github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 08:02:14 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #14501] Ollama 500 Error with Qwen3.5:35b-a3b and qwen3.5:27b-q4_K_M Models #55920

New Issue
Closed
opened 2026-04-29 09:57:01 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-29 09:57:01 -05:00
Owner
Copy Link

Originally created by @lyvs2012 on GitHub (Feb 27, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/14501

What is the issue?

Issue Report: Qwen3.5 Model Fails with 500 Error on Ollama
System Context:
I can successfully run other models of similar and larger sizes on my local machine (e.g., Qwen3 48GB, GPT-OSS:120B 60GB, GLM-4.7-Flash 19GB).
The hardware specifications are sufficient for heavy models.
Problem Description:
I am encountering a 500 Internal Server Error when attempting to use the Qwen3.5 model via the Ollama client.
Reproduction Steps & Symptoms:
Initial Run: The model loads and opens a conversation session successfully the very first time.
Subsequent Runs: After the initial session, any further attempts to call the model or open a new dialog result in an immediate 500 error.
Model Sizes: The local instances of these models are approximately 17GB and 23GB respectively.
Question:
Why is the Qwen3.5 model specifically failing with a 500 error after the first use, while other significantly larger models function normally?

Relevant log output

Couldn't find 'C:\Users\LiYong\.ollama\id_ed25519'. Generating new private key.
Your new public key is: 

ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAILMrLVPjXm1m8rK8DKoebe3/5pXoygVk8dw2y8QZEsod

time=2026-02-27T16:38:11.468+08:00 level=INFO source=routes.go:1663 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:0 OLLAMA_DEBUG:INFO OLLAMA_EDITOR: OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\\Users\\LiYong\\.ollama\\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NO_CLOUD:false OLLAMA_NUM_PARALLEL:1 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_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false OLLAMA_VULKAN:false ROCR_VISIBLE_DEVICES:]"
time=2026-02-27T16:38:11.471+08:00 level=INFO source=routes.go:1665 msg="Ollama cloud disabled: false"
time=2026-02-27T16:38:11.472+08:00 level=INFO source=images.go:473 msg="total blobs: 0"
time=2026-02-27T16:38:11.472+08:00 level=INFO source=images.go:480 msg="total unused blobs removed: 0"
time=2026-02-27T16:38:11.472+08:00 level=INFO source=routes.go:1718 msg="Listening on 127.0.0.1:11434 (version 0.17.4)"
time=2026-02-27T16:38:11.473+08:00 level=INFO source=runner.go:67 msg="discovering available GPUs..."
time=2026-02-27T16:38:11.483+08:00 level=INFO source=runner.go:106 msg="experimental Vulkan support disabled.  To enable, set OLLAMA_VULKAN=1"
time=2026-02-27T16:38:11.490+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 13703"
time=2026-02-27T16:38:11.768+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 13708"
time=2026-02-27T16:38:11.968+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 13713"
time=2026-02-27T16:38:12.175+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 13720"
time=2026-02-27T16:38:12.175+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 13721"
time=2026-02-27T16:38:12.370+08:00 level=INFO source=types.go:42 msg="inference compute" id=GPU-22e15c16-4a8d-b130-9dda-d5c054898aee filter_id="" library=CUDA compute=12.0 name=CUDA0 description="NVIDIA GeForce RTX 5070 Ti" libdirs=ollama,cuda_v13 driver=13.1 pci_id=0000:01:00.0 type=discrete total="15.9 GiB" available="14.5 GiB"
time=2026-02-27T16:38:12.370+08:00 level=INFO source=routes.go:1768 msg="vram-based default context" total_vram="15.9 GiB" default_num_ctx=4096
[GIN] 2026/02/27 - 16:38:12 | 200 |            0s |       127.0.0.1 | GET      "/api/version"
[GIN] 2026/02/27 - 16:39:09 | 200 |            0s |       127.0.0.1 | GET      "/api/version"
[GIN] 2026/02/27 - 16:39:09 | 200 |            0s |       127.0.0.1 | GET      "/api/version"
[GIN] 2026/02/27 - 16:39:09 | 200 |      2.1133ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:39:09 | 404 |      1.9031ms |       127.0.0.1 | POST     "/api/show"
[GIN] 2026/02/27 - 16:39:10 | 401 |    388.1598ms |       127.0.0.1 | POST     "/api/me"
[GIN] 2026/02/27 - 16:39:10 | 401 |    511.7081ms |       127.0.0.1 | POST     "/api/me"
[GIN] 2026/02/27 - 16:39:14 | 200 |    130.5896ms |       127.0.0.1 | POST     "/api/show"
[GIN] 2026/02/27 - 16:39:17 | 200 |      1.0376ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:39:17 | 200 |    121.9582ms |       127.0.0.1 | POST     "/api/show"
[GIN] 2026/02/27 - 16:39:17 | 200 |    121.0927ms |       127.0.0.1 | POST     "/api/show"
time=2026-02-27T16:39:17.782+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 12315"
time=2026-02-27T16:39:17.938+08:00 level=INFO source=cpu_windows.go:148 msg=packages count=1
time=2026-02-27T16:39:17.938+08:00 level=INFO source=cpu_windows.go:195 msg="" package=0 cores=8 efficiency=0 threads=16
time=2026-02-27T16:39:18.012+08:00 level=INFO source=server.go:247 msg="enabling flash attention"
time=2026-02-27T16:39:18.013+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --model C:\\Users\\LiYong\\.ollama\\models\\blobs\\sha256-7935de6e08f9444536d0edcacf19d2166b34bef8ddb4ac7ce9263ff5cad0693b --port 12320"
time=2026-02-27T16:39:18.024+08:00 level=INFO source=sched.go:491 msg="system memory" total="143.7 GiB" free="130.1 GiB" free_swap="135.4 GiB"
time=2026-02-27T16:39:18.024+08:00 level=INFO source=sched.go:498 msg="gpu memory" id=GPU-22e15c16-4a8d-b130-9dda-d5c054898aee library=CUDA available="13.9 GiB" free="14.3 GiB" minimum="457.0 MiB" overhead="0 B"
time=2026-02-27T16:39:18.024+08:00 level=INFO source=server.go:757 msg="loading model" "model layers"=65 requested=-1
time=2026-02-27T16:39:18.054+08:00 level=INFO source=runner.go:1411 msg="starting ollama engine"
time=2026-02-27T16:39:18.058+08:00 level=INFO source=runner.go:1446 msg="Server listening on 127.0.0.1:12320"
time=2026-02-27T16:39:18.066+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:65[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:65(0..64)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
time=2026-02-27T16:39:18.095+08:00 level=INFO source=ggml.go:136 msg="" architecture=qwen35 file_type=Q4_K_M name="" description="" num_tensors=1307 num_key_values=53
load_backend: loaded CPU backend from C:\Users\LiYong\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 5070 Ti, compute capability 12.0, VMM: yes, ID: GPU-22e15c16-4a8d-b130-9dda-d5c054898aee
load_backend: loaded CUDA backend from C:\Users\LiYong\AppData\Local\Programs\Ollama\lib\ollama\cuda_v13\ggml-cuda.dll
time=2026-02-27T16:39:18.175+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)
time=2026-02-27T16:39:18.644+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:48[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:48(16..63)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
time=2026-02-27T16:39:18.886+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:alloc LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:48[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:48(16..63)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:commit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:48[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:48(16..63)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=ggml.go:482 msg="offloading 48 repeating layers to GPU"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=ggml.go:486 msg="offloading output layer to CPU"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=ggml.go:494 msg="offloaded 48/65 layers to GPU"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:240 msg="model weights" device=CUDA0 size="10.1 GiB"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:245 msg="model weights" device=CPU size="6.2 GiB"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:251 msg="kv cache" device=CUDA0 size="2.9 GiB"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:256 msg="kv cache" device=CPU size="999.2 MiB"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:262 msg="compute graph" device=CUDA0 size="799.3 MiB"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:267 msg="compute graph" device=CPU size="1.0 GiB"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:272 msg="total memory" size="21.9 GiB"
time=2026-02-27T16:39:19.548+08:00 level=INFO source=sched.go:566 msg="loaded runners" count=1
time=2026-02-27T16:39:19.549+08:00 level=INFO source=server.go:1350 msg="waiting for llama runner to start responding"
time=2026-02-27T16:39:19.555+08:00 level=INFO source=server.go:1384 msg="waiting for server to become available" status="llm server loading model"
time=2026-02-27T16:39:21.810+08:00 level=INFO source=server.go:1388 msg="llama runner started in 3.79 seconds"
[GIN] 2026/02/27 - 16:39:47 | 200 |      1.5064ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:40:17 | 200 |      3.1242ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:40:47 | 200 |      1.5386ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:41:04 | 200 |         1m47s |       127.0.0.1 | POST     "/api/chat"
[GIN] 2026/02/27 - 16:41:17 | 200 |      1.0201ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:41:47 | 200 |      1.0384ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:42:17 | 200 |      1.0917ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:42:47 | 200 |      1.0375ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:43:17 | 200 |      1.0341ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:43:45 | 200 |    127.4648ms |       127.0.0.1 | POST     "/api/show"
[GIN] 2026/02/27 - 16:43:45 | 200 |    115.6221ms |       127.0.0.1 | POST     "/api/show"
CUDA error: invalid argument
  current device: 0, in function ggml_cuda_cpy at C:\a\ollama\ollama\ml\backend\ggml\ggml\src\ggml-cuda\cpy.cu:438
  cudaMemcpyAsyncReserve(src1_ddc, src0_ddc, ggml_nbytes(src0), cudaMemcpyDeviceToDevice, main_stream)
C:\a\ollama\ollama\ml\backend\ggml\ggml\src\ggml-cuda\ggml-cuda.cu:94: CUDA error
time=2026-02-27T16:43:46.061+08:00 level=ERROR source=server.go:1610 msg="post predict" error="Post \"http://127.0.0.1:12320/completion\": read tcp 127.0.0.1:14337->127.0.0.1:12320: wsarecv: An existing connection was forcibly closed by the remote host."
[GIN] 2026/02/27 - 16:43:46 | 500 |    458.4568ms |       127.0.0.1 | POST     "/api/chat"
time=2026-02-27T16:43:46.633+08:00 level=ERROR source=server.go:304 msg="llama runner terminated" error="exit status 1"
[GIN] 2026/02/27 - 16:43:47 | 200 |      1.3126ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:44:04 | 200 |    123.7783ms |       127.0.0.1 | POST     "/api/show"
[GIN] 2026/02/27 - 16:44:08 | 200 |    126.5724ms |       127.0.0.1 | POST     "/api/show"
[GIN] 2026/02/27 - 16:44:08 | 200 |    113.2701ms |       127.0.0.1 | POST     "/api/show"
time=2026-02-27T16:44:08.940+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 14367"
time=2026-02-27T16:44:09.104+08:00 level=INFO source=cpu_windows.go:148 msg=packages count=1
time=2026-02-27T16:44:09.104+08:00 level=INFO source=cpu_windows.go:195 msg="" package=0 cores=8 efficiency=0 threads=16
time=2026-02-27T16:44:09.123+08:00 level=INFO source=sched.go:689 msg="updated VRAM based on existing loaded models" gpu=GPU-22e15c16-4a8d-b130-9dda-d5c054898aee library=CUDA total="15.9 GiB" available="2.2 GiB"
time=2026-02-27T16:44:09.180+08:00 level=INFO source=server.go:247 msg="enabling flash attention"
time=2026-02-27T16:44:09.182+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --model C:\\Users\\LiYong\\.ollama\\models\\blobs\\sha256-2abd0d805943fa113f934d1ae4f2d5a749b5d4fe2a0a9c64b645c1df15868da7 --port 14372"
time=2026-02-27T16:44:09.190+08:00 level=INFO source=sched.go:491 msg="system memory" total="143.7 GiB" free="129.4 GiB" free_swap="135.3 GiB"
time=2026-02-27T16:44:09.191+08:00 level=INFO source=sched.go:498 msg="gpu memory" id=GPU-22e15c16-4a8d-b130-9dda-d5c054898aee library=CUDA available="1.7 GiB" free="2.2 GiB" minimum="457.0 MiB" overhead="0 B"
time=2026-02-27T16:44:09.191+08:00 level=INFO source=server.go:757 msg="loading model" "model layers"=41 requested=-1
time=2026-02-27T16:44:09.221+08:00 level=INFO source=runner.go:1411 msg="starting ollama engine"
time=2026-02-27T16:44:09.226+08:00 level=INFO source=runner.go:1446 msg="Server listening on 127.0.0.1:14372"
time=2026-02-27T16:44:09.234+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:41[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:41(0..40)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
time=2026-02-27T16:44:09.262+08:00 level=INFO source=ggml.go:136 msg="" architecture=qwen35moe file_type=Q4_K_M name="" description="" num_tensors=1959 num_key_values=57
load_backend: loaded CPU backend from C:\Users\LiYong\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 5070 Ti, compute capability 12.0, VMM: yes, ID: GPU-22e15c16-4a8d-b130-9dda-d5c054898aee
load_backend: loaded CUDA backend from C:\Users\LiYong\AppData\Local\Programs\Ollama\lib\ollama\cuda_v13\ggml-cuda.dll
time=2026-02-27T16:44:09.340+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)
time=2026-02-27T16:44:09.830+08:00 level=INFO source=server.go:1029 msg="model requires more gpu memory than is currently available, evicting a model to make space" "loaded layers"=2
time=2026-02-27T16:44:09.830+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:close LoraPath:[] Parallel:0 BatchSize:0 FlashAttention:Disabled KvSize:0 KvCacheType: NumThreads:0 GPULayers:[] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:240 msg="model weights" device=CUDA0 size="21.9 GiB"
time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:245 msg="model weights" device=CPU size="277.3 MiB"
time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:251 msg="kv cache" device=CUDA0 size="1.6 GiB"
time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:262 msg="compute graph" device=CUDA0 size="554.5 MiB"
time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:267 msg="compute graph" device=CPU size="15.8 MiB"
time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:272 msg="total memory" size="24.4 GiB"
time=2026-02-27T16:44:09.841+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 14379"
time=2026-02-27T16:44:10.234+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10027"
time=2026-02-27T16:44:10.485+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10032"
time=2026-02-27T16:44:10.735+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10037"
time=2026-02-27T16:44:10.985+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10042"
time=2026-02-27T16:44:11.235+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10057"
time=2026-02-27T16:44:11.484+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10062"
time=2026-02-27T16:44:11.735+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10067"
time=2026-02-27T16:44:11.985+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10072"
time=2026-02-27T16:44:12.236+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10077"
time=2026-02-27T16:44:12.485+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10088"
time=2026-02-27T16:44:12.735+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10101"
time=2026-02-27T16:44:12.985+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10113"
time=2026-02-27T16:44:13.235+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10118"
time=2026-02-27T16:44:13.486+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10123"
time=2026-02-27T16:44:13.734+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10128"
time=2026-02-27T16:44:13.987+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10133"
time=2026-02-27T16:44:14.235+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10138"
time=2026-02-27T16:44:14.486+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10145"
time=2026-02-27T16:44:14.735+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10150"
time=2026-02-27T16:44:14.984+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10155"
time=2026-02-27T16:44:14.993+08:00 level=INFO source=runner.go:464 msg="failure during GPU discovery" OLLAMA_LIBRARY_PATH="[C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\lib\\ollama C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\lib\\ollama\\cuda_v13]" extra_envs=map[] error="failed to finish discovery before timeout"
time=2026-02-27T16:44:14.993+08:00 level=WARN source=runner.go:356 msg="unable to refresh free memory, using old values"
time=2026-02-27T16:44:14.995+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10156"
time=2026-02-27T16:44:15.138+08:00 level=INFO source=cpu_windows.go:148 msg=packages count=1
time=2026-02-27T16:44:15.138+08:00 level=INFO source=cpu_windows.go:195 msg="" package=0 cores=8 efficiency=0 threads=16
time=2026-02-27T16:44:15.191+08:00 level=INFO source=sched.go:491 msg="system memory" total="143.7 GiB" free="128.6 GiB" free_swap="134.1 GiB"
time=2026-02-27T16:44:15.191+08:00 level=INFO source=sched.go:498 msg="gpu memory" id=GPU-22e15c16-4a8d-b130-9dda-d5c054898aee library=CUDA available="14.1 GiB" free="14.6 GiB" minimum="457.0 MiB" overhead="0 B"
time=2026-02-27T16:44:15.191+08:00 level=INFO source=server.go:757 msg="loading model" "model layers"=41 requested=-1
time=2026-02-27T16:44:15.191+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:24[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:24(16..39)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
time=2026-02-27T16:44:15.418+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:alloc LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:24[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:24(16..39)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:commit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:24[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:24(16..39)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=ggml.go:482 msg="offloading 24 repeating layers to GPU"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=ggml.go:486 msg="offloading output layer to CPU"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=ggml.go:494 msg="offloaded 24/41 layers to GPU"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:240 msg="model weights" device=CUDA0 size="12.2 GiB"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:245 msg="model weights" device=CPU size="10.0 GiB"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:251 msg="kv cache" device=CUDA0 size="990.2 MiB"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:256 msg="kv cache" device=CPU size="660.1 MiB"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:262 msg="compute graph" device=CUDA0 size="589.2 MiB"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:267 msg="compute graph" device=CPU size="630.8 MiB"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:272 msg="total memory" size="25.0 GiB"
time=2026-02-27T16:44:15.854+08:00 level=INFO source=sched.go:566 msg="loaded runners" count=1
time=2026-02-27T16:44:15.854+08:00 level=INFO source=server.go:1350 msg="waiting for llama runner to start responding"
time=2026-02-27T16:44:15.855+08:00 level=INFO source=server.go:1384 msg="waiting for server to become available" status="llm server loading model"
[GIN] 2026/02/27 - 16:44:17 | 200 |      1.6005ms |       127.0.0.1 | GET      "/api/tags"
time=2026-02-27T16:44:18.612+08:00 level=INFO source=server.go:1388 msg="llama runner started in 9.42 seconds"
CUDA error: invalid argument
  current device: 0, in function ggml_cuda_cpy at C:\a\ollama\ollama\ml\backend\ggml\ggml\src\ggml-cuda\cpy.cu:438
  cudaMemcpyAsyncReserve(src1_ddc, src0_ddc, ggml_nbytes(src0), cudaMemcpyDeviceToDevice, main_stream)
C:\a\ollama\ollama\ml\backend\ggml\ggml\src\ggml-cuda\ggml-cuda.cu:94: CUDA error
time=2026-02-27T16:44:18.825+08:00 level=ERROR source=server.go:1610 msg="post predict" error="Post \"http://127.0.0.1:14372/completion\": read tcp 127.0.0.1:14377->127.0.0.1:14372: wsarecv: An existing connection was forcibly closed by the remote host."
[GIN] 2026/02/27 - 16:44:18 | 500 |   10.0025406s |       127.0.0.1 | POST     "/api/chat"
time=2026-02-27T16:44:19.097+08:00 level=ERROR source=server.go:304 msg="llama runner terminated" error="exit status 1"
[GIN] 2026/02/27 - 16:44:47 | 200 |       530.3µs |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:45:17 | 200 |       1.017ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:45:47 | 200 |      1.2218ms |       127.0.0.1 | GET      "/api/tags"
[GIN] 2026/02/27 - 16:46:17 | 200 |      1.2302ms |       127.0.0.1 | GET      "/api/tags"

OS

Windows

GPU

Nvidia

CPU

AMD

Ollama version

v0.17.4/v0.17.3/v0.17.2/v0.17.1/v0.17.0

Originally created by @lyvs2012 on GitHub (Feb 27, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/14501 ### What is the issue? Issue Report: Qwen3.5 Model Fails with 500 Error on Ollama System Context: I can successfully run other models of similar and larger sizes on my local machine (e.g., Qwen3 48GB, GPT-OSS:120B 60GB, GLM-4.7-Flash 19GB). The hardware specifications are sufficient for heavy models. Problem Description: I am encountering a 500 Internal Server Error when attempting to use the Qwen3.5 model via the Ollama client. Reproduction Steps & Symptoms: Initial Run: The model loads and opens a conversation session successfully the very first time. Subsequent Runs: After the initial session, any further attempts to call the model or open a new dialog result in an immediate 500 error. Model Sizes: The local instances of these models are approximately 17GB and 23GB respectively. Question: Why is the Qwen3.5 model specifically failing with a 500 error after the first use, while other significantly larger models function normally? ### Relevant log output ```shell Couldn't find 'C:\Users\LiYong\.ollama\id_ed25519'. Generating new private key. Your new public key is: ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAILMrLVPjXm1m8rK8DKoebe3/5pXoygVk8dw2y8QZEsod time=2026-02-27T16:38:11.468+08:00 level=INFO source=routes.go:1663 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GGML_VK_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:0 OLLAMA_DEBUG:INFO OLLAMA_EDITOR: OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\\Users\\LiYong\\.ollama\\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NO_CLOUD:false OLLAMA_NUM_PARALLEL:1 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_REMOTES:[ollama.com] OLLAMA_SCHED_SPREAD:false OLLAMA_VULKAN:false ROCR_VISIBLE_DEVICES:]" time=2026-02-27T16:38:11.471+08:00 level=INFO source=routes.go:1665 msg="Ollama cloud disabled: false" time=2026-02-27T16:38:11.472+08:00 level=INFO source=images.go:473 msg="total blobs: 0" time=2026-02-27T16:38:11.472+08:00 level=INFO source=images.go:480 msg="total unused blobs removed: 0" time=2026-02-27T16:38:11.472+08:00 level=INFO source=routes.go:1718 msg="Listening on 127.0.0.1:11434 (version 0.17.4)" time=2026-02-27T16:38:11.473+08:00 level=INFO source=runner.go:67 msg="discovering available GPUs..." time=2026-02-27T16:38:11.483+08:00 level=INFO source=runner.go:106 msg="experimental Vulkan support disabled. To enable, set OLLAMA_VULKAN=1" time=2026-02-27T16:38:11.490+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 13703" time=2026-02-27T16:38:11.768+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 13708" time=2026-02-27T16:38:11.968+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 13713" time=2026-02-27T16:38:12.175+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 13720" time=2026-02-27T16:38:12.175+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 13721" time=2026-02-27T16:38:12.370+08:00 level=INFO source=types.go:42 msg="inference compute" id=GPU-22e15c16-4a8d-b130-9dda-d5c054898aee filter_id="" library=CUDA compute=12.0 name=CUDA0 description="NVIDIA GeForce RTX 5070 Ti" libdirs=ollama,cuda_v13 driver=13.1 pci_id=0000:01:00.0 type=discrete total="15.9 GiB" available="14.5 GiB" time=2026-02-27T16:38:12.370+08:00 level=INFO source=routes.go:1768 msg="vram-based default context" total_vram="15.9 GiB" default_num_ctx=4096 [GIN] 2026/02/27 - 16:38:12 | 200 | 0s | 127.0.0.1 | GET "/api/version" [GIN] 2026/02/27 - 16:39:09 | 200 | 0s | 127.0.0.1 | GET "/api/version" [GIN] 2026/02/27 - 16:39:09 | 200 | 0s | 127.0.0.1 | GET "/api/version" [GIN] 2026/02/27 - 16:39:09 | 200 | 2.1133ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:39:09 | 404 | 1.9031ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/02/27 - 16:39:10 | 401 | 388.1598ms | 127.0.0.1 | POST "/api/me" [GIN] 2026/02/27 - 16:39:10 | 401 | 511.7081ms | 127.0.0.1 | POST "/api/me" [GIN] 2026/02/27 - 16:39:14 | 200 | 130.5896ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/02/27 - 16:39:17 | 200 | 1.0376ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:39:17 | 200 | 121.9582ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/02/27 - 16:39:17 | 200 | 121.0927ms | 127.0.0.1 | POST "/api/show" time=2026-02-27T16:39:17.782+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 12315" time=2026-02-27T16:39:17.938+08:00 level=INFO source=cpu_windows.go:148 msg=packages count=1 time=2026-02-27T16:39:17.938+08:00 level=INFO source=cpu_windows.go:195 msg="" package=0 cores=8 efficiency=0 threads=16 time=2026-02-27T16:39:18.012+08:00 level=INFO source=server.go:247 msg="enabling flash attention" time=2026-02-27T16:39:18.013+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --model C:\\Users\\LiYong\\.ollama\\models\\blobs\\sha256-7935de6e08f9444536d0edcacf19d2166b34bef8ddb4ac7ce9263ff5cad0693b --port 12320" time=2026-02-27T16:39:18.024+08:00 level=INFO source=sched.go:491 msg="system memory" total="143.7 GiB" free="130.1 GiB" free_swap="135.4 GiB" time=2026-02-27T16:39:18.024+08:00 level=INFO source=sched.go:498 msg="gpu memory" id=GPU-22e15c16-4a8d-b130-9dda-d5c054898aee library=CUDA available="13.9 GiB" free="14.3 GiB" minimum="457.0 MiB" overhead="0 B" time=2026-02-27T16:39:18.024+08:00 level=INFO source=server.go:757 msg="loading model" "model layers"=65 requested=-1 time=2026-02-27T16:39:18.054+08:00 level=INFO source=runner.go:1411 msg="starting ollama engine" time=2026-02-27T16:39:18.058+08:00 level=INFO source=runner.go:1446 msg="Server listening on 127.0.0.1:12320" time=2026-02-27T16:39:18.066+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:65[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:65(0..64)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2026-02-27T16:39:18.095+08:00 level=INFO source=ggml.go:136 msg="" architecture=qwen35 file_type=Q4_K_M name="" description="" num_tensors=1307 num_key_values=53 load_backend: loaded CPU backend from C:\Users\LiYong\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 5070 Ti, compute capability 12.0, VMM: yes, ID: GPU-22e15c16-4a8d-b130-9dda-d5c054898aee load_backend: loaded CUDA backend from C:\Users\LiYong\AppData\Local\Programs\Ollama\lib\ollama\cuda_v13\ggml-cuda.dll time=2026-02-27T16:39:18.175+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) time=2026-02-27T16:39:18.644+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:48[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:48(16..63)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2026-02-27T16:39:18.886+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:alloc LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:48[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:48(16..63)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2026-02-27T16:39:19.548+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:commit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:48[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:48(16..63)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2026-02-27T16:39:19.548+08:00 level=INFO source=ggml.go:482 msg="offloading 48 repeating layers to GPU" time=2026-02-27T16:39:19.548+08:00 level=INFO source=ggml.go:486 msg="offloading output layer to CPU" time=2026-02-27T16:39:19.548+08:00 level=INFO source=ggml.go:494 msg="offloaded 48/65 layers to GPU" time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:240 msg="model weights" device=CUDA0 size="10.1 GiB" time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:245 msg="model weights" device=CPU size="6.2 GiB" time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:251 msg="kv cache" device=CUDA0 size="2.9 GiB" time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:256 msg="kv cache" device=CPU size="999.2 MiB" time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:262 msg="compute graph" device=CUDA0 size="799.3 MiB" time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:267 msg="compute graph" device=CPU size="1.0 GiB" time=2026-02-27T16:39:19.548+08:00 level=INFO source=device.go:272 msg="total memory" size="21.9 GiB" time=2026-02-27T16:39:19.548+08:00 level=INFO source=sched.go:566 msg="loaded runners" count=1 time=2026-02-27T16:39:19.549+08:00 level=INFO source=server.go:1350 msg="waiting for llama runner to start responding" time=2026-02-27T16:39:19.555+08:00 level=INFO source=server.go:1384 msg="waiting for server to become available" status="llm server loading model" time=2026-02-27T16:39:21.810+08:00 level=INFO source=server.go:1388 msg="llama runner started in 3.79 seconds" [GIN] 2026/02/27 - 16:39:47 | 200 | 1.5064ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:40:17 | 200 | 3.1242ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:40:47 | 200 | 1.5386ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:41:04 | 200 | 1m47s | 127.0.0.1 | POST "/api/chat" [GIN] 2026/02/27 - 16:41:17 | 200 | 1.0201ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:41:47 | 200 | 1.0384ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:42:17 | 200 | 1.0917ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:42:47 | 200 | 1.0375ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:43:17 | 200 | 1.0341ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:43:45 | 200 | 127.4648ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/02/27 - 16:43:45 | 200 | 115.6221ms | 127.0.0.1 | POST "/api/show" CUDA error: invalid argument current device: 0, in function ggml_cuda_cpy at C:\a\ollama\ollama\ml\backend\ggml\ggml\src\ggml-cuda\cpy.cu:438 cudaMemcpyAsyncReserve(src1_ddc, src0_ddc, ggml_nbytes(src0), cudaMemcpyDeviceToDevice, main_stream) C:\a\ollama\ollama\ml\backend\ggml\ggml\src\ggml-cuda\ggml-cuda.cu:94: CUDA error time=2026-02-27T16:43:46.061+08:00 level=ERROR source=server.go:1610 msg="post predict" error="Post \"http://127.0.0.1:12320/completion\": read tcp 127.0.0.1:14337->127.0.0.1:12320: wsarecv: An existing connection was forcibly closed by the remote host." [GIN] 2026/02/27 - 16:43:46 | 500 | 458.4568ms | 127.0.0.1 | POST "/api/chat" time=2026-02-27T16:43:46.633+08:00 level=ERROR source=server.go:304 msg="llama runner terminated" error="exit status 1" [GIN] 2026/02/27 - 16:43:47 | 200 | 1.3126ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:44:04 | 200 | 123.7783ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/02/27 - 16:44:08 | 200 | 126.5724ms | 127.0.0.1 | POST "/api/show" [GIN] 2026/02/27 - 16:44:08 | 200 | 113.2701ms | 127.0.0.1 | POST "/api/show" time=2026-02-27T16:44:08.940+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 14367" time=2026-02-27T16:44:09.104+08:00 level=INFO source=cpu_windows.go:148 msg=packages count=1 time=2026-02-27T16:44:09.104+08:00 level=INFO source=cpu_windows.go:195 msg="" package=0 cores=8 efficiency=0 threads=16 time=2026-02-27T16:44:09.123+08:00 level=INFO source=sched.go:689 msg="updated VRAM based on existing loaded models" gpu=GPU-22e15c16-4a8d-b130-9dda-d5c054898aee library=CUDA total="15.9 GiB" available="2.2 GiB" time=2026-02-27T16:44:09.180+08:00 level=INFO source=server.go:247 msg="enabling flash attention" time=2026-02-27T16:44:09.182+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --model C:\\Users\\LiYong\\.ollama\\models\\blobs\\sha256-2abd0d805943fa113f934d1ae4f2d5a749b5d4fe2a0a9c64b645c1df15868da7 --port 14372" time=2026-02-27T16:44:09.190+08:00 level=INFO source=sched.go:491 msg="system memory" total="143.7 GiB" free="129.4 GiB" free_swap="135.3 GiB" time=2026-02-27T16:44:09.191+08:00 level=INFO source=sched.go:498 msg="gpu memory" id=GPU-22e15c16-4a8d-b130-9dda-d5c054898aee library=CUDA available="1.7 GiB" free="2.2 GiB" minimum="457.0 MiB" overhead="0 B" time=2026-02-27T16:44:09.191+08:00 level=INFO source=server.go:757 msg="loading model" "model layers"=41 requested=-1 time=2026-02-27T16:44:09.221+08:00 level=INFO source=runner.go:1411 msg="starting ollama engine" time=2026-02-27T16:44:09.226+08:00 level=INFO source=runner.go:1446 msg="Server listening on 127.0.0.1:14372" time=2026-02-27T16:44:09.234+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:41[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:41(0..40)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2026-02-27T16:44:09.262+08:00 level=INFO source=ggml.go:136 msg="" architecture=qwen35moe file_type=Q4_K_M name="" description="" num_tensors=1959 num_key_values=57 load_backend: loaded CPU backend from C:\Users\LiYong\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 5070 Ti, compute capability 12.0, VMM: yes, ID: GPU-22e15c16-4a8d-b130-9dda-d5c054898aee load_backend: loaded CUDA backend from C:\Users\LiYong\AppData\Local\Programs\Ollama\lib\ollama\cuda_v13\ggml-cuda.dll time=2026-02-27T16:44:09.340+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) time=2026-02-27T16:44:09.830+08:00 level=INFO source=server.go:1029 msg="model requires more gpu memory than is currently available, evicting a model to make space" "loaded layers"=2 time=2026-02-27T16:44:09.830+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:close LoraPath:[] Parallel:0 BatchSize:0 FlashAttention:Disabled KvSize:0 KvCacheType: NumThreads:0 GPULayers:[] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:240 msg="model weights" device=CUDA0 size="21.9 GiB" time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:245 msg="model weights" device=CPU size="277.3 MiB" time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:251 msg="kv cache" device=CUDA0 size="1.6 GiB" time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:262 msg="compute graph" device=CUDA0 size="554.5 MiB" time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:267 msg="compute graph" device=CPU size="15.8 MiB" time=2026-02-27T16:44:09.830+08:00 level=INFO source=device.go:272 msg="total memory" size="24.4 GiB" time=2026-02-27T16:44:09.841+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 14379" time=2026-02-27T16:44:10.234+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10027" time=2026-02-27T16:44:10.485+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10032" time=2026-02-27T16:44:10.735+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10037" time=2026-02-27T16:44:10.985+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10042" time=2026-02-27T16:44:11.235+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10057" time=2026-02-27T16:44:11.484+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10062" time=2026-02-27T16:44:11.735+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10067" time=2026-02-27T16:44:11.985+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10072" time=2026-02-27T16:44:12.236+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10077" time=2026-02-27T16:44:12.485+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10088" time=2026-02-27T16:44:12.735+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10101" time=2026-02-27T16:44:12.985+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10113" time=2026-02-27T16:44:13.235+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10118" time=2026-02-27T16:44:13.486+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10123" time=2026-02-27T16:44:13.734+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10128" time=2026-02-27T16:44:13.987+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10133" time=2026-02-27T16:44:14.235+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10138" time=2026-02-27T16:44:14.486+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10145" time=2026-02-27T16:44:14.735+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10150" time=2026-02-27T16:44:14.984+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10155" time=2026-02-27T16:44:14.993+08:00 level=INFO source=runner.go:464 msg="failure during GPU discovery" OLLAMA_LIBRARY_PATH="[C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\lib\\ollama C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\lib\\ollama\\cuda_v13]" extra_envs=map[] error="failed to finish discovery before timeout" time=2026-02-27T16:44:14.993+08:00 level=WARN source=runner.go:356 msg="unable to refresh free memory, using old values" time=2026-02-27T16:44:14.995+08:00 level=INFO source=server.go:431 msg="starting runner" cmd="C:\\Users\\LiYong\\AppData\\Local\\Programs\\Ollama\\ollama.exe runner --ollama-engine --port 10156" time=2026-02-27T16:44:15.138+08:00 level=INFO source=cpu_windows.go:148 msg=packages count=1 time=2026-02-27T16:44:15.138+08:00 level=INFO source=cpu_windows.go:195 msg="" package=0 cores=8 efficiency=0 threads=16 time=2026-02-27T16:44:15.191+08:00 level=INFO source=sched.go:491 msg="system memory" total="143.7 GiB" free="128.6 GiB" free_swap="134.1 GiB" time=2026-02-27T16:44:15.191+08:00 level=INFO source=sched.go:498 msg="gpu memory" id=GPU-22e15c16-4a8d-b130-9dda-d5c054898aee library=CUDA available="14.1 GiB" free="14.6 GiB" minimum="457.0 MiB" overhead="0 B" time=2026-02-27T16:44:15.191+08:00 level=INFO source=server.go:757 msg="loading model" "model layers"=41 requested=-1 time=2026-02-27T16:44:15.191+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:24[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:24(16..39)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2026-02-27T16:44:15.418+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:alloc LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:24[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:24(16..39)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2026-02-27T16:44:15.854+08:00 level=INFO source=runner.go:1284 msg=load request="{Operation:commit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:4096 KvCacheType: NumThreads:8 GPULayers:24[ID:GPU-22e15c16-4a8d-b130-9dda-d5c054898aee Layers:24(16..39)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" time=2026-02-27T16:44:15.854+08:00 level=INFO source=ggml.go:482 msg="offloading 24 repeating layers to GPU" time=2026-02-27T16:44:15.854+08:00 level=INFO source=ggml.go:486 msg="offloading output layer to CPU" time=2026-02-27T16:44:15.854+08:00 level=INFO source=ggml.go:494 msg="offloaded 24/41 layers to GPU" time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:240 msg="model weights" device=CUDA0 size="12.2 GiB" time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:245 msg="model weights" device=CPU size="10.0 GiB" time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:251 msg="kv cache" device=CUDA0 size="990.2 MiB" time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:256 msg="kv cache" device=CPU size="660.1 MiB" time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:262 msg="compute graph" device=CUDA0 size="589.2 MiB" time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:267 msg="compute graph" device=CPU size="630.8 MiB" time=2026-02-27T16:44:15.854+08:00 level=INFO source=device.go:272 msg="total memory" size="25.0 GiB" time=2026-02-27T16:44:15.854+08:00 level=INFO source=sched.go:566 msg="loaded runners" count=1 time=2026-02-27T16:44:15.854+08:00 level=INFO source=server.go:1350 msg="waiting for llama runner to start responding" time=2026-02-27T16:44:15.855+08:00 level=INFO source=server.go:1384 msg="waiting for server to become available" status="llm server loading model" [GIN] 2026/02/27 - 16:44:17 | 200 | 1.6005ms | 127.0.0.1 | GET "/api/tags" time=2026-02-27T16:44:18.612+08:00 level=INFO source=server.go:1388 msg="llama runner started in 9.42 seconds" CUDA error: invalid argument current device: 0, in function ggml_cuda_cpy at C:\a\ollama\ollama\ml\backend\ggml\ggml\src\ggml-cuda\cpy.cu:438 cudaMemcpyAsyncReserve(src1_ddc, src0_ddc, ggml_nbytes(src0), cudaMemcpyDeviceToDevice, main_stream) C:\a\ollama\ollama\ml\backend\ggml\ggml\src\ggml-cuda\ggml-cuda.cu:94: CUDA error time=2026-02-27T16:44:18.825+08:00 level=ERROR source=server.go:1610 msg="post predict" error="Post \"http://127.0.0.1:14372/completion\": read tcp 127.0.0.1:14377->127.0.0.1:14372: wsarecv: An existing connection was forcibly closed by the remote host." [GIN] 2026/02/27 - 16:44:18 | 500 | 10.0025406s | 127.0.0.1 | POST "/api/chat" time=2026-02-27T16:44:19.097+08:00 level=ERROR source=server.go:304 msg="llama runner terminated" error="exit status 1" [GIN] 2026/02/27 - 16:44:47 | 200 | 530.3µs | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:45:17 | 200 | 1.017ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:45:47 | 200 | 1.2218ms | 127.0.0.1 | GET "/api/tags" [GIN] 2026/02/27 - 16:46:17 | 200 | 1.2302ms | 127.0.0.1 | GET "/api/tags" ``` ### OS Windows ### GPU Nvidia ### CPU AMD ### Ollama version v0.17.4/v0.17.3/v0.17.2/v0.17.1/v0.17.0
GiteaMirror added the bug label 2026-04-29 09:57:01 -05:00
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#55920
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?