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

[PR #11301] [CLOSED] model: precompute special tokens in NewVocabulary to avoid repeated alloc #75796

New Issue
Closed
opened 2026-05-05 08:13:38 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-05-05 08:13:38 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/11301
Author: @Stogas
Created: 7/4/2025
Status: ❌ Closed

Base: main ← Head: main

📝 Commits (2)

  • a4fd04a model: precompute special tokens in NewVocabulary to avoid repeated allocations
  • 0bb3dcd model: add benchmark for SpecialVocabulary()

📊 Changes

10 files changed (+125 additions, -77 deletions)

View changed files

📝 model/bytepairencoding_test.go (+1 -5)
📝 model/models/llama/model.go (+10 -9)
📝 model/models/llama4/model.go (+16 -13)
📝 model/models/mistral3/model.go (+10 -9)
📝 model/models/qwen2/model.go (+10 -9)
📝 model/models/qwen25vl/model.go (+10 -9)
📝 model/models/qwen3/model.go (+10 -9)
📝 model/vocabulary.go (+22 -10)
➕ model/vocabulary_bench_test.go (+31 -0)
📝 model/vocabulary_test.go (+5 -4)

📄 Description

This moves the special token list creation into the constructor, so that SpecialVocabulary() doesn't need to run sync.Once.Do() repeatedly. Doing this only once per model load avoids excessive memory allocations during chat requests when sync.Once.Do() is called with a Closure / Anonymous function.

Fixes #11299

🔄 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/11301 **Author:** [@Stogas](https://github.com/Stogas) **Created:** 7/4/2025 **Status:** ❌ Closed **Base:** `main` ← **Head:** `main` --- ### 📝 Commits (2) - [`a4fd04a`](https://github.com/ollama/ollama/commit/a4fd04a3b9677bfe1816ef370a24f30438f20ac3) model: precompute special tokens in NewVocabulary to avoid repeated allocations - [`0bb3dcd`](https://github.com/ollama/ollama/commit/0bb3dcd8d0e5ab833f11e1239ecfa43d0b6aa030) model: add benchmark for SpecialVocabulary() ### 📊 Changes **10 files changed** (+125 additions, -77 deletions) <details> <summary>View changed files</summary> 📝 `model/bytepairencoding_test.go` (+1 -5) 📝 `model/models/llama/model.go` (+10 -9) 📝 `model/models/llama4/model.go` (+16 -13) 📝 `model/models/mistral3/model.go` (+10 -9) 📝 `model/models/qwen2/model.go` (+10 -9) 📝 `model/models/qwen25vl/model.go` (+10 -9) 📝 `model/models/qwen3/model.go` (+10 -9) 📝 `model/vocabulary.go` (+22 -10) ➕ `model/vocabulary_bench_test.go` (+31 -0) 📝 `model/vocabulary_test.go` (+5 -4) </details> ### 📄 Description This moves the special token list creation into the constructor, so that `SpecialVocabulary()` doesn't need to run `sync.Once.Do()` repeatedly. Doing this only once per model load avoids excessive memory allocations during chat requests when `sync.Once.Do()` is called with a Closure / Anonymous function. Fixes #11299 --- <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 2026-05-05 08:13:38 -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#75796
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?