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 #11659] Allow users to set context length like we used to instead of using terrible GUI presets every single launch #69766

New Issue
Closed
opened 2026-05-04 19:07:53 -05:00 by GiteaMirror · 5 comments
GiteaMirror commented 2026-05-04 19:07:53 -05:00
Owner
Copy Link

Originally created by @FatheredPuma81 on GitHub (Aug 5, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11659

OLLAMA_CONTEXT_LENGTH no longer functions after 0.9.6. The GUI presets are not a good alternative when you need something between 32k and 64k context length. It's also way more inconvenient than double clicking a script that sets it to exactly what you want.
If this was intentional my feature request is to bring OLLAMA_CONTEXT_LENGTH variable back and make it override whatever the GUI setting is.

Originally created by @FatheredPuma81 on GitHub (Aug 5, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11659 OLLAMA_CONTEXT_LENGTH no longer functions after 0.9.6. The GUI presets are not a good alternative when you need something between 32k and 64k context length. It's also way more inconvenient than double clicking a script that sets it to exactly what you want. If this was intentional my feature request is to bring OLLAMA_CONTEXT_LENGTH variable back and make it override whatever the GUI setting is.
GiteaMirror added the bugfeature request labels 2026-05-04 19:07:53 -05:00
GiteaMirror commented 2026-05-04 19:07:58 -05:00
Author
Owner
Copy Link

@FatheredPuma81 commented on GitHub (Aug 5, 2025):

Well in the meantime I guess I'll be upgrading to version 0.9.6 I guess.

<!-- gh-comment-id:3153115794 --> @FatheredPuma81 commented on GitHub (Aug 5, 2025): Well in the meantime I guess I'll be upgrading to version 0.9.6 I guess.
GiteaMirror commented 2026-05-04 19:07:58 -05:00
Author
Owner
Copy Link

@FatheredPuma81 commented on GitHub (Aug 5, 2025):

@jmorganca If it's a bug then I plead the "The AI told me it was intentionally depreciated and I decided to update halfway through what I was doing and spent and hour trying to figure out wth was going on"
I fixed the OP.

<!-- gh-comment-id:3154159632 --> @FatheredPuma81 commented on GitHub (Aug 5, 2025): @jmorganca If it's a bug then I plead the "The AI told me it was intentionally depreciated and I decided to update halfway through what I was doing and spent and hour trying to figure out wth was going on" I fixed the OP.
GiteaMirror commented 2026-05-04 19:07:59 -05:00
Author
Owner
Copy Link

@av commented on GitHub (Aug 5, 2025):

Also stopped working for me in Docker, no app in sight

<!-- gh-comment-id:3156496876 --> @av commented on GitHub (Aug 5, 2025): Also stopped working for me in Docker, no app in sight
GiteaMirror commented 2026-05-04 19:07:59 -05:00
Author
Owner
Copy Link

@blackfeather9 commented on GitHub (Aug 11, 2025):

I recently started using the OLLAMA_CONTEXT_LENGTH variable and it wasn't working at all (running ollama with docker compose). I set it to 128k using the following format in ollama.env:

OLLAMA_CONTEXT_LENGTH value="128000 ollama serve"

Perhaps I was reading this documentation wrong https://github.com/ollama/ollama/blob/main/docs/faq.md#how-can-i-specify-the-context-window-size

When I ran the container, it gave me the following error:

time=2025-08-11T11:29:42.338Z level=WARN source=config.go:210 msg="invalid environment variable, using default" key=OLLAMA_CONTEXT_LENGTH value="128000 ollama serve" default=4096

The solution this morning was to change my env variable to simply OLLAMA_CONTEXT_LENGTH value=128000 and then run docker compose down and docker compose up -d --force-recreate which loaded the new values. Loading a model into memory then shows it following whatever context I defined, i.e. llama_context: n_ctx = 128000

Using the ollama image from 8/7 which I believe is running 0.11.4. Maybe this will help you.

<!-- gh-comment-id:3174529073 --> @blackfeather9 commented on GitHub (Aug 11, 2025): I recently started using the `OLLAMA_CONTEXT_LENGTH` variable and it wasn't working at all (running ollama with docker compose). I set it to 128k using the following format in `ollama.env`: ``` OLLAMA_CONTEXT_LENGTH value="128000 ollama serve" ``` Perhaps I was reading this documentation wrong https://github.com/ollama/ollama/blob/main/docs/faq.md#how-can-i-specify-the-context-window-size When I ran the container, it gave me the following error: ``` time=2025-08-11T11:29:42.338Z level=WARN source=config.go:210 msg="invalid environment variable, using default" key=OLLAMA_CONTEXT_LENGTH value="128000 ollama serve" default=4096 ``` The solution this morning was to change my env variable to simply `OLLAMA_CONTEXT_LENGTH value=128000` and then run `docker compose down` and `docker compose up -d --force-recreate` which loaded the new values. Loading a model into memory then shows it following whatever context I defined, i.e. `llama_context: n_ctx = 128000` Using the ollama image from 8/7 which I believe is running ` 0.11.4`. Maybe this will help you.
GiteaMirror commented 2026-05-04 19:08:00 -05:00
Author
Owner
Copy Link

@fictiontoreality commented on GitHub (Jan 17, 2026):

OLLAMA_CONTEXT_LENGTH works fine for me on ollama 0.14.2 running in a Docker container.

$ docker compose exec ollama ollama ps
NAME           ID              SIZE     PROCESSOR    CONTEXT    UNTIL              
gpt-oss:20b    17052f91a42e    17 GB    100% GPU     128000     2 minutes from now

My compose.yaml

services:
  # Easy setup of local LLM.
  ollama:
    # Requires Nvidia Container Toolkit for Nvidia GPU support.
    # Also supports AMD GPUs via a different approach.
    # For details: https://hub.docker.com/r/ollama/ollama
    image: ollama/ollama:0.14.2
    container_name: ollama
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: 1
              capabilities: [gpu]
    volumes:
      - /home/user/data/ollama:/root/.ollama
    environment:
      OLLAMA_CONTEXT_LENGTH: 128000
    ports:
      - 11434:11434
    # privileged: true
    restart: unless-stopped
<!-- gh-comment-id:3764202008 --> @fictiontoreality commented on GitHub (Jan 17, 2026): `OLLAMA_CONTEXT_LENGTH` works fine for me on ollama 0.14.2 running in a Docker container. ``` $ docker compose exec ollama ollama ps NAME ID SIZE PROCESSOR CONTEXT UNTIL gpt-oss:20b 17052f91a42e 17 GB 100% GPU 128000 2 minutes from now ``` My compose.yaml ``` services: # Easy setup of local LLM. ollama: # Requires Nvidia Container Toolkit for Nvidia GPU support. # Also supports AMD GPUs via a different approach. # For details: https://hub.docker.com/r/ollama/ollama image: ollama/ollama:0.14.2 container_name: ollama deploy: resources: reservations: devices: - driver: nvidia count: 1 capabilities: [gpu] volumes: - /home/user/data/ollama:/root/.ollama environment: OLLAMA_CONTEXT_LENGTH: 128000 ports: - 11434:11434 # privileged: true restart: unless-stopped ```
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 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#69766
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?