github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-15 12:48:05 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #8465] Conflicted description of command-r7b #5446

New Issue
Closed
opened 2026-04-12 16:40:45 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-12 16:40:45 -05:00
Owner
Copy Link

Originally created by @vYLQs6 on GitHub (Jan 17, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8465

What is the issue?

https://www.ollama.com/library/command-r7b

Model readme said 128k context length, but Metadata only shows 8k, which is correct?

Context length: Command R7B supports a context length of 128K.

Image

cohere2.context_length 8192

Image

OS

Windows

GPU

Nvidia

CPU

AMD

Ollama version

0.5.5

Originally created by @vYLQs6 on GitHub (Jan 17, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8465 ### What is the issue? ### https://www.ollama.com/library/command-r7b ### Model readme said 128k context length, but Metadata only shows 8k, which is correct? > Context length: Command R7B supports a context length of 128K. ![Image](https://github.com/user-attachments/assets/b4aa49b1-2c21-46aa-b415-3f035f744260) > cohere2.context_length 8192 ![Image](https://github.com/user-attachments/assets/aabb2b49-6a4a-47ed-ba13-46146ecfdb7b) ### OS Windows ### GPU Nvidia ### CPU AMD ### Ollama version 0.5.5
GiteaMirror added the bug label 2026-04-12 16:40:45 -05:00
GiteaMirror commented 2026-04-12 16:40:46 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 17, 2025):

The base model from HF has a context size of 8192, but the model also has 4 sliding windows to improve attention. A simple needle-in-a-haystack test got 100% for an input token length of 40K, so it seems pretty solid. You can increase the context size during completions by setting num_ctx in the API call or by adjusting the Modelfile with PARAMETER num_ctx.

<!-- gh-comment-id:2597452462 --> @rick-github commented on GitHub (Jan 17, 2025): The base model from HF has a [context size](https://huggingface.co/CohereForAI/c4ai-command-r7b-12-2024/blob/6e95067bc61560b05eb014eef8af443034d94ae9/config.json#L18) of 8192, but the model also has 4 sliding windows to improve attention. A simple [needle-in-a-haystack](https://github.com/chigkim/haystack-test) test got 100% for an input token length of 40K, so it seems pretty solid. You can increase the context size during completions by setting `num_ctx` in the API call or by adjusting the Modelfile with `PARAMETER num_ctx`.
GiteaMirror commented 2026-04-12 16:40:46 -05:00
Author
Owner
Copy Link

@mchiang0610 commented on GitHub (Jan 19, 2025):

@vYLQs6 sorry for the confusion. We work with what Cohere has given us.

https://huggingface.co/CohereForAI/c4ai-command-r7b-12-2024/blob/main/config.json#L18

I've fixed the readme for now.

<!-- gh-comment-id:2601041468 --> @mchiang0610 commented on GitHub (Jan 19, 2025): @vYLQs6 sorry for the confusion. We work with what Cohere has given us. https://huggingface.co/CohereForAI/c4ai-command-r7b-12-2024/blob/main/config.json#L18 I've fixed the readme for now.
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
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#5446
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?