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

[GH-ISSUE #13618] Qwen3VL issue on AMD GPU / Rocm #8961

New Issue
Closed
opened 2026-04-12 21:47:37 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-04-12 21:47:37 -05:00
Owner
Copy Link

Originally created by @FR-Mister-T on GitHub (Jan 4, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/13618

What is the issue?

I'm running ubuntu24.04/python 3.13.11/ the lastest ollama: rocm 7.0.2 (the most stable for my setup currently)

There is a bug in Ollama's Qwen3VL vision model implementation. The Conv3D layer is trying to access a nil pointer when processing the vision model components.

I tested both directly in terminal and throught comfyui OllamaConnectivityV2 node with the same issue
Trying to remake the modelfile with mmproj failed

model is available here https://huggingface.co/Phr00t/Qwen3-VL-32B-Instruct-heretic-v2-iQ5KS-GGUF

Relevant log output

zeuss194@zeuss194-ThinkStation-P520:~$ journalctl -u ollama -n 50 --no-pager
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: constructing llama_context
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_seq_max     = 1
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_ctx         = 4096
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_ctx_seq     = 4096
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_batch       = 512
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_ubatch      = 512
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: causal_attn   = 1
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: flash_attn    = auto
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: kv_unified    = false
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: freq_base     = 100000000.0
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: freq_scale    = 1
Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_ctx_seq (4096) < n_ctx_train (32768) -- the full capacity of the model will not be utilized
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context:  ROCm_Host  output buffer size =     0.52 MiB
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_kv_cache:      ROCm0 KV buffer size =   640.00 MiB
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_kv_cache: size =  640.00 MiB (  4096 cells,  40 layers,  1/1 seqs), K (f16):  320.00 MiB, V (f16):  320.00 MiB
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: Flash Attention was auto, set to enabled
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context:      ROCm0 compute buffer size =   266.00 MiB
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context:  ROCm_Host compute buffer size =    18.01 MiB
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: graph nodes  = 1247
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: graph splits = 2
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:02:36.449+01:00 level=INFO source=server.go:1376 msg="llama runner started in 6.05 seconds"
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:02:36.449+01:00 level=INFO source=sched.go:517 msg="loaded runners" count=1
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:02:36.449+01:00 level=INFO source=server.go:1338 msg="waiting for llama runner to start responding"
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:02:36.450+01:00 level=INFO source=server.go:1376 msg="llama runner started in 6.05 seconds"
Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: check_double_bos_eos: Added a BOS token to the prompt as specified by the model but the prompt also starts with a BOS token. So now the final prompt starts with 2 BOS tokens. Are you sure this is what you want?
Jan 04 06:03:47 zeuss194-ThinkStation-P520 ollama[352603]: [GIN] 2026/01/04 - 06:03:47 | 200 |         1m18s |       127.0.0.1 | POST     "/api/chat"
Jan 04 06:08:47 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:08:47.058+01:00 level=INFO source=server.go:429 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 35401"
Jan 04 06:08:48 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:08:48.881+01:00 level=INFO source=server.go:429 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 34445"
Jan 04 06:12:10 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:10.557+01:00 level=INFO source=server.go:429 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 41917"
Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.195+01:00 level=INFO source=server.go:245 msg="enabling flash attention"
Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.195+01:00 level=INFO source=server.go:429 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --model /usr/share/ollama/.ollama/models/blobs/sha256-7e8c3957c3ed7475b652ae1c66848fb98265e3c1c31d127c927eb99532e0bdfa --port 38935"
Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.196+01:00 level=INFO source=sched.go:443 msg="system memory" total="62.4 GiB" free="54.8 GiB" free_swap="5.6 GiB"
Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.196+01:00 level=INFO source=sched.go:450 msg="gpu memory" id=GPU-5c5383ed20b3fc7d library=ROCm available="30.9 GiB" free="31.4 GiB" minimum="457.0 MiB" overhead="0 B"
Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.196+01:00 level=INFO source=server.go:746 msg="loading model" "model layers"=65 requested=-1
Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.212+01:00 level=INFO source=runner.go:1405 msg="starting ollama engine"
Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.212+01:00 level=INFO source=runner.go:1440 msg="Server listening on 127.0.0.1:38935"
Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.218+01:00 level=INFO source=runner.go:1278 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:8192 KvCacheType: NumThreads:8 GPULayers:65[ID:GPU-5c5383ed20b3fc7d Layers:65(0..64)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}"
Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.255+01:00 level=INFO source=ggml.go:136 msg="" architecture=qwen3vl file_type=Q5_K_S name="Q3 Vl H2" description="" num_tensors=707 num_key_values=39
Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: load_backend: loaded CPU backend from /usr/local/lib/ollama/libggml-cpu-skylakex.so
Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: ggml_cuda_init: GGML_CUDA_FORCE_MMQ:    no
Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: ggml_cuda_init: found 1 ROCm devices:
Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]:   Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32, ID: GPU-5c5383ed20b3fc7d
Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: load_backend: loaded ROCm backend from /usr/local/lib/ollama/rocm/libggml-hip.so
Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:13.798+01:00 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.LLAMAFILE=1 CPU.1.LLAMAFILE=1 ROCm.0.NO_VMM=1 ROCm.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(gcc)
Jan 04 06:12:14 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:14.197+01:00 level=INFO source=server.go:3634 msg="http: panic serving 127.0.0.1:54742: runtime error: invalid memory address or nil pointer dereference\ngoroutine 13 [running]:\nnet/http.(*conn).serve.func1()\n\tnet/http/server.go:1947 +0xbe\npanic({0x583da72782c0?, 0x583da7c00410?})\n\truntime/panic.go:792 +0x132\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).allocModel.func1()\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1187 +0x11a\npanic({0x583da72782c0?, 0x583da7c00410?})\n\truntime/panic.go:792 +0x132\ngithub.com/ollama/ollama/ml/nn.(*Conv3D).Forward(0x0, {0x583da73f4250, 0xc000f57d80}, {0x583da73feb20?, 0xc000f68048?}, 0x10?, 0xc000600808?, 0xc000667840?, 0xc000047190?, 0x0, ...)\n\tgithub.com/ollama/ollama/ml/nn/convolution.go:25 +0x3a\ngithub.com/ollama/ollama/model/models/qwen3vl.(*VisionModel).Forward(0xc00013c0c0, {0x583da73f4250, 0xc000f57d80}, {0x583da73feb20, 0xc000f68030}, 0xc000f40000)\n\tgithub.com/ollama/ollama/model/models/qwen3vl/model_vision.go:224 +0x118\ngithub.com/ollama/ollama/model/models/qwen3vl.(*Model).EncodeMultimodal(0xc000d7e270, {0x583da73f4250, 0xc000f57d80}, {0xc001cc8000, 0x400436, 0x700000})\n\tgithub.com/ollama/ollama/model/models/qwen3vl/model.go:43 +0x14e\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).reserveWorstCaseGraph(0xc000236f00, 0x1)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1098 +0x34e\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).allocModel(0xc000236f00, {0x7fffc5285d04?, 0x583da612841a?}, {0x0, 0x8, {0xc000714140, 0x1, 0x1}, 0x1}, {0x0, ...}, ...)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1226 +0x391\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).load(0xc000236f00, {0x583da73e6fa0, 0xc00078e000}, 0xc000788000)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1305 +0x54b\nnet/http.HandlerFunc.ServeHTTP(0xc00013d5c0?, {0x583da73e6fa0?, 0xc00078e000?}, 0xc000555b60?)\n\tnet/http/server.go:2294 +0x29\nnet/http.(*ServeMux).ServeHTTP(0x583da5dd88c5?, {0x583da73e6fa0, 0xc00078e000}, 0xc000788000)\n\tnet/http/server.go:2822 +0x1c4\nnet/http.serverHandler.ServeHTTP({0x583da73e3590?}, {0x583da73e6fa0?, 0xc00078e000?}, 0x1?)\n\tnet/http/server.go:3301 +0x8e\nnet/http.(*conn).serve(0xc0005203f0, {0x583da73e93d8, 0xc000717140})\n\tnet/http/server.go:2102 +0x625\ncreated by net/http.(*Server).Serve in goroutine 1\n\tnet/http/server.go:3454 +0x485"
Jan 04 06:12:14 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:14.198+01:00 level=INFO source=runner.go:1278 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}"
Jan 04 06:12:14 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:14.198+01:00 level=INFO source=sched.go:470 msg="Load failed" model=/usr/share/ollama/.ollama/models/blobs/sha256-7e8c3957c3ed7475b652ae1c66848fb98265e3c1c31d127c927eb99532e0bdfa error="do load request: Post \"http://127.0.0.1:38935/load\": EOF"
Jan 04 06:12:14 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:14.240+01:00 level=ERROR source=server.go:302 msg="llama runner terminated" error="signal: killed"
Jan 04 06:12:14 zeuss194-ThinkStation-P520 ollama[352603]: [GIN] 2026/01/04 - 06:12:14 | 500 |  3.787155736s |       127.0.0.1 | POST     "/api/chat"

OS

Linux

GPU

AMD

CPU

Intel

Ollama version

0.13.5

Originally created by @FR-Mister-T on GitHub (Jan 4, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/13618 ### What is the issue? I'm running ubuntu24.04/python 3.13.11/ the lastest ollama: rocm 7.0.2 (the most stable for my setup currently) There is a bug in Ollama's Qwen3VL vision model implementation. The Conv3D layer is trying to access a nil pointer when processing the vision model components. I tested both directly in terminal and throught comfyui OllamaConnectivityV2 node with the same issue Trying to remake the modelfile with mmproj failed model is available here https://huggingface.co/Phr00t/Qwen3-VL-32B-Instruct-heretic-v2-iQ5KS-GGUF ### Relevant log output ```shell zeuss194@zeuss194-ThinkStation-P520:~$ journalctl -u ollama -n 50 --no-pager Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: constructing llama_context Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_seq_max = 1 Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_ctx = 4096 Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_ctx_seq = 4096 Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_batch = 512 Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_ubatch = 512 Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: causal_attn = 1 Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: flash_attn = auto Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: kv_unified = false Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: freq_base = 100000000.0 Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: freq_scale = 1 Jan 04 06:02:35 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: n_ctx_seq (4096) < n_ctx_train (32768) -- the full capacity of the model will not be utilized Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: ROCm_Host output buffer size = 0.52 MiB Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_kv_cache: ROCm0 KV buffer size = 640.00 MiB Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_kv_cache: size = 640.00 MiB ( 4096 cells, 40 layers, 1/1 seqs), K (f16): 320.00 MiB, V (f16): 320.00 MiB Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: Flash Attention was auto, set to enabled Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: ROCm0 compute buffer size = 266.00 MiB Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: ROCm_Host compute buffer size = 18.01 MiB Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: graph nodes = 1247 Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: llama_context: graph splits = 2 Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:02:36.449+01:00 level=INFO source=server.go:1376 msg="llama runner started in 6.05 seconds" Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:02:36.449+01:00 level=INFO source=sched.go:517 msg="loaded runners" count=1 Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:02:36.449+01:00 level=INFO source=server.go:1338 msg="waiting for llama runner to start responding" Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:02:36.450+01:00 level=INFO source=server.go:1376 msg="llama runner started in 6.05 seconds" Jan 04 06:02:36 zeuss194-ThinkStation-P520 ollama[352603]: check_double_bos_eos: Added a BOS token to the prompt as specified by the model but the prompt also starts with a BOS token. So now the final prompt starts with 2 BOS tokens. Are you sure this is what you want? Jan 04 06:03:47 zeuss194-ThinkStation-P520 ollama[352603]: [GIN] 2026/01/04 - 06:03:47 | 200 | 1m18s | 127.0.0.1 | POST "/api/chat" Jan 04 06:08:47 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:08:47.058+01:00 level=INFO source=server.go:429 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 35401" Jan 04 06:08:48 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:08:48.881+01:00 level=INFO source=server.go:429 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 34445" Jan 04 06:12:10 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:10.557+01:00 level=INFO source=server.go:429 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --port 41917" Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.195+01:00 level=INFO source=server.go:245 msg="enabling flash attention" Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.195+01:00 level=INFO source=server.go:429 msg="starting runner" cmd="/usr/local/bin/ollama runner --ollama-engine --model /usr/share/ollama/.ollama/models/blobs/sha256-7e8c3957c3ed7475b652ae1c66848fb98265e3c1c31d127c927eb99532e0bdfa --port 38935" Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.196+01:00 level=INFO source=sched.go:443 msg="system memory" total="62.4 GiB" free="54.8 GiB" free_swap="5.6 GiB" Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.196+01:00 level=INFO source=sched.go:450 msg="gpu memory" id=GPU-5c5383ed20b3fc7d library=ROCm available="30.9 GiB" free="31.4 GiB" minimum="457.0 MiB" overhead="0 B" Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.196+01:00 level=INFO source=server.go:746 msg="loading model" "model layers"=65 requested=-1 Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.212+01:00 level=INFO source=runner.go:1405 msg="starting ollama engine" Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.212+01:00 level=INFO source=runner.go:1440 msg="Server listening on 127.0.0.1:38935" Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.218+01:00 level=INFO source=runner.go:1278 msg=load request="{Operation:fit LoraPath:[] Parallel:1 BatchSize:512 FlashAttention:Enabled KvSize:8192 KvCacheType: NumThreads:8 GPULayers:65[ID:GPU-5c5383ed20b3fc7d Layers:65(0..64)] MultiUserCache:false ProjectorPath: MainGPU:0 UseMmap:false}" Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:12.255+01:00 level=INFO source=ggml.go:136 msg="" architecture=qwen3vl file_type=Q5_K_S name="Q3 Vl H2" description="" num_tensors=707 num_key_values=39 Jan 04 06:12:12 zeuss194-ThinkStation-P520 ollama[352603]: load_backend: loaded CPU backend from /usr/local/lib/ollama/libggml-cpu-skylakex.so Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: ggml_cuda_init: found 1 ROCm devices: Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32, ID: GPU-5c5383ed20b3fc7d Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: load_backend: loaded ROCm backend from /usr/local/lib/ollama/rocm/libggml-hip.so Jan 04 06:12:13 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:13.798+01:00 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.LLAMAFILE=1 CPU.1.LLAMAFILE=1 ROCm.0.NO_VMM=1 ROCm.0.PEER_MAX_BATCH_SIZE=128 compiler=cgo(gcc) Jan 04 06:12:14 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:14.197+01:00 level=INFO source=server.go:3634 msg="http: panic serving 127.0.0.1:54742: runtime error: invalid memory address or nil pointer dereference\ngoroutine 13 [running]:\nnet/http.(*conn).serve.func1()\n\tnet/http/server.go:1947 +0xbe\npanic({0x583da72782c0?, 0x583da7c00410?})\n\truntime/panic.go:792 +0x132\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).allocModel.func1()\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1187 +0x11a\npanic({0x583da72782c0?, 0x583da7c00410?})\n\truntime/panic.go:792 +0x132\ngithub.com/ollama/ollama/ml/nn.(*Conv3D).Forward(0x0, {0x583da73f4250, 0xc000f57d80}, {0x583da73feb20?, 0xc000f68048?}, 0x10?, 0xc000600808?, 0xc000667840?, 0xc000047190?, 0x0, ...)\n\tgithub.com/ollama/ollama/ml/nn/convolution.go:25 +0x3a\ngithub.com/ollama/ollama/model/models/qwen3vl.(*VisionModel).Forward(0xc00013c0c0, {0x583da73f4250, 0xc000f57d80}, {0x583da73feb20, 0xc000f68030}, 0xc000f40000)\n\tgithub.com/ollama/ollama/model/models/qwen3vl/model_vision.go:224 +0x118\ngithub.com/ollama/ollama/model/models/qwen3vl.(*Model).EncodeMultimodal(0xc000d7e270, {0x583da73f4250, 0xc000f57d80}, {0xc001cc8000, 0x400436, 0x700000})\n\tgithub.com/ollama/ollama/model/models/qwen3vl/model.go:43 +0x14e\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).reserveWorstCaseGraph(0xc000236f00, 0x1)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1098 +0x34e\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).allocModel(0xc000236f00, {0x7fffc5285d04?, 0x583da612841a?}, {0x0, 0x8, {0xc000714140, 0x1, 0x1}, 0x1}, {0x0, ...}, ...)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1226 +0x391\ngithub.com/ollama/ollama/runner/ollamarunner.(*Server).load(0xc000236f00, {0x583da73e6fa0, 0xc00078e000}, 0xc000788000)\n\tgithub.com/ollama/ollama/runner/ollamarunner/runner.go:1305 +0x54b\nnet/http.HandlerFunc.ServeHTTP(0xc00013d5c0?, {0x583da73e6fa0?, 0xc00078e000?}, 0xc000555b60?)\n\tnet/http/server.go:2294 +0x29\nnet/http.(*ServeMux).ServeHTTP(0x583da5dd88c5?, {0x583da73e6fa0, 0xc00078e000}, 0xc000788000)\n\tnet/http/server.go:2822 +0x1c4\nnet/http.serverHandler.ServeHTTP({0x583da73e3590?}, {0x583da73e6fa0?, 0xc00078e000?}, 0x1?)\n\tnet/http/server.go:3301 +0x8e\nnet/http.(*conn).serve(0xc0005203f0, {0x583da73e93d8, 0xc000717140})\n\tnet/http/server.go:2102 +0x625\ncreated by net/http.(*Server).Serve in goroutine 1\n\tnet/http/server.go:3454 +0x485" Jan 04 06:12:14 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:14.198+01:00 level=INFO source=runner.go:1278 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}" Jan 04 06:12:14 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:14.198+01:00 level=INFO source=sched.go:470 msg="Load failed" model=/usr/share/ollama/.ollama/models/blobs/sha256-7e8c3957c3ed7475b652ae1c66848fb98265e3c1c31d127c927eb99532e0bdfa error="do load request: Post \"http://127.0.0.1:38935/load\": EOF" Jan 04 06:12:14 zeuss194-ThinkStation-P520 ollama[352603]: time=2026-01-04T06:12:14.240+01:00 level=ERROR source=server.go:302 msg="llama runner terminated" error="signal: killed" Jan 04 06:12:14 zeuss194-ThinkStation-P520 ollama[352603]: [GIN] 2026/01/04 - 06:12:14 | 500 | 3.787155736s | 127.0.0.1 | POST "/api/chat" ``` ### OS Linux ### GPU AMD ### CPU Intel ### Ollama version 0.13.5
GiteaMirror added the needs more infobug labels 2026-04-12 21:47:37 -05:00
GiteaMirror commented 2026-04-12 21:47:39 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 4, 2026):

Does it work with the original model?

<!-- gh-comment-id:3707774126 --> @rick-github commented on GitHub (Jan 4, 2026): Does it work with the [original model](https://ollama.com/library/qwen3-vl:32b)?
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 needs more info
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#8961
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?