error ollama run

#1
by boss003i - opened

ollama run hf.co/Ex0bit/Elbaz-GLM-4.6V-Flash-PRISM
Error: 500 Internal Server Error: llama runner process has terminated: exit status

server log
time=2025-12-12T06:14:44.946-05:00 level=INFO source=sched.go:470 msg="Load failed" model=/home/ubuntu/.ollama/models/blobs/sha256-c7427e77baf811bf04d749116b9ffa443ea0f6acfd5d446665ac64f3fe73c9cf error="llama runner process has terminated: exit status 2"
[GIN] 2025/12/12 - 06:14:44 | 500 | 1.424141105s | 127.0.0.1 | POST "/api/generate"

@boss003i - Can I trouble you to confirm the machine/specs and ollama version? (ollama --version)

Ex0bit changed discussion status to closed
ollama version is 0.13.5```

`Device 0: NVIDIA GeForce RTX 4090, compute capability 8.9, VMM: yes, ID: GPU`

--- vision hparams ---
load_hparams: image_size: 336
load_hparams: patch_size: 14
load_hparams: has_llava_proj: 0
load_hparams: minicpmv_version: 0
load_hparams: n_merge: 2
load_hparams: n_wa_pattern: 0
load_hparams: image_min_pixels: 6272
load_hparams: image_max_pixels: 3211264

load_hparams: model size: 1707.61 MiB
load_hparams: metadata size: 0.06 MiB
clip_init: failed to load model '/home/ubuntu/.ollama/models/blobs/sha256-88b81ee0ca10a823d7c7aa363a4badf80b43f3b76e0914959bdf9928afb206dd': operator(): unable to find tensor mm.model.fc.weight

SIGSEGV: segmentation violation
PC=0x7fa7119527d1 m=5 sigcode=1 addr=0x7fb496b8c1b8
signal arrived during cgo execution

fixed with new pull

Original vision mmproj was patch fixed prior to official llama.cpp support. Removing the old blobs and redownloading should do it.

Sign up or log in to comment