github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 08:30:05 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[PR #1963] [MERGED] trim chat prompt based on llm context size #21279

New Issue
Closed
opened 2026-04-19 15:33:30 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-19 15:33:30 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/1963
Author: @BruceMacD
Created: 1/12/2024
Status: ✅ Merged
Merged: 1/30/2024
Merged by: @BruceMacD

Base: main ← Head: brucemacd/template-token-smart

📝 Commits (8)

📊 Changes

4 files changed (+440 additions, -57 deletions)

View changed files

📝 server/images.go (+24 -27)
📝 server/images_test.go (+56 -28)
📝 server/routes.go (+105 -2)
📝 server/routes_test.go (+255 -0)

📄 Description

When trimming the input chat prompt we need to make sure we keep the prompt template in the expected format. Without this the prompt will be trimmed without accounting for the model template when the maximum context length is reached, which can result in unexpected behavior from the model.

  • update the ChatPrompt function to return a list of prompt variable, to allow the calling function to append them into the final prompt
  • create the final prompt based on the loaded LLM's context window size, while preserving the prompt template formatting and system message in the first message of the new context window

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/ollama/ollama/pull/1963 **Author:** [@BruceMacD](https://github.com/BruceMacD) **Created:** 1/12/2024 **Status:** ✅ Merged **Merged:** 1/30/2024 **Merged by:** [@BruceMacD](https://github.com/BruceMacD) **Base:** `main` ← **Head:** `brucemacd/template-token-smart` --- ### 📝 Commits (8) - [`04ec4f3`](https://github.com/ollama/ollama/commit/04ec4f340ba52ed2cd34a88beae97758fbff412f) trim chat prompt based on llm context size - [`87eeb4e`](https://github.com/ollama/ollama/commit/87eeb4e885102bbcf0a451817d0d3a1a43b4cd78) Update images_test.go - [`c930fe8`](https://github.com/ollama/ollama/commit/c930fe889fd0cfef52a81a6c68616212ec2a1d6f) refactor contextLimitPrompt - [`94c21b5`](https://github.com/ollama/ollama/commit/94c21b5510aa013b385d750b094b64dd1cdcc26c) maintain system message in chat history - [`28e0293`](https://github.com/ollama/ollama/commit/28e0293ce12c1778a9df2fadc9d87719df529fc1) lint fix - [`171e22b`](https://github.com/ollama/ollama/commit/171e22b0c42697b84a48829c074051cc8590899f) fix lint - [`4f6f68d`](https://github.com/ollama/ollama/commit/4f6f68d475556d2b833009b277415e9295f500ca) formatting - [`b8608ff`](https://github.com/ollama/ollama/commit/b8608ff3b9503a593ed351d567c322ca229846ff) refactor ### 📊 Changes **4 files changed** (+440 additions, -57 deletions) <details> <summary>View changed files</summary> 📝 `server/images.go` (+24 -27) 📝 `server/images_test.go` (+56 -28) 📝 `server/routes.go` (+105 -2) 📝 `server/routes_test.go` (+255 -0) </details> ### 📄 Description When trimming the input chat prompt we need to make sure we keep the prompt template in the expected format. Without this the prompt will be trimmed without accounting for the model template when the maximum context length is reached, which can result in unexpected behavior from the model. - update the `ChatPrompt` function to return a list of prompt variable, to allow the calling function to append them into the final prompt - create the final prompt based on the loaded LLM's context window size, while preserving the prompt template formatting and system message in the first message of the new context window --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-04-19 15:33:30 -05:00
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 pull-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#21279
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?