github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-22 21:51:37 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #14483] Proposal for Text-Only LLM Parameter Loading in Qwen3.5 #87083

New Issue
Open
opened 2026-05-10 04:53:21 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-05-10 04:53:21 -05:00
Owner
Copy Link

Originally created by @lordofkey on GitHub (Feb 27, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/14483

A text-only LLM parameter loading option should be added to the Qwen3.5 model. For pure text-based LLM tasks, this would significantly reduce memory consumption. For example, when loading a 4-bit quantized 27B model, a 4090 with 24GB VRAM cannot accommodate all parameters, resulting in degraded inference speed. By loading only the text inference portion of the parameters, the entire model could fit within GPU memory, substantially improving inference performance.

Originally created by @lordofkey on GitHub (Feb 27, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/14483 A text-only LLM parameter loading option should be added to the Qwen3.5 model. For pure text-based LLM tasks, this would significantly reduce memory consumption. For example, when loading a 4-bit quantized 27B model, a 4090 with 24GB VRAM cannot accommodate all parameters, resulting in degraded inference speed. By loading only the text inference portion of the parameters, the entire model could fit within GPU memory, substantially improving inference performance.
GiteaMirror added the feature request label 2026-05-10 04:53:21 -05:00
GiteaMirror commented 2026-05-10 04:54:05 -05:00
Author
Owner
Copy Link

@mdemirci10 commented on GitHub (Mar 23, 2026):

Qwen3.5:9b model sometimes tries to generate image. For example, examine the following output:

**Here are the monthly financial performance (amount) data for 2026:

[Image blocked: Monthly Financial Performance in 2026]

This graph shows the total financial performance (amount) information for each month in 2026.**

I want to force to return text-only result, because if it tries to generate and display image, image-blocked comment displayed. Does ollama support image result, or not?

<!-- gh-comment-id:4108872095 --> @mdemirci10 commented on GitHub (Mar 23, 2026): Qwen3.5:9b model sometimes tries to generate image. For example, examine the following output: **Here are the monthly financial performance (amount) data for 2026: [Image blocked: Monthly Financial Performance in 2026] This graph shows the total financial performance (amount) information for each month in 2026.** I want to force to return text-only result, because if it tries to generate and display image, image-blocked comment displayed. Does ollama support image result, or not?
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 feature request
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#87083
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?