[PR #12934] [MERGED] chore: update models to use slice/chunk/chunksections #39876

Closed
opened 2026-04-23 00:53:20 -05:00 by GiteaMirror · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/12934
Author: @mxyng
Created: 11/3/2025
Status: Merged
Merged: 11/13/2025
Merged by: @mxyng

Base: mainHead: mxyng/use-slice


📝 Commits (10+)

📊 Changes

13 files changed (+59 additions, -135 deletions)

View changed files

📝 convert/convert_gptoss.go (+6 -3)
📝 ml/backend.go (+0 -3)
📝 ml/backend/ggml/ggml.go (+0 -28)
📝 ml/nn/pooling/pooling.go (+2 -3)
📝 model/models/bert/embed.go (+1 -1)
📝 model/models/deepseek2/model.go (+14 -26)
📝 model/models/gemma3n/model_text.go (+7 -7)
📝 model/models/gptoss/model.go (+5 -33)
📝 model/models/llama4/model.go (+2 -6)
📝 model/models/llama4/model_vision.go (+9 -13)
📝 model/models/mistral3/model_vision.go (+3 -3)
📝 model/models/qwen25vl/model_vision.go (+3 -3)
📝 model/models/qwen3vl/model_vision.go (+7 -6)

📄 Description

where it's appropriate, use (*Tensor).Slice, (*Tensor).Chunk and (*Tensor).ChunkSections. some places such as summing experts is inappropriate since the view manipulates the shape at the same time


🔄 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/12934 **Author:** [@mxyng](https://github.com/mxyng) **Created:** 11/3/2025 **Status:** ✅ Merged **Merged:** 11/13/2025 **Merged by:** [@mxyng](https://github.com/mxyng) **Base:** `main` ← **Head:** `mxyng/use-slice` --- ### 📝 Commits (10+) - [`5697906`](https://github.com/ollama/ollama/commit/56979066133be0ad0032a8dc795289085ebd696e) use slice/chunks - [`fb820d9`](https://github.com/ollama/ollama/commit/fb820d9865a208330d26c8ecbc080f41c33d926e) bert - [`b9fdc89`](https://github.com/ollama/ollama/commit/b9fdc897800cfadb47fcdc44f6e23f207577e6b9) llama4 - [`5cef355`](https://github.com/ollama/ollama/commit/5cef355a6df61d4c8860723f8717ceee29d3635d) gemma3n - [`956e39f`](https://github.com/ollama/ollama/commit/956e39f3a14857f02719c6aa66e87a2cad1c88dd) gptoss - [`46e6080`](https://github.com/ollama/ollama/commit/46e6080228de832d0c944397368390a79a368658) mistral3 - [`2abe682`](https://github.com/ollama/ollama/commit/2abe68239b34735af2889b3a364bcf50dfb15515) qwen3vl - [`4435f56`](https://github.com/ollama/ollama/commit/4435f56dbb2a0dd314f87dbe4b27fe74adc09370) qwen25vl - [`4bbce47`](https://github.com/ollama/ollama/commit/4bbce473dd0f4d30cfa61909db796fdd3a34eeb9) deepseek2 - [`086cb3a`](https://github.com/ollama/ollama/commit/086cb3ad176af92487a703e20e3242bd4f64f633) remove unused ops ### 📊 Changes **13 files changed** (+59 additions, -135 deletions) <details> <summary>View changed files</summary> 📝 `convert/convert_gptoss.go` (+6 -3) 📝 `ml/backend.go` (+0 -3) 📝 `ml/backend/ggml/ggml.go` (+0 -28) 📝 `ml/nn/pooling/pooling.go` (+2 -3) 📝 `model/models/bert/embed.go` (+1 -1) 📝 `model/models/deepseek2/model.go` (+14 -26) 📝 `model/models/gemma3n/model_text.go` (+7 -7) 📝 `model/models/gptoss/model.go` (+5 -33) 📝 `model/models/llama4/model.go` (+2 -6) 📝 `model/models/llama4/model_vision.go` (+9 -13) 📝 `model/models/mistral3/model_vision.go` (+3 -3) 📝 `model/models/qwen25vl/model_vision.go` (+3 -3) 📝 `model/models/qwen3vl/model_vision.go` (+7 -6) </details> ### 📄 Description where it's appropriate, use `(*Tensor).Slice`, `(*Tensor).Chunk` and `(*Tensor).ChunkSections`. some places such as summing experts is inappropriate since the view manipulates the shape at the same time --- <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 2026-04-23 00:53:20 -05:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#39876