systemctl restart ollama ``` And make sure CUDA is working with ``` nvidia-smi ``` if nvidia-smi isn't found it is likely a driver issue.
GiteaMirror commented 2026-04-12 21:36:20 -05:00
Author
Owner
Copy Link

@dipping5115 commented on GitHub (Dec 6, 2025):

It’s not an issue with the drivers. I’m running other models, including gemma and qwen3-vl, which can run on the GPU. You can also find evidence of the GPU working in the above logs.

$ ollama run ministral-3:3b --num-gpu 27
Error: unknown flag: --num-gpu
$ OLLAMA_NUM_GPU=-1 ollama run ministral-3:3b
Error: 500 Internal Server Error: model requires more system memory (9.3 GiB) than is available (7.0 GiB)
<!-- gh-comment-id:3621068577 --> @dipping5115 commented on GitHub (Dec 6, 2025): It’s not an issue with the drivers. I’m running other models, including gemma and qwen3-vl, which can run on the GPU. You can also find evidence of the GPU working in the above logs. ```bash $ ollama run ministral-3:3b --num-gpu 27 Error: unknown flag: --num-gpu $ OLLAMA_NUM_GPU=-1 ollama run ministral-3:3b Error: 500 Internal Server Error: model requires more system memory (9.3 GiB) than is available (7.0 GiB) ```
GiteaMirror commented 2026-04-12 21:36:20 -05:00
Author
Owner
Copy Link

@Tyguy047 commented on GitHub (Dec 8, 2025):

Could be a laptop issue too or something deeper in Debian. I was paying more attention to this part of your original post:

time=2025-12-07T00:37:12.298+08:00 level=WARN source=server.go:989 msg="model request too large for system" requested="9.3 GiB" available="3.8 GiB" total="15.2 GiB" free="3.8 GiB" swap="8.2 MiB"

Seems like it isn't detecting your GPU's RAM (VRAM).

<!-- gh-comment-id:3624303466 --> @Tyguy047 commented on GitHub (Dec 8, 2025): Could be a laptop issue too or something deeper in Debian. I was paying more attention to this part of your original post: ``` time=2025-12-07T00:37:12.298+08:00 level=WARN source=server.go:989 msg="model request too large for system" requested="9.3 GiB" available="3.8 GiB" total="15.2 GiB" free="3.8 GiB" swap="8.2 MiB" ``` Seems like it isn't detecting your GPU's RAM (VRAM).
GiteaMirror commented 2026-04-12 21:36:21 -05:00
Author
Owner
Copy Link

@dipping5115 commented on GitHub (Dec 8, 2025):

No, the problem isn't just this one, there's another issue: it can't run on the CPU either.

$ ollama run ministral-3:3b
Error: 500 Internal Server Error: model requires more system memory (11.7 GiB) than is available (11.6 GiB)

What you're saying about not detecting the GPU memory is also strange, as I mentioned above, I can run other models directly using my GPU, and I noticed that this model requires the latest version of ollama. I suspect that the latest version of ollama has made specific optimizations for this model, which has indirectly caused this problem to occur.

<!-- gh-comment-id:3625575505 --> @dipping5115 commented on GitHub (Dec 8, 2025): No, the problem isn't just this one, there's another issue: it can't run on the CPU either. ```shell $ ollama run ministral-3:3b Error: 500 Internal Server Error: model requires more system memory (11.7 GiB) than is available (11.6 GiB) ``` What you're saying about not detecting the GPU memory is also strange, as I mentioned above, I can run other models directly using my GPU, and I noticed that this model requires the latest version of ollama. I suspect that the latest version of ollama has made specific optimizations for this model, which has indirectly caused this problem to occur.
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#8821
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?