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 #1981] Feature request - support symlink to GGUF in custom model instead of GGUF 1:1 copy #1143

New Issue
Closed
opened 2026-04-12 10:53:41 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-12 10:53:41 -05:00
Owner
Copy Link

Originally created by @mirekjany on GitHub (Jan 13, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/1981

Hello there,

maybe I'm missing something from the documentation. I am working with lots of custom models where the only difference is in System prompts but the custom models are always based on the same GGUF file. So, having ollama to always copy/duplicate it again and again when I create new model is

  1. time-consuming
  2. eats disk-space very quickly

Now, after model is created, I delete the duplicate blob and manually symlink it to the source GGUF, which is a little bit inconvenient.

Would it be possible to implement this? Something like a parameter called SYMLINK (besides FROM).

Thanks!

Originally created by @mirekjany on GitHub (Jan 13, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/1981 Hello there, maybe I'm missing something from the documentation. I am working with lots of custom models where the only difference is in System prompts but the custom models are always based on the same GGUF file. So, having ollama to always copy/duplicate it again and again when I create new model is 1) time-consuming 2) eats disk-space very quickly Now, after model is created, I delete the duplicate blob and manually symlink it to the source GGUF, which is a little bit inconvenient. Would it be possible to implement this? Something like a parameter called SYMLINK (besides FROM). Thanks!
GiteaMirror commented 2026-04-12 10:53:42 -05:00
Author
Owner
Copy Link

@easp commented on GitHub (Jan 15, 2024):

You only need to specify the GGUF once, for the first model you create. Any time you create a new model, reference that model by name, rather than the location of the GGUF.

For example, say you first ollama create my-base .... If you want to create another model based on the same GGUF, use FROM my-base, not the path to the GGUF.

Beyond that, Ollama tracks the uploaded weights using a hash of the actual data. If you upload the same file twice, they'll have the same hash and so only a single copy of the data will be stored.

<!-- gh-comment-id:1891186925 --> @easp commented on GitHub (Jan 15, 2024): You only need to specify the GGUF once, for the first model you create. Any time you create a new model, reference that model by name, rather than the location of the GGUF. For example, say you first `ollama create my-base ...`. If you want to create another model based on the same GGUF, use `FROM my-base`, not the path to the GGUF. Beyond that, Ollama tracks the uploaded weights using a hash of the actual data. If you upload the same file twice, they'll have the same hash and so only a single copy of the data will be stored.
GiteaMirror commented 2026-04-12 10:53:43 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Jan 15, 2024):

@mirekjany as @easp was saying, Ollama does de-duplication automatically for any layer. If a layer is the same between models, only one copy will be saved. It does this using content addressability; the layers are stored by their sha256 hashes in the models/blobs/ directory, and the manifest for the model always references the data by that hash.

<!-- gh-comment-id:1892833471 --> @pdevine commented on GitHub (Jan 15, 2024): @mirekjany as @easp was saying, Ollama does de-duplication automatically for any layer. If a layer is the same between models, only one copy will be saved. It does this using content addressability; the layers are stored by their sha256 hashes in the `models/blobs/` directory, and the manifest for the model always references the data by that hash.
GiteaMirror commented 2026-04-12 10:53:43 -05:00
Author
Owner
Copy Link

@mirekjany commented on GitHub (Feb 15, 2024):

Appreciate the clarification, works as expected, closing the issue.

<!-- gh-comment-id:1945711026 --> @mirekjany commented on GitHub (Feb 15, 2024): Appreciate the clarification, works as expected, closing the issue.
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
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#1143
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?