github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-16 13:19:40 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #9412] The phi4-mini model cannot run properly #52647

New Issue
Closed
opened 2026-04-28 23:56:10 -05:00 by GiteaMirror · 15 comments
GiteaMirror commented 2026-04-28 23:56:10 -05:00
Owner
Copy Link

Originally created by @51762344 on GitHub (Feb 28, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/9412

What is the issue?

I have pulled Ollama 0.5.13, but the model cannot respond normally.

Image

Relevant log output

OS

No response

GPU

No response

CPU

No response

Ollama version

No response

Originally created by @51762344 on GitHub (Feb 28, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/9412 ### What is the issue? I have pulled Ollama 0.5.13, but the model cannot respond normally. ![Image](https://github.com/user-attachments/assets/2fd4dacd-cde2-4d63-9801-31e46c5aa07a) ### Relevant log output ```shell ``` ### OS _No response_ ### GPU _No response_ ### CPU _No response_ ### Ollama version _No response_
GiteaMirror added the bug label 2026-04-28 23:56:10 -05:00
GiteaMirror commented 2026-04-28 23:56:12 -05:00
Author
Owner
Copy Link

@ynot01 commented on GitHub (Feb 28, 2025):

Experiencing similar issues

Running latest (572.60) NVIDIA drivers on RTX 4070 (12 GB) on Windows 11

ollama version is 0.5.13-rc1

Image

<!-- gh-comment-id:2690154683 --> @ynot01 commented on GitHub (Feb 28, 2025): Experiencing similar issues Running latest (572.60) NVIDIA drivers on RTX 4070 (12 GB) on Windows 11 ollama version is 0.5.13-rc1 ![Image](https://github.com/user-attachments/assets/28f9b338-7831-4a83-bfcf-09ef7dc13b36)
GiteaMirror commented 2026-04-28 23:56:12 -05:00
Author
Owner
Copy Link

@MMaturax commented on GitHub (Feb 28, 2025):

Image

Image

Additionally, there is a significant slowdown in inference speed. While I was getting 148 tokens per second for Gemma 2 9B Q4 with an RTX 5090, now the evaluation rate is only 35 tokens per second.

<!-- gh-comment-id:2690195502 --> @MMaturax commented on GitHub (Feb 28, 2025): ![Image](https://github.com/user-attachments/assets/a8f3c9b9-f474-4661-8820-4e6da6389515) ![Image](https://github.com/user-attachments/assets/db73d3a4-759d-47c9-99f6-d2f03e744ae2) Additionally, there is a significant slowdown in inference speed. While I was getting 148 tokens per second for Gemma 2 9B Q4 with an RTX 5090, now the evaluation rate is only 35 tokens per second.
GiteaMirror commented 2026-04-28 23:56:13 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 28, 2025):

It looks like support for phi4 was merged after 0.5.13-rc1 was built, so the next release might work better.

<!-- gh-comment-id:2690218847 --> @rick-github commented on GitHub (Feb 28, 2025): It looks like support for [phi4](https://github.com/ollama/ollama/pull/9403) was merged after 0.5.13-rc1 was built, so the next release might work better.
GiteaMirror commented 2026-04-28 23:56:13 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 28, 2025):

Nope, built from source (with phi4 support merged) and the model does not work.

<!-- gh-comment-id:2690351254 --> @rick-github commented on GitHub (Feb 28, 2025): Nope, built from source (with phi4 support merged) and the model does not work.
GiteaMirror commented 2026-04-28 23:56:13 -05:00
Author
Owner
Copy Link

@nh-99 commented on GitHub (Feb 28, 2025):

I've tried using phi4-mini:3.8b on both an M2 MacBook Air and my Windows PC with an RTX 5080. I was attempting to do tool calling with LangChain, and the queries to Ollama were taking upwards of 10 mins and I had to kill them. The call I was making should have only had ~1500 tokens.

EDIT: Just wanted to note that I am using the Ollama pre-release version too, 0.5.13

<!-- gh-comment-id:2690640670 --> @nh-99 commented on GitHub (Feb 28, 2025): I've tried using `phi4-mini:3.8b` on both an M2 MacBook Air and my Windows PC with an RTX 5080. I was attempting to do tool calling with LangChain, and the queries to Ollama were taking upwards of 10 mins and I had to kill them. The call I was making should have only had ~1500 tokens. EDIT: Just wanted to note that I am using the Ollama pre-release version too, `0.5.13`
GiteaMirror commented 2026-04-28 23:56:14 -05:00
Author
Owner
Copy Link

@Jellepepe commented on GitHub (Feb 28, 2025):

Model on ollama or ollama implementation seems broken, slow performance and mostly gibberish, nowhere near what I've seen demoed. Tested on M4 pro and RTX 3080 (windows)

<!-- gh-comment-id:2690674059 --> @Jellepepe commented on GitHub (Feb 28, 2025): Model on ollama or ollama implementation seems broken, slow performance and mostly gibberish, nowhere near what I've seen demoed. Tested on M4 pro and RTX 3080 (windows)
GiteaMirror commented 2026-04-28 23:56:15 -05:00
Author
Owner
Copy Link

@shaktisd commented on GitHub (Feb 28, 2025):

It is writing gibberish . I am on ollama version is 0.5.13-rc1

Image

<!-- gh-comment-id:2690983642 --> @shaktisd commented on GitHub (Feb 28, 2025): It is writing gibberish . I am on ollama version is 0.5.13-rc1 ![Image](https://github.com/user-attachments/assets/c30ce564-d0e2-408a-9a6b-95bc52aecac0)
GiteaMirror commented 2026-04-28 23:56:16 -05:00
Author
Owner
Copy Link

@AntonIXO commented on GitHub (Feb 28, 2025):

The same on CPU on Archlinux, built from the latest source.

<!-- gh-comment-id:2690992516 --> @AntonIXO commented on GitHub (Feb 28, 2025): The same on CPU on Archlinux, built from the latest source.
GiteaMirror commented 2026-04-28 23:56:19 -05:00
Author
Owner
Copy Link

@mswcap commented on GitHub (Feb 28, 2025):

Same here. Running the latest RC version of Ollama but here also Phi4-mini goes bezerk on me. It's funny to see, but not what I expected.

Image

<!-- gh-comment-id:2691276425 --> @mswcap commented on GitHub (Feb 28, 2025): Same here. Running the latest RC version of Ollama but here also Phi4-mini goes bezerk on me. It's funny to see, but not what I expected. ![Image](https://github.com/user-attachments/assets/58062706-ff0c-4772-a59e-0a01bfc4c64b)
GiteaMirror commented 2026-04-28 23:56:20 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Feb 28, 2025):

Hi there, sorry about this. Re-downloading the model ollama pull phi4-mini will fix the issue (let me know if it doesn't).

<!-- gh-comment-id:2691445122 --> @jmorganca commented on GitHub (Feb 28, 2025): Hi there, sorry about this. Re-downloading the model `ollama pull phi4-mini` will fix the issue (let me know if it doesn't).
GiteaMirror commented 2026-04-28 23:56:20 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Feb 28, 2025):

@rick-github thanks for helping converge issues to this one

<!-- gh-comment-id:2691445432 --> @jmorganca commented on GitHub (Feb 28, 2025): @rick-github thanks for helping converge issues to this one
GiteaMirror commented 2026-04-28 23:56:20 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 28, 2025):

Can confirm the model works now.

<!-- gh-comment-id:2691473066 --> @rick-github commented on GitHub (Feb 28, 2025): Can confirm the model works now.
GiteaMirror commented 2026-04-28 23:56:20 -05:00
Author
Owner
Copy Link

@laurentpayot commented on GitHub (Feb 28, 2025):

Working now on my Ubuntu 24.10 machine. Thanks @jmorganca! 👍

<!-- gh-comment-id:2691486882 --> @laurentpayot commented on GitHub (Feb 28, 2025): Working now on my Ubuntu 24.10 machine. Thanks @jmorganca! 👍
GiteaMirror commented 2026-04-28 23:56:20 -05:00
Author
Owner
Copy Link

@shaktisd commented on GitHub (Mar 1, 2025):

@jmorganca - Thanks, It is working now

Image

<!-- gh-comment-id:2691806397 --> @shaktisd commented on GitHub (Mar 1, 2025): @jmorganca - Thanks, It is working now ![Image](https://github.com/user-attachments/assets/a55e608d-2312-4bd2-b392-8467bdf3f224)
GiteaMirror commented 2026-04-28 23:56:21 -05:00
Author
Owner
Copy Link

@mswcap commented on GitHub (Mar 1, 2025):

Thank you @jmorganca, it works now.

<!-- gh-comment-id:2692037492 --> @mswcap commented on GitHub (Mar 1, 2025): Thank you @jmorganca, it works now.
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#52647
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?