ollama - 💡(How to fix) Fix bug: 0.22.0 image generation working from DMG but not Homebrew with mlx response parse and mlx scanner errors [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
ollama/ollama#15882Fetched 2026-04-30 06:18:43
View on GitHub
Comments
0
Participants
1
Timeline
2
Reactions
0
Participants
Timeline (top)
labeled ×1unsubscribed ×1

Error Message

Below log excerpt show errors only seen with Ollama 0.22.0 installed by Homebrew, tested using both models cited in experimental announcement in Jan-2026, x/z-image-turbo and x/flux2-klein. Command execution will accept and process, within 1-2 second reach 100% and stdout will show blank new line without any error message. level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line="" level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line="" level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line="" level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line="" level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line="" level=WARN source=server.go:135 msg=mlx-runner msg="SIGSEGV: segmentation violation" level=WARN source=server.go:135 msg=mlx-runner msg="PC=0x115423130 m=11 sigcode=2 addr=0x0" level=WARN source=server.go:135 msg=mlx-runner msg="signal arrived during cgo execution" level=WARN source=server.go:135 msg=mlx-runner msg="fault 0x0" level=ERROR source=server.go:362 msg="mlx scanner error" error="unexpected EOF"

Code Example

ollama run x/z-image-turbo "a cat"
ollama run x/flux2-klein "a cat"

---

level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line=""
level=INFO source=server.go:128 msg=mlx-runner msg="    step 1: 0.00s (JIT warmup), peak 5.3 GB"
level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line=""
level=INFO source=server.go:128 msg=mlx-runner msg="    step 2: 0.00s, peak 5.3 GB"
level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line=""
level=INFO source=server.go:128 msg=mlx-runner msg="    step 3: 0.00s, peak 5.3 GB"
level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line=""
level=INFO source=server.go:128 msg=mlx-runner msg="    step 4: 0.00s, peak 5.3 GB"
level=INFO source=server.go:128 msg=mlx-runner msg="  Denoised 4 steps in 0.01s (0.00s/step), peak 5.3 GB"
level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line=""
level=WARN source=server.go:135 msg=mlx-runner msg="SIGSEGV: segmentation violation"
level=WARN source=server.go:135 msg=mlx-runner msg="PC=0x115423130 m=11 sigcode=2 addr=0x0"
level=WARN source=server.go:135 msg=mlx-runner msg="signal arrived during cgo execution"
....
....
....
level=WARN source=server.go:135 msg=mlx-runner msg="fault   0x0"
level=INFO source=server.go:128 msg=mlx-runner msg="  Decoding VAE... "
level=ERROR source=server.go:362 msg="mlx scanner error" error="unexpected EOF"
RAW_BUFFERClick to expand / collapse

What is the issue?

Ollama 0.22.0 image generation is working from DMG, but not when installation is performed using Homebrew.

May have some relation to https://github.com/ollama/ollama/issues/15107

Below log excerpt show errors only seen with Ollama 0.22.0 installed by Homebrew, tested using both models cited in experimental announcement in Jan-2026, x/z-image-turbo and x/flux2-klein. Command execution will accept and process, within 1-2 second reach 100% and stdout will show blank new line without any error message.

Test commands:

ollama run x/z-image-turbo "a cat"
ollama run x/flux2-klein "a cat"

Relevant log output

level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line=""
level=INFO source=server.go:128 msg=mlx-runner msg="    step 1: 0.00s (JIT warmup), peak 5.3 GB"
level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line=""
level=INFO source=server.go:128 msg=mlx-runner msg="    step 2: 0.00s, peak 5.3 GB"
level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line=""
level=INFO source=server.go:128 msg=mlx-runner msg="    step 3: 0.00s, peak 5.3 GB"
level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line=""
level=INFO source=server.go:128 msg=mlx-runner msg="    step 4: 0.00s, peak 5.3 GB"
level=INFO source=server.go:128 msg=mlx-runner msg="  Denoised 4 steps in 0.01s (0.00s/step), peak 5.3 GB"
level=DEBUG source=server.go:331 msg="mlx response parse error" error="unexpected end of JSON input" line=""
level=WARN source=server.go:135 msg=mlx-runner msg="SIGSEGV: segmentation violation"
level=WARN source=server.go:135 msg=mlx-runner msg="PC=0x115423130 m=11 sigcode=2 addr=0x0"
level=WARN source=server.go:135 msg=mlx-runner msg="signal arrived during cgo execution"
....
....
....
level=WARN source=server.go:135 msg=mlx-runner msg="fault   0x0"
level=INFO source=server.go:128 msg=mlx-runner msg="  Decoding VAE... "
level=ERROR source=server.go:362 msg="mlx scanner error" error="unexpected EOF"

OS

macOS

GPU

Apple

CPU

Apple

Ollama version

0.22.0

extent analysis

TL;DR

The issue with Ollama 0.22.0 image generation when installed via Homebrew on macOS may be related to a JSON parsing error or a segmentation violation, suggesting a potential compatibility or configuration problem.

Guidance

  • Investigate the mlx response parse error with unexpected end of JSON input to determine if there's an issue with the JSON data being processed or if it's a symptom of a larger problem.
  • Examine the SIGSEGV: segmentation violation error and the subsequent warnings to identify if there's a memory-related issue or a compatibility problem with the Apple GPU or CPU.
  • Compare the installation and execution environments between the DMG installation (which works) and the Homebrew installation to identify any differences that could be causing the issue.
  • Consider checking the issue tracker for any updates or fixes related to the mentioned https://github.com/ollama/ollama/issues/15107, as it may be relevant to the problem at hand.

Notes

The provided log output suggests a complex issue that could be related to either a software bug, a compatibility problem with the Apple hardware, or an issue with the Homebrew installation process. Without more specific details about the environment and the exact commands used for installation, it's challenging to provide a definitive fix.

Recommendation

Apply workaround: Given the complexity and the potential for this being a compatibility or configuration issue, attempting to replicate the working environment of the DMG installation or waiting for an update that addresses the mentioned issue might be the most straightforward path forward.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

ollama - 💡(How to fix) Fix bug: 0.22.0 image generation working from DMG but not Homebrew with mlx response parse and mlx scanner errors [1 participants]