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 #4003] Ollama.com - Pull Statistics can be easily fooled #2481

New Issue
Closed
opened 2026-04-12 12:48:42 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-12 12:48:42 -05:00
Owner
Copy Link

Originally created by @electricalgorithm on GitHub (Apr 28, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/4003

Originally assigned to: @bmizerany on GitHub.

What is the issue?

Ollama.com model statistics include a pull count. The pull count statistic provides users with the "popularity" of models, where it seems that it is easy to increase the pull count.

Code Reapply
By using the following Python code snippet, you can increase the amount of pulls a hundred times.

import subprocess
import time
import signal
import os

for i in range(1000):
    process = subprocess.Popen(["ollama", "pull", "USERNAME/MODELNAME"])
    time.sleep(1)
    try:
        os.kill(process.pid, signal.SIGTERM)
    except ProcessLookupError:
        pass

Note that 2-3 minutes is needed to see the changes on the UI.

Suggestion
Pull statistics must be increased after the whole model is pulled into a machine, not after starting a pull request.

I'm not a Go dev, don't think that I can provide a fix in a short time.

OS

macOS

GPU

Apple

CPU

Apple

Ollama version

0.1.32

Originally created by @electricalgorithm on GitHub (Apr 28, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/4003 Originally assigned to: @bmizerany on GitHub. ### What is the issue? Ollama.com model statistics include a pull count. The pull count statistic provides users with the "popularity" of models, where it seems that it is easy to increase the pull count. **Code Reapply** By using the following Python code snippet, you can increase the amount of pulls a hundred times. ```python import subprocess import time import signal import os for i in range(1000): process = subprocess.Popen(["ollama", "pull", "USERNAME/MODELNAME"]) time.sleep(1) try: os.kill(process.pid, signal.SIGTERM) except ProcessLookupError: pass ``` Note that 2-3 minutes is needed to see the changes on the UI. **Suggestion** Pull statistics must be increased after the whole model is pulled into a machine, not after starting a pull request. I'm not a Go dev, don't think that I can provide a fix in a short time. ### OS macOS ### GPU Apple ### CPU Apple ### Ollama version 0.1.32
GiteaMirror added the bug label 2026-04-12 12:48:42 -05:00
GiteaMirror commented 2026-04-12 12:48:43 -05:00
Author
Owner
Copy Link

@electricalgorithm commented on GitHub (Apr 28, 2024):

I've applied the technique to a model of mine. Within the bug, we can achieve more than 1k pulls in under 20 minutes. Here is the example model: https://ollama.com/electricalgorithm/hippomistral. I stopped the script on 1,368 pulls.

resim

You can imagine that no one would find a model published an hour ago & install it quickly, right? 😄

<!-- gh-comment-id:2081488519 --> @electricalgorithm commented on GitHub (Apr 28, 2024): I've applied the technique to a model of mine. Within the bug, we can achieve more than 1k pulls in under 20 minutes. Here is the example model: https://ollama.com/electricalgorithm/hippomistral. I stopped the script on 1,368 pulls. <img width="805" alt="resim" src="https://github.com/ollama/ollama/assets/27111270/dcdbba0c-2a70-4b42-ad90-edade4b08224"> You can imagine that no one would find a model published an hour ago & install it quickly, right? 😄
GiteaMirror commented 2026-04-12 12:48:44 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Apr 29, 2024):

Hey @electricalgorithm , we're actually counting pulls the same way as Docker counts them, i.e. if the manifest gets pulled it counts as a pull. The reason why is because models can share content, and it's not always necessary to re-pull an entire contents of a model.

I think the solution will be to introduce throttling which will prevent you from hitting that API repeatedly. I'm reluctant to stop people from pulling though until it's clear that it's being abused.

<!-- gh-comment-id:2082825311 --> @pdevine commented on GitHub (Apr 29, 2024): Hey @electricalgorithm , we're actually counting pulls the same way as Docker counts them, i.e. if the manifest gets pulled it counts as a pull. The reason why is because models can share content, and it's not always necessary to re-pull an entire contents of a model. I think the solution will be to introduce throttling which will prevent you from hitting that API repeatedly. I'm reluctant to _stop_ people from pulling though until it's clear that it's being abused.
GiteaMirror commented 2026-04-12 12:48:44 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (May 9, 2024):

Hi @electricalgorithm thank you so much for the issue. Will close this for now if that's ok just since there are quite a few issues open and we are hoping to lower that 😊 . That said don't hesitate to create new issues!

<!-- gh-comment-id:2103438329 --> @jmorganca commented on GitHub (May 9, 2024): Hi @electricalgorithm thank you so much for the issue. Will close this for now if that's ok just since there are quite a few issues open and we are hoping to lower that 😊 . That said don't hesitate to create new issues!
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#2481
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?