github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 08:30:05 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #9934] ollama v0.6.2 gemma 3 OOM: Killed process #32261

New Issue
Closed
opened 2026-04-22 13:21:35 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-22 13:21:35 -05:00
Owner
Copy Link

Originally created by @akshaal on GitHub (Mar 22, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/9934

Originally assigned to: @jessegross on GitHub.

What is the issue?

Gemma 3 IQ4 XS, num_ctx 32k, num_predict 32k. nvidia 4090 24g, 128g RAM

OLLAMA_FLASH_ATTENTION=1 OLLAMA_KV_CACHE_TYPE=q8_0 ./ollama serve

Image
Image

Relevant log output

ollama stdout+stderr:
  2025/03/21 19:47:08    routes.go:1230: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:2048 OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:true OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE:q8_0 OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/home/user/.ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* 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:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]"
time=2025-03-21T19:47:08.925+01:00 level=INFO source=images.go:432 msg="total blobs: 40"
time=2025-03-21T19:47:08.926+01:00 level=INFO source=images.go:439 msg="total unused blobs removed: 0"
time=2025-03-21T19:47:08.927+01:00 level=INFO source=routes.go:1297 msg="Listening on 127.0.0.1:11434 (version 0.6.2)"
time=2025-03-21T19:47:08.927+01:00 level=INFO source=gpu.go:217 msg="looking for compatible GPUs"
time=2025-03-21T19:47:09.049+01:00 level=INFO source=types.go:130 msg="inference compute" id=GPU-22eaf998-1aa8-14bc-3c72-c7275965de5e library=cuda variant=v12 compute=8.9 driver=12.2 name="NVIDIA GeForce RTX 4090" total="23.6 GiB" available="22.1 GiB"
time=2025-03-21T19:47:20.652+01:00 level=INFO source=server.go:105 msg="system memory" total="125.7 GiB" free="115.3 GiB" free_swap="0 B"
time=2025-03-21T19:47:20.732+01:00 level=INFO source=server.go:138 msg=offload library=cuda layers.requested=-1 layers.model=63 layers.offload=62 layers.split="" memory.available="[22.1 GiB]" memory.gpu_overhead="0 B" memory.required.full="22.8 GiB" memory.required.partial="21.7 GiB" memory.required.kv="5.8 GiB" memory.required.allocations="[21.7 GiB]" memory.weights.total="12.7 GiB" memory.weights.repeating="12.7 GiB" memory.weights.nonrepeating="1.1 GiB" memory.graph.full="1.6 GiB" memory.graph.partial="1.7 GiB" projector.weights="818.0 MiB" projector.graph="0 B"
time=2025-03-21T19:47:20.732+01:00 level=INFO source=server.go:185 msg="enabling flash attention"
time=2025-03-21T19:47:20.798+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.pretokenizer default="(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+"
time=2025-03-21T19:47:20.813+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.add_eot_token default=false
time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.image_size default=0
time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.patch_size default=0
time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.num_channels default=0
time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.block_count default=0
time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.embedding_length default=0
time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.attention.head_count default=0
time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.image_size default=0
time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.patch_size default=0
time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.attention.layer_norm_epsilon default=0
time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.pretokenizer default="(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+"
time=2025-03-21T19:47:20.823+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.local.freq_base default=10000
time=2025-03-21T19:47:20.823+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.global.freq_base default=1e+06
time=2025-03-21T19:47:20.823+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.freq_scale default=1
time=2025-03-21T19:47:20.823+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.mm_tokens_per_image default=256
time=2025-03-21T19:47:20.824+01:00 level=INFO source=server.go:405 msg="starting llama server" cmd="/home/user/Downloads/ollama/bin/ollama runner --ollama-engine --model /home/user/.ollama/models/blobs/sha256-bd2f188c66d8ccb0bffcb0c91e4dbbb72754bb1732e0bca323a2f266a35e01c8 --ctx-size 24576 --batch-size 512 --n-gpu-layers 62 --threads 12 --flash-attn --kv-cache-type q8_0 --parallel 1 --port 41731"
time=2025-03-21T19:47:20.824+01:00 level=INFO source=sched.go:450 msg="loaded runners" count=1
time=2025-03-21T19:47:20.824+01:00 level=INFO source=server.go:580 msg="waiting for llama runner to start responding"
time=2025-03-21T19:47:20.825+01:00 level=INFO source=server.go:614 msg="waiting for server to become available" status="llm server error"
time=2025-03-21T19:47:20.837+01:00 level=INFO source=runner.go:763 msg="starting ollama engine"
time=2025-03-21T19:47:20.837+01:00 level=INFO source=runner.go:823 msg="Server listening on 127.0.0.1:41731"
time=2025-03-21T19:47:20.902+01:00 level=WARN source=ggml.go:149 msg="key not found" key=general.description default=""
time=2025-03-21T19:47:20.902+01:00 level=INFO source=ggml.go:67 msg="" architecture=gemma3 file_type=IQ4_XS name="Gemma 3 27b It" description="" num_tensors=808 num_key_values=45
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 4090, compute capability 8.9, VMM: yes
load_backend: loaded CUDA backend from /home/user/Downloads/ollama/lib/ollama/cuda_v12/libggml-cuda.so
load_backend: loaded CPU backend from /home/user/Downloads/ollama/lib/ollama/libggml-cpu-haswell.so
time=2025-03-21T19:47:20.942+01:00 level=INFO source=ggml.go:109 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.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=500,600,610,700,750,800,860,870,890,900,1200 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(gcc)
time=2025-03-21T19:47:21.001+01:00 level=INFO source=ggml.go:289 msg="model weights" buffer=CPU size="2.2 GiB"
time=2025-03-21T19:47:21.001+01:00 level=INFO source=ggml.go:289 msg="model weights" buffer=CUDA0 size="12.7 GiB"
time=2025-03-21T19:47:21.276+01:00 level=INFO source=server.go:614 msg="waiting for server to become available" status="llm server not responding"
time=2025-03-21T19:47:22.009+01:00 level=INFO source=server.go:614 msg="waiting for server to become available" status="llm server loading model"
time=2025-03-21T19:47:24.735+01:00 level=INFO source=ggml.go:358 msg="compute graph" backend=CUDA0 buffer_type=CUDA0
time=2025-03-21T19:47:24.735+01:00 level=INFO source=ggml.go:358 msg="compute graph" backend=CPU buffer_type=CUDA_Host
time=2025-03-21T19:47:24.735+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.pretokenizer default="(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+"
time=2025-03-21T19:47:24.738+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.add_eot_token default=false
time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.image_size default=0
time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.patch_size default=0
time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.num_channels default=0
time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.block_count default=0
time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.embedding_length default=0
time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.attention.head_count default=0
time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.image_size default=0
time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.patch_size default=0
time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.attention.layer_norm_epsilon default=0
time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.pretokenizer default="(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+"
time=2025-03-21T19:47:24.747+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.local.freq_base default=10000
time=2025-03-21T19:47:24.747+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.global.freq_base default=1e+06
time=2025-03-21T19:47:24.747+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.freq_scale default=1
time=2025-03-21T19:47:24.747+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.mm_tokens_per_image default=256
time=2025-03-21T19:47:24.775+01:00 level=INFO source=server.go:619 msg="llama runner started in 3.95 seconds"
[GIN] 2025/03/21 - 19:56:26 | 200 |      22.462µs |       127.0.0.1 | HEAD     "/"
[GIN] 2025/03/21 - 19:56:26 | 200 |      30.257µs |       127.0.0.1 | GET      "/api/ps"
since it's killed by kernel, there is no error

journald:
[2597940.603198] Out of memory: Killed process 4044306 (ollama) total-vm:248418888kB, anon-rss:9889244kB, file-rss:71688kB, shmem-rss:106964460kB, UID:1000 pgtables:324024kB oom_score_adj:0

OS

Linux

GPU

Nvidia

CPU

AMD

Ollama version

0.6.2

Originally created by @akshaal on GitHub (Mar 22, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/9934 Originally assigned to: @jessegross on GitHub. ### What is the issue? Gemma 3 IQ4 XS, num_ctx 32k, num_predict 32k. nvidia 4090 24g, 128g RAM ``` OLLAMA_FLASH_ATTENTION=1 OLLAMA_KV_CACHE_TYPE=q8_0 ./ollama serve ``` ![Image](https://github.com/user-attachments/assets/87b39473-7007-4591-b241-07a68106c1c5) ![Image](https://github.com/user-attachments/assets/0802e37f-0bb5-4ee6-8771-67827c320e9d) ### Relevant log output ```shell ollama stdout+stderr: 2025/03/21 19:47:08 routes.go:1230: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:2048 OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:true OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE:q8_0 OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:/home/user/.ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* 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:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES: http_proxy: https_proxy: no_proxy:]" time=2025-03-21T19:47:08.925+01:00 level=INFO source=images.go:432 msg="total blobs: 40" time=2025-03-21T19:47:08.926+01:00 level=INFO source=images.go:439 msg="total unused blobs removed: 0" time=2025-03-21T19:47:08.927+01:00 level=INFO source=routes.go:1297 msg="Listening on 127.0.0.1:11434 (version 0.6.2)" time=2025-03-21T19:47:08.927+01:00 level=INFO source=gpu.go:217 msg="looking for compatible GPUs" time=2025-03-21T19:47:09.049+01:00 level=INFO source=types.go:130 msg="inference compute" id=GPU-22eaf998-1aa8-14bc-3c72-c7275965de5e library=cuda variant=v12 compute=8.9 driver=12.2 name="NVIDIA GeForce RTX 4090" total="23.6 GiB" available="22.1 GiB" time=2025-03-21T19:47:20.652+01:00 level=INFO source=server.go:105 msg="system memory" total="125.7 GiB" free="115.3 GiB" free_swap="0 B" time=2025-03-21T19:47:20.732+01:00 level=INFO source=server.go:138 msg=offload library=cuda layers.requested=-1 layers.model=63 layers.offload=62 layers.split="" memory.available="[22.1 GiB]" memory.gpu_overhead="0 B" memory.required.full="22.8 GiB" memory.required.partial="21.7 GiB" memory.required.kv="5.8 GiB" memory.required.allocations="[21.7 GiB]" memory.weights.total="12.7 GiB" memory.weights.repeating="12.7 GiB" memory.weights.nonrepeating="1.1 GiB" memory.graph.full="1.6 GiB" memory.graph.partial="1.7 GiB" projector.weights="818.0 MiB" projector.graph="0 B" time=2025-03-21T19:47:20.732+01:00 level=INFO source=server.go:185 msg="enabling flash attention" time=2025-03-21T19:47:20.798+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.pretokenizer default="(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+" time=2025-03-21T19:47:20.813+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.add_eot_token default=false time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.image_size default=0 time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.patch_size default=0 time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.num_channels default=0 time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.block_count default=0 time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.embedding_length default=0 time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.attention.head_count default=0 time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.image_size default=0 time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.patch_size default=0 time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.attention.layer_norm_epsilon default=0 time=2025-03-21T19:47:20.818+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.pretokenizer default="(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+" time=2025-03-21T19:47:20.823+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.local.freq_base default=10000 time=2025-03-21T19:47:20.823+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.global.freq_base default=1e+06 time=2025-03-21T19:47:20.823+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.freq_scale default=1 time=2025-03-21T19:47:20.823+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.mm_tokens_per_image default=256 time=2025-03-21T19:47:20.824+01:00 level=INFO source=server.go:405 msg="starting llama server" cmd="/home/user/Downloads/ollama/bin/ollama runner --ollama-engine --model /home/user/.ollama/models/blobs/sha256-bd2f188c66d8ccb0bffcb0c91e4dbbb72754bb1732e0bca323a2f266a35e01c8 --ctx-size 24576 --batch-size 512 --n-gpu-layers 62 --threads 12 --flash-attn --kv-cache-type q8_0 --parallel 1 --port 41731" time=2025-03-21T19:47:20.824+01:00 level=INFO source=sched.go:450 msg="loaded runners" count=1 time=2025-03-21T19:47:20.824+01:00 level=INFO source=server.go:580 msg="waiting for llama runner to start responding" time=2025-03-21T19:47:20.825+01:00 level=INFO source=server.go:614 msg="waiting for server to become available" status="llm server error" time=2025-03-21T19:47:20.837+01:00 level=INFO source=runner.go:763 msg="starting ollama engine" time=2025-03-21T19:47:20.837+01:00 level=INFO source=runner.go:823 msg="Server listening on 127.0.0.1:41731" time=2025-03-21T19:47:20.902+01:00 level=WARN source=ggml.go:149 msg="key not found" key=general.description default="" time=2025-03-21T19:47:20.902+01:00 level=INFO source=ggml.go:67 msg="" architecture=gemma3 file_type=IQ4_XS name="Gemma 3 27b It" description="" num_tensors=808 num_key_values=45 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 4090, compute capability 8.9, VMM: yes load_backend: loaded CUDA backend from /home/user/Downloads/ollama/lib/ollama/cuda_v12/libggml-cuda.so load_backend: loaded CPU backend from /home/user/Downloads/ollama/lib/ollama/libggml-cpu-haswell.so time=2025-03-21T19:47:20.942+01:00 level=INFO source=ggml.go:109 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.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=500,600,610,700,750,800,860,870,890,900,1200 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(gcc) time=2025-03-21T19:47:21.001+01:00 level=INFO source=ggml.go:289 msg="model weights" buffer=CPU size="2.2 GiB" time=2025-03-21T19:47:21.001+01:00 level=INFO source=ggml.go:289 msg="model weights" buffer=CUDA0 size="12.7 GiB" time=2025-03-21T19:47:21.276+01:00 level=INFO source=server.go:614 msg="waiting for server to become available" status="llm server not responding" time=2025-03-21T19:47:22.009+01:00 level=INFO source=server.go:614 msg="waiting for server to become available" status="llm server loading model" time=2025-03-21T19:47:24.735+01:00 level=INFO source=ggml.go:358 msg="compute graph" backend=CUDA0 buffer_type=CUDA0 time=2025-03-21T19:47:24.735+01:00 level=INFO source=ggml.go:358 msg="compute graph" backend=CPU buffer_type=CUDA_Host time=2025-03-21T19:47:24.735+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.pretokenizer default="(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+" time=2025-03-21T19:47:24.738+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.add_eot_token default=false time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.image_size default=0 time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.patch_size default=0 time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.num_channels default=0 time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.block_count default=0 time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.embedding_length default=0 time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.attention.head_count default=0 time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.image_size default=0 time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.patch_size default=0 time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.vision.attention.layer_norm_epsilon default=0 time=2025-03-21T19:47:24.741+01:00 level=WARN source=ggml.go:149 msg="key not found" key=tokenizer.ggml.pretokenizer default="(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?\\p{L}+|\\p{N}{1,3}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+" time=2025-03-21T19:47:24.747+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.local.freq_base default=10000 time=2025-03-21T19:47:24.747+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.global.freq_base default=1e+06 time=2025-03-21T19:47:24.747+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.rope.freq_scale default=1 time=2025-03-21T19:47:24.747+01:00 level=WARN source=ggml.go:149 msg="key not found" key=gemma3.mm_tokens_per_image default=256 time=2025-03-21T19:47:24.775+01:00 level=INFO source=server.go:619 msg="llama runner started in 3.95 seconds" [GIN] 2025/03/21 - 19:56:26 | 200 | 22.462µs | 127.0.0.1 | HEAD "/" [GIN] 2025/03/21 - 19:56:26 | 200 | 30.257µs | 127.0.0.1 | GET "/api/ps" since it's killed by kernel, there is no error journald: [2597940.603198] Out of memory: Killed process 4044306 (ollama) total-vm:248418888kB, anon-rss:9889244kB, file-rss:71688kB, shmem-rss:106964460kB, UID:1000 pgtables:324024kB oom_score_adj:0 ``` ### OS Linux ### GPU Nvidia ### CPU AMD ### Ollama version 0.6.2
GiteaMirror added the bug label 2026-04-22 13:21:35 -05:00
GiteaMirror commented 2026-04-22 13:21:40 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Mar 24, 2025):

There are going to be a couple of issues running bartowski's quants on ollama:

  • the memory estimation is going to be off w/ cuda because we don't (yet) optimize this for the kvcache (it should work fine w/ unified memory)
  • we haven't tested iquants/imatrix based weights w/ the ollama engine (the gemma3 architecture uses the ollama engine, and not llama.cpp)
  • bartowski doesn't include the vision tensors in the main model, so you're only going to get the text part of the model. I'm not 100% sure if this will work correctly (it may), but you certainly won't get the vision part
<!-- gh-comment-id:2749121936 --> @pdevine commented on GitHub (Mar 24, 2025): There are going to be a couple of issues running bartowski's quants on ollama: * the memory estimation is going to be off w/ cuda because we don't (yet) optimize this for the kvcache (it should work fine w/ unified memory) * we haven't tested iquants/imatrix based weights w/ the ollama engine (the gemma3 architecture uses the ollama engine, and *not* llama.cpp) * bartowski doesn't include the vision tensors in the main model, so you're only going to get the text part of the model. I'm not 100% sure if this will work correctly (it may), but you certainly won't get the vision part
GiteaMirror commented 2026-04-22 13:21:41 -05:00
Author
Owner
Copy Link

@jetnet commented on GitHub (Mar 29, 2025):

Same issue, when sending lots of requests to /api/generate.
Workaround: a cronjob to stop the model every hour :(

<!-- gh-comment-id:2763329145 --> @jetnet commented on GitHub (Mar 29, 2025): Same issue, when sending lots of requests to `/api/generate`. Workaround: a cronjob to stop the model every hour :(
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#32261
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?