github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-03-09 07:16:38 -05:00
Code Issues Packages Projects Releases Wiki Activity

model: support for qwen3.5 architecture (#14378)

Browse Source
This commit is contained in:
Jeffrey Morgan
2026-02-24 20:08:05 -08:00
committed by GitHub
parent 9d902d63ce
commit da70c3222e
31 changed files with 1902 additions and 1628 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
fs/ggml/ggml.go
View File

@@ -290,6 +290,7 @@ func (kv KV) OllamaEngineRequired() bool {
"olmo3",
"qwen25vl",
"qwen3", "qwen3moe",
"qwen35", "qwen35moe",
"qwen3next",
"qwen3vl", "qwen3vlmoe",
"glm4moelite",
@@ -868,7 +869,12 @@ func (f GGML) SupportsFlashAttention() bool {
return false
}
if arch := f.KV().Architecture(); slices.Contains([]string{"gemma2"}, arch) {
arch := f.KV().Architecture()
if slices.Contains([]string{"qwen35", "qwen35moe", "qwen3next"}, arch) {
return true
}
if slices.Contains([]string{"gemma2"}, arch) {
return false
}
@@ -892,6 +898,7 @@ func (f GGML) FlashAttention() bool {
"nemotron_h", "nemotron_h_moe",
"olmo3",
"qwen3", "qwen3moe",
"qwen35", "qwen35moe",
"qwen3next",
"qwen3vl", "qwen3vlmoe",
}, f.KV().String("general.architecture"))