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 #5329] clip models fail to load with unicode characters in OLLAMA_MODELS path on windows #49846

New Issue
Closed
opened 2026-04-28 13:10:36 -05:00 by GiteaMirror · 9 comments
GiteaMirror commented 2026-04-28 13:10:36 -05:00
Owner
Copy Link

Originally created by @Derix76 on GitHub (Jun 27, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/5329

Originally assigned to: @dhiltgen on GitHub.

What is the issue?

I tried to start llava:1.6 (or any similar llava based modell) an the llama server terminated.
llama3 modell or different non llava models work just fine.

GPU is: NVIDIA GeForce RTX 4060" total="8.0 GiB" available="6.9 GiB"
CPU is: AMD Ryzen 7 4700G with Radeon GPU (ignored by Ollama) ("unsupported Radeon iGPU detected skipping" id=0 name="AMD Radeon(TM) Graphics")
OS in Win 11
Geforce Drivers: 555.99 Studio

Log extract:
time=2024-06-27T16:53:15.204+02:00 level=INFO source=memory.go:309 msg="offload to cuda" layers.requested=-1 layers.model=33 layers.offload=33 layers.split="" memory.available="[7.2 GiB]" memory.required.full="5.3 GiB" memory.required.partial="5.3 GiB" memory.required.kv="256.0 MiB" memory.required.allocations="[5.3 GiB]" memory.weights.total="3.9 GiB" memory.weights.repeating="3.8 GiB" memory.weights.nonrepeating="102.6 MiB" memory.graph.full="164.0 MiB" memory.graph.partial="181.0 MiB"
time=2024-06-27T16:53:15.205+02:00 level=WARN source=server.go:241 msg="multimodal models don't support parallel requests yet"
time=2024-06-27T16:53:15.208+02:00 level=INFO source=server.go:368 msg="starting llama server" cmd="C:\Users\Stefan Hüttner\AppData\Local\Programs\Ollama\ollama_runners\cuda_v11.3\ollama_llama_server.exe --model C:\Users\Stefan Hüttner\.ollama\models\blobs\sha256-170370233dd5c5415250a2ecd5c71586352850729062ccef1496385647293868 --ctx-size 2048 --batch-size 512 --embedding --log-disable --n-gpu-layers 33 --mmproj C:\Users\Stefan Hüttner\.ollama\models\blobs\sha256-72d6f08a42f656d36b356dbe0920675899a99ce21192fd66266fb7d82ed07539 --no-mmap --parallel 1 --port 51274"
time=2024-06-27T16:53:15.742+02:00 level=INFO source=sched.go:382 msg="loaded runners" count=1
time=2024-06-27T16:53:15.742+02:00 level=INFO source=server.go:556 msg="waiting for llama runner to start responding"
time=2024-06-27T16:53:15.743+02:00 level=INFO source=server.go:594 msg="waiting for server to become available" status="llm server error"
INFO [wmain] build info | build=3171 commit="7c26775a" tid="17384" timestamp=1719499996
INFO [wmain] system info | n_threads=8 n_threads_batch=-1 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | " tid="17384" timestamp=1719499996 total_threads=16
INFO [wmain] HTTP server listening | hostname="127.0.0.1" n_threads_http="15" port="51274" tid="17384" timestamp=1719499996
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: CUDA_USE_TENSOR_CORES: yes
ggml_cuda_init: found 1 CUDA devices:
Device 0: NVIDIA GeForce RTX 4060, compute capability 8.9, VMM: yes
ERROR [load_model] unable to load clip model | model="C:\Users\Stefan Hüttner\.ollama\models\blobs\sha256-72d6f08a42f656d36b356dbe0920675899a99ce21192fd66266fb7d82ed07539" tid="17384" timestamp=1719499996
time=2024-06-27T16:53:16.512+02:00 level=ERROR source=sched.go:388 msg="error loading llama server" error="llama runner process has terminated: exit status 0xc0000409 "

OS

Windows

GPU

Nvidia

CPU

AMD

Ollama version

0.1.47

Originally created by @Derix76 on GitHub (Jun 27, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/5329 Originally assigned to: @dhiltgen on GitHub. ### What is the issue? I tried to start llava:1.6 (or any similar llava based modell) an the llama server terminated. llama3 modell or different non llava models work just fine. GPU is: NVIDIA GeForce RTX 4060" total="8.0 GiB" available="6.9 GiB" CPU is: AMD Ryzen 7 4700G with Radeon GPU (ignored by Ollama) ("unsupported Radeon iGPU detected skipping" id=0 name="AMD Radeon(TM) Graphics") OS in Win 11 Geforce Drivers: 555.99 Studio Log extract: time=2024-06-27T16:53:15.204+02:00 level=INFO source=memory.go:309 msg="offload to cuda" layers.requested=-1 layers.model=33 layers.offload=33 layers.split="" memory.available="[7.2 GiB]" memory.required.full="5.3 GiB" memory.required.partial="5.3 GiB" memory.required.kv="256.0 MiB" memory.required.allocations="[5.3 GiB]" memory.weights.total="3.9 GiB" memory.weights.repeating="3.8 GiB" memory.weights.nonrepeating="102.6 MiB" memory.graph.full="164.0 MiB" memory.graph.partial="181.0 MiB" time=2024-06-27T16:53:15.205+02:00 level=WARN source=server.go:241 msg="multimodal models don't support parallel requests yet" time=2024-06-27T16:53:15.208+02:00 level=INFO source=server.go:368 msg="starting llama server" cmd="C:\\Users\\Stefan Hüttner\\AppData\\Local\\Programs\\Ollama\\ollama_runners\\cuda_v11.3\\ollama_llama_server.exe --model C:\\Users\\Stefan Hüttner\\.ollama\\models\\blobs\\sha256-170370233dd5c5415250a2ecd5c71586352850729062ccef1496385647293868 --ctx-size 2048 --batch-size 512 --embedding --log-disable --n-gpu-layers 33 --mmproj C:\\Users\\Stefan Hüttner\\.ollama\\models\\blobs\\sha256-72d6f08a42f656d36b356dbe0920675899a99ce21192fd66266fb7d82ed07539 --no-mmap --parallel 1 --port 51274" time=2024-06-27T16:53:15.742+02:00 level=INFO source=sched.go:382 msg="loaded runners" count=1 time=2024-06-27T16:53:15.742+02:00 level=INFO source=server.go:556 msg="waiting for llama runner to start responding" time=2024-06-27T16:53:15.743+02:00 level=INFO source=server.go:594 msg="waiting for server to become available" status="llm server error" INFO [wmain] build info | build=3171 commit="7c26775a" tid="17384" timestamp=1719499996 INFO [wmain] system info | n_threads=8 n_threads_batch=-1 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | " tid="17384" timestamp=1719499996 total_threads=16 INFO [wmain] HTTP server listening | hostname="127.0.0.1" n_threads_http="15" port="51274" tid="17384" timestamp=1719499996 ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no ggml_cuda_init: CUDA_USE_TENSOR_CORES: yes ggml_cuda_init: found 1 CUDA devices: Device 0: NVIDIA GeForce RTX 4060, compute capability 8.9, VMM: yes ERROR [load_model] unable to load clip model | model="C:\\Users\\Stefan Hüttner\\.ollama\\models\\blobs\\sha256-72d6f08a42f656d36b356dbe0920675899a99ce21192fd66266fb7d82ed07539" tid="17384" timestamp=1719499996 time=2024-06-27T16:53:16.512+02:00 level=ERROR source=sched.go:388 msg="error loading llama server" error="llama runner process has terminated: exit status 0xc0000409 " ### OS Windows ### GPU Nvidia ### CPU AMD ### Ollama version 0.1.47
GiteaMirror added the bugwindows labels 2026-04-28 13:10:37 -05:00
GiteaMirror commented 2026-04-28 13:10:42 -05:00
Author
Owner
Copy Link

@jsoncode commented on GitHub (Jun 28, 2024):

I also encountered the same problem
微信截图_20240628211157

<!-- gh-comment-id:2196876992 --> @jsoncode commented on GitHub (Jun 28, 2024): I also encountered the same problem ![微信截图_20240628211157](https://github.com/ollama/ollama/assets/13176273/0270077e-32d0-4ef2-ae3c-abc684ee18e8)
GiteaMirror commented 2026-04-28 13:10:43 -05:00
Author
Owner
Copy Link

@jsoncode commented on GitHub (Jun 28, 2024):

I updated ollama and it worked.

<!-- gh-comment-id:2196897658 --> @jsoncode commented on GitHub (Jun 28, 2024): I updated ollama and it worked.
GiteaMirror commented 2026-04-28 13:10:45 -05:00
Author
Owner
Copy Link

@Derix76 commented on GitHub (Jun 30, 2024):

Thanks, but error still persists after update.

<!-- gh-comment-id:2198502904 --> @Derix76 commented on GitHub (Jun 30, 2024): Thanks, but error still persists after update.
GiteaMirror commented 2026-04-28 13:10:46 -05:00
Author
Owner
Copy Link

@gaduffl commented on GitHub (Jun 30, 2024):

The error occured after updating to the latest version. I get the issue now with llama3 which was working fine before.

<!-- gh-comment-id:2198662513 --> @gaduffl commented on GitHub (Jun 30, 2024): The error occured after updating to the latest version. I get the issue now with llama3 which was working fine before.
GiteaMirror commented 2026-04-28 13:10:51 -05:00
Author
Owner
Copy Link

@Marten-Ka commented on GitHub (Jul 2, 2024):

Same here, opened another issue https://github.com/ollama/ollama/issues/5431

<!-- gh-comment-id:2202941630 --> @Marten-Ka commented on GitHub (Jul 2, 2024): Same here, opened another issue [https://github.com/ollama/ollama/issues/5431](url)
GiteaMirror commented 2026-04-28 13:10:53 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jul 3, 2024):

It looks like there's a bug in the clip model loading code in C++ that doesn't handle unicode characters properly.

Until we can get this fixed, a workaround is create a model directory on your C: with only ASCII characters and set OLLAMA_MODELS to that path for the server.

https://github.com/ollama/ollama/blob/main/docs/faq.md#how-do-i-configure-ollama-server

<!-- gh-comment-id:2204771134 --> @dhiltgen commented on GitHub (Jul 3, 2024): It looks like there's a bug in the clip model loading code in C++ that doesn't handle unicode characters properly. Until we can get this fixed, a workaround is create a model directory on your `C:` with only ASCII characters and set `OLLAMA_MODELS` to that path for the server. https://github.com/ollama/ollama/blob/main/docs/faq.md#how-do-i-configure-ollama-server
GiteaMirror commented 2026-04-28 13:10:54 -05:00
Author
Owner
Copy Link

@Derix76 commented on GitHub (Jul 3, 2024):

Dear @dhiltgen - thanks a lot. This works like a charm. Unexpected that unicode is not handled, but workaround is fine. Many thanks for our great work!

<!-- gh-comment-id:2205416761 --> @Derix76 commented on GitHub (Jul 3, 2024): Dear @dhiltgen - thanks a lot. This works like a charm. Unexpected that unicode is not handled, but workaround is fine. Many thanks for our great work!
GiteaMirror commented 2026-04-28 13:10:56 -05:00
Author
Owner
Copy Link

@Derix76 commented on GitHub (Jul 3, 2024):

Item closed, workaround is doing fine

<!-- gh-comment-id:2205418203 --> @Derix76 commented on GitHub (Jul 3, 2024): Item closed, workaround is doing fine
GiteaMirror commented 2026-04-28 13:10:57 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jul 3, 2024):

Thanks for submitting the issue. This definitely should work, so we'll get this fixed properly.

<!-- gh-comment-id:2206598731 --> @dhiltgen commented on GitHub (Jul 3, 2024): Thanks for submitting the issue. This definitely should work, so we'll get this fixed properly.
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 windows
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#49846
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?