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 #10454] unknown model architecture: 'qwen3' #68931

New Issue
Closed
opened 2026-05-04 15:53:12 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-05-04 15:53:12 -05:00
Owner
Copy Link

Originally created by @cam-narzt on GitHub (Apr 29, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10454

What is the issue?

when I try to use the qwen3 model in the latest tagged ollama docker image I get the error: llama_model_load: error loading model: error loading model architecture: unknown model architecture: 'qwen3'.

Relevant log output

ollama            | time=2025-04-29T00:37:26.067Z level=WARN source=ggml.go:152 msg="key not found" key=qwen3.vision.block_count default=0
ollama            | time=2025-04-29T00:37:26.067Z level=WARN source=ggml.go:152 msg="key not found" key=qwen3.vision.block_count default=0
ollama            | time=2025-04-29T00:37:26.068Z level=WARN source=ggml.go:152 msg="key not found" key=qwen3.vision.block_count default=0
ollama            | time=2025-04-29T00:37:26.068Z level=WARN source=ggml.go:152 msg="key not found" key=qwen3.vision.block_count default=0
ollama            | time=2025-04-29T00:37:26.128Z level=INFO source=server.go:105 msg="system memory" total="62.0 GiB" free="55.3 GiB" free_swap="7.9 GiB"
ollama            | time=2025-04-29T00:37:26.128Z level=WARN source=ggml.go:152 msg="key not found" key=qwen3.vision.block_count default=0
ollama            | time=2025-04-29T00:37:26.129Z level=INFO source=server.go:138 msg=offload library=cuda layers.requested=-1 layers.model=41 layers.offload=15 layers.split="" memory.available="[11.5 GiB]" memory.gpu_overhead="0 B" memory.required.full="20.8 GiB" memory.required.partial="11.2 GiB" memory.required.kv="6.2 GiB" memory.required.allocations="[11.2 GiB]" memory.weights.total="8.2 GiB" memory.weights.repeating="7.6 GiB" memory.weights.nonrepeating="608.6 MiB" memory.graph.full="5.2 GiB" memory.graph.partial="5.2 GiB"
ollama            | llama_model_loader: loaded meta data with 27 key-value pairs and 443 tensors from /root/.ollama/models/blobs/sha256-a8cc1361f3145dc01f6d77c6c82c9116b9ffe3c97b34716fe20418455876c40e (version GGUF V3 (latest))
ollama            | llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
ollama            | llama_model_loader: - kv   0:                       general.architecture str              = qwen3
ollama            | llama_model_loader: - kv   1:                               general.type str              = model
ollama            | llama_model_loader: - kv   2:                               general.name str              = Qwen3 14B
ollama            | llama_model_loader: - kv   3:                           general.basename str              = Qwen3
ollama            | llama_model_loader: - kv   4:                         general.size_label str              = 14B
ollama            | llama_model_loader: - kv   5:                          qwen3.block_count u32              = 40
ollama            | llama_model_loader: - kv   6:                       qwen3.context_length u32              = 40960
ollama            | llama_model_loader: - kv   7:                     qwen3.embedding_length u32              = 5120
ollama            | llama_model_loader: - kv   8:                  qwen3.feed_forward_length u32              = 17408
ollama            | llama_model_loader: - kv   9:                 qwen3.attention.head_count u32              = 40
ollama            | llama_model_loader: - kv  10:              qwen3.attention.head_count_kv u32              = 8
ollama            | llama_model_loader: - kv  11:                       qwen3.rope.freq_base f32              = 1000000.000000
ollama            | llama_model_loader: - kv  12:     qwen3.attention.layer_norm_rms_epsilon f32              = 0.000001
ollama            | llama_model_loader: - kv  13:                 qwen3.attention.key_length u32              = 128
ollama            | llama_model_loader: - kv  14:               qwen3.attention.value_length u32              = 128
ollama            | llama_model_loader: - kv  15:                       tokenizer.ggml.model str              = gpt2
ollama            | llama_model_loader: - kv  16:                         tokenizer.ggml.pre str              = qwen2
ollama            | llama_model_loader: - kv  17:                      tokenizer.ggml.tokens arr[str,151936]  = ["!", "\"", "#", "$", "%", "&", "'", ...
ollama            | llama_model_loader: - kv  18:                  tokenizer.ggml.token_type arr[i32,151936]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
ollama            | llama_model_loader: - kv  19:                      tokenizer.ggml.merges arr[str,151387]  = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
ollama            | llama_model_loader: - kv  20:                tokenizer.ggml.eos_token_id u32              = 151645
ollama            | llama_model_loader: - kv  21:            tokenizer.ggml.padding_token_id u32              = 151643
ollama            | llama_model_loader: - kv  22:                tokenizer.ggml.bos_token_id u32              = 151643
ollama            | llama_model_loader: - kv  23:               tokenizer.ggml.add_bos_token bool             = false
ollama            | llama_model_loader: - kv  24:                    tokenizer.chat_template str              = {%- if tools %}\n    {{- '<|im_start|>...
ollama            | llama_model_loader: - kv  25:               general.quantization_version u32              = 2
ollama            | llama_model_loader: - kv  26:                          general.file_type u32              = 15
ollama            | llama_model_loader: - type  f32:  161 tensors
ollama            | llama_model_loader: - type  f16:   40 tensors
ollama            | llama_model_loader: - type q4_K:  221 tensors
ollama            | llama_model_loader: - type q6_K:   21 tensors
ollama            | print_info: file format = GGUF V3 (latest)
ollama            | print_info: file type   = Q4_K - Medium
ollama            | print_info: file size   = 8.63 GiB (5.02 BPW)
ollama            | llama_model_load: error loading model: error loading model architecture: unknown model architecture: 'qwen3'
ollama            | llama_model_load_from_file_impl: failed to load model
ollama            | time=2025-04-29T00:37:26.205Z level=INFO source=sched.go:430 msg="NewLlamaServer failed" model=/root/.ollama/models/blobs/sha256-a8cc1361f3145dc01f6d77c6c82c9116b9ffe3c97b34716fe20418455876c40e error="unable to load model: /root/.ollama/models/blobs/sha256-a8cc1361f3145dc01f6d77c6c82c9116b9ffe3c97b34716fe20418455876c40e"
ollama            | [GIN] 2025/04/29 - 00:37:26 | 500 |   258.25722ms |      172.18.0.4 | POST     "/api/chat"

OS

Linux

GPU

Nvidia

CPU

AMD

Ollama version

0.6.5

Originally created by @cam-narzt on GitHub (Apr 29, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10454 ### What is the issue? when I try to use the qwen3 model in the latest tagged ollama docker image I get the error: `llama_model_load: error loading model: error loading model architecture: unknown model architecture: 'qwen3'`. ### Relevant log output ```shell ollama | time=2025-04-29T00:37:26.067Z level=WARN source=ggml.go:152 msg="key not found" key=qwen3.vision.block_count default=0 ollama | time=2025-04-29T00:37:26.067Z level=WARN source=ggml.go:152 msg="key not found" key=qwen3.vision.block_count default=0 ollama | time=2025-04-29T00:37:26.068Z level=WARN source=ggml.go:152 msg="key not found" key=qwen3.vision.block_count default=0 ollama | time=2025-04-29T00:37:26.068Z level=WARN source=ggml.go:152 msg="key not found" key=qwen3.vision.block_count default=0 ollama | time=2025-04-29T00:37:26.128Z level=INFO source=server.go:105 msg="system memory" total="62.0 GiB" free="55.3 GiB" free_swap="7.9 GiB" ollama | time=2025-04-29T00:37:26.128Z level=WARN source=ggml.go:152 msg="key not found" key=qwen3.vision.block_count default=0 ollama | time=2025-04-29T00:37:26.129Z level=INFO source=server.go:138 msg=offload library=cuda layers.requested=-1 layers.model=41 layers.offload=15 layers.split="" memory.available="[11.5 GiB]" memory.gpu_overhead="0 B" memory.required.full="20.8 GiB" memory.required.partial="11.2 GiB" memory.required.kv="6.2 GiB" memory.required.allocations="[11.2 GiB]" memory.weights.total="8.2 GiB" memory.weights.repeating="7.6 GiB" memory.weights.nonrepeating="608.6 MiB" memory.graph.full="5.2 GiB" memory.graph.partial="5.2 GiB" ollama | llama_model_loader: loaded meta data with 27 key-value pairs and 443 tensors from /root/.ollama/models/blobs/sha256-a8cc1361f3145dc01f6d77c6c82c9116b9ffe3c97b34716fe20418455876c40e (version GGUF V3 (latest)) ollama | llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output. ollama | llama_model_loader: - kv 0: general.architecture str = qwen3 ollama | llama_model_loader: - kv 1: general.type str = model ollama | llama_model_loader: - kv 2: general.name str = Qwen3 14B ollama | llama_model_loader: - kv 3: general.basename str = Qwen3 ollama | llama_model_loader: - kv 4: general.size_label str = 14B ollama | llama_model_loader: - kv 5: qwen3.block_count u32 = 40 ollama | llama_model_loader: - kv 6: qwen3.context_length u32 = 40960 ollama | llama_model_loader: - kv 7: qwen3.embedding_length u32 = 5120 ollama | llama_model_loader: - kv 8: qwen3.feed_forward_length u32 = 17408 ollama | llama_model_loader: - kv 9: qwen3.attention.head_count u32 = 40 ollama | llama_model_loader: - kv 10: qwen3.attention.head_count_kv u32 = 8 ollama | llama_model_loader: - kv 11: qwen3.rope.freq_base f32 = 1000000.000000 ollama | llama_model_loader: - kv 12: qwen3.attention.layer_norm_rms_epsilon f32 = 0.000001 ollama | llama_model_loader: - kv 13: qwen3.attention.key_length u32 = 128 ollama | llama_model_loader: - kv 14: qwen3.attention.value_length u32 = 128 ollama | llama_model_loader: - kv 15: tokenizer.ggml.model str = gpt2 ollama | llama_model_loader: - kv 16: tokenizer.ggml.pre str = qwen2 ollama | llama_model_loader: - kv 17: tokenizer.ggml.tokens arr[str,151936] = ["!", "\"", "#", "$", "%", "&", "'", ... ollama | llama_model_loader: - kv 18: tokenizer.ggml.token_type arr[i32,151936] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... ollama | llama_model_loader: - kv 19: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",... ollama | llama_model_loader: - kv 20: tokenizer.ggml.eos_token_id u32 = 151645 ollama | llama_model_loader: - kv 21: tokenizer.ggml.padding_token_id u32 = 151643 ollama | llama_model_loader: - kv 22: tokenizer.ggml.bos_token_id u32 = 151643 ollama | llama_model_loader: - kv 23: tokenizer.ggml.add_bos_token bool = false ollama | llama_model_loader: - kv 24: tokenizer.chat_template str = {%- if tools %}\n {{- '<|im_start|>... ollama | llama_model_loader: - kv 25: general.quantization_version u32 = 2 ollama | llama_model_loader: - kv 26: general.file_type u32 = 15 ollama | llama_model_loader: - type f32: 161 tensors ollama | llama_model_loader: - type f16: 40 tensors ollama | llama_model_loader: - type q4_K: 221 tensors ollama | llama_model_loader: - type q6_K: 21 tensors ollama | print_info: file format = GGUF V3 (latest) ollama | print_info: file type = Q4_K - Medium ollama | print_info: file size = 8.63 GiB (5.02 BPW) ollama | llama_model_load: error loading model: error loading model architecture: unknown model architecture: 'qwen3' ollama | llama_model_load_from_file_impl: failed to load model ollama | time=2025-04-29T00:37:26.205Z level=INFO source=sched.go:430 msg="NewLlamaServer failed" model=/root/.ollama/models/blobs/sha256-a8cc1361f3145dc01f6d77c6c82c9116b9ffe3c97b34716fe20418455876c40e error="unable to load model: /root/.ollama/models/blobs/sha256-a8cc1361f3145dc01f6d77c6c82c9116b9ffe3c97b34716fe20418455876c40e" ollama | [GIN] 2025/04/29 - 00:37:26 | 500 | 258.25722ms | 172.18.0.4 | POST "/api/chat" ``` ### OS Linux ### GPU Nvidia ### CPU AMD ### Ollama version 0.6.5
GiteaMirror added the bug label 2026-05-04 15:53:12 -05:00
GiteaMirror commented 2026-05-04 15:53:15 -05:00
Author
Owner
Copy Link

@zhou668899 commented on GitHub (Apr 29, 2025):

the same to you

<!-- gh-comment-id:2837161597 --> @zhou668899 commented on GitHub (Apr 29, 2025): the same to you
GiteaMirror commented 2026-05-04 15:53:16 -05:00
Author
Owner
Copy Link

@nandakho commented on GitHub (Apr 29, 2025):

Updating Ollama to 0.6.6 should fix that issue.
As stated in Qwen3 Ollama's model page: This model requires Ollama 0.6.6 or later

<!-- gh-comment-id:2837175276 --> @nandakho commented on GitHub (Apr 29, 2025): Updating Ollama to 0.6.6 should fix that issue. As stated in Qwen3 Ollama's model page: [This model requires Ollama 0.6.6 or later](https://ollama.com/library/qwen3)
GiteaMirror commented 2026-05-04 15:53:17 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Apr 29, 2025):

We're adding a check for this which will tell you to upgrade when you run the model. I'll go ahead and close the issue.

cc @BruceMacD

<!-- gh-comment-id:2839565494 --> @pdevine commented on GitHub (Apr 29, 2025): We're adding a check for this which will tell you to upgrade when you run the model. I'll go ahead and close the issue. cc @BruceMacD
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#68931
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?