Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

model: add dots3-note (#27060)

Browse Source 
* text: conversion

* init impl

* address review comments

* fix rope

* move to a new llama_kv_cache_dsa_iswa
This commit is contained in:
Xuan-Son Nguyen
2026-08-21 19:52:34 +02:00
committed by GitHub
parent 873e5d8e39
commit 5a32f7b66e
20 changed files with 1412 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-kv-cache.cpp
+2 -1
View File
@@ -323,7 +323,8 @@ llama_kv_cache::llama_kv_cache(
hparams.n_embd_head_k() % 64 == 0;
// always create Hadamard rotation tensors for DeepSeek lightning indexers
if ((model.arch == LLM_ARCH_DEEPSEEK32 || model.arch == LLM_ARCH_DEEPSEEK4 || model.arch == LLM_ARCH_GLM_DSA) &&
if ((model.arch == LLM_ARCH_DEEPSEEK32 || model.arch == LLM_ARCH_DEEPSEEK4 ||
model.arch == LLM_ARCH_GLM_DSA || model.arch == LLM_ARCH_DOTS3NOTE) &&
hparams.n_embd_head_k_full == hparams.indexer_head_size) {
attn_rot_k = true;
}