github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #6772] Using the qwen2-7B-q8 model, the GPU can reach 14-15GB when accessed with the Dify API #50779

New Issue
Closed
opened 2026-04-28 17:06:53 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-28 17:06:53 -05:00
Owner
Copy Link

Originally created by @bingbing6 on GitHub (Sep 12, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/6772

What is the issue?

When using the qwen2-7B-q8 model on Dify's API, the GPU is up to 15GB, but directly using Ollama's API is the normal 9GB, and Dify says that they also call Ollama's API normally, without doing any other processing. See my question to them for details:
https://github.com/langgenius/dify/issues/8294

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

ollama version is 0.3.10

Originally created by @bingbing6 on GitHub (Sep 12, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/6772 ### What is the issue? When using the qwen2-7B-q8 model on Dify's API, the GPU is up to 15GB, but directly using Ollama's API is the normal 9GB, and Dify says that they also call Ollama's API normally, without doing any other processing. See my question to them for details: https://github.com/langgenius/dify/issues/8294 ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version ollama version is 0.3.10
GiteaMirror added the question label 2026-04-28 17:06:53 -05:00
GiteaMirror commented 2026-04-28 17:06:55 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Sep 13, 2024):

Dify may be setting a large context window with num_ctx. Server logs from your ollama server may help in debugging.

<!-- gh-comment-id:2347837664 --> @rick-github commented on GitHub (Sep 13, 2024): Dify may be setting a large context window with `num_ctx`. [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) from your ollama server may help in debugging.
GiteaMirror commented 2026-04-28 17:06:55 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Sep 16, 2024):

settings screenshot

From this it looks like you are setting context window (token limit) to 128000. The default value via the API is 2048. This is why the model is taking so much more VRAM.

<!-- gh-comment-id:2351907200 --> @rick-github commented on GitHub (Sep 16, 2024): ![settings screenshot](https://private-user-images.githubusercontent.com/51957370/366754820-d3c19613-f1a2-4399-b532-5957568a8082.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.Sn6nhrun5gxV2-sRgbnRIAmuiT94cYyXWIK-qEmLzv4) From this it looks like you are setting context window (token limit) to 128000. The default value via the API is 2048. This is why the model is taking so much more VRAM.
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 question
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#50779
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?