[PR #6846] [MERGED] add solar pro (preview) #12243

Closed
opened 2026-04-12 23:52:46 -05:00 by GiteaMirror · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/6846
Author: @mxyng
Created: 9/17/2024
Status: Merged
Merged: 9/18/2024
Merged by: @jmorganca

Base: mainHead: mxyng/solar-pro


📝 Commits (2)

  • b8770a6 add solar pro
  • 0531fe7 Rename 1000-solar-pro.patch to 0008-solar-pro.patch

📊 Changes

1 file changed (+402 additions, -0 deletions)

View changed files

llm/patches/0008-solar-pro.patch (+402 -0)

📄 Description

solar-pro introduces block skip connections where blocks are connected to other, non-sequential blocks with a scale multiple

this change adds 4 new keys to store the skip connections and one new tensor to store the scalar. the scalar is implemented as a 1-dimensional tensor with 2 elements derived from the model's bskcn_tv configuration. in general, the values are bskcn_tv, 1 - bskcn_tv

https://huggingface.co/upstage/solar-pro-preview-instruct


🔄 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/6846 **Author:** [@mxyng](https://github.com/mxyng) **Created:** 9/17/2024 **Status:** ✅ Merged **Merged:** 9/18/2024 **Merged by:** [@jmorganca](https://github.com/jmorganca) **Base:** `main` ← **Head:** `mxyng/solar-pro` --- ### 📝 Commits (2) - [`b8770a6`](https://github.com/ollama/ollama/commit/b8770a6997dfdd70b7bd0d3f07e82735b4c9ea77) add solar pro - [`0531fe7`](https://github.com/ollama/ollama/commit/0531fe7dde0af88bf239d1ec454b17f3c614ce49) Rename 1000-solar-pro.patch to 0008-solar-pro.patch ### 📊 Changes **1 file changed** (+402 additions, -0 deletions) <details> <summary>View changed files</summary> ➕ `llm/patches/0008-solar-pro.patch` (+402 -0) </details> ### 📄 Description solar-pro introduces block skip connections where blocks are connected to other, non-sequential blocks with a scale multiple this change adds 4 new keys to store the skip connections and one new tensor to store the scalar. the scalar is implemented as a 1-dimensional tensor with 2 elements derived from the model's `bskcn_tv` configuration. in general, the values are `bskcn_tv, 1 - bskcn_tv` https://huggingface.co/upstage/solar-pro-preview-instruct --- <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 2026-04-12 23:52:46 -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#12243