mirror
/
ollama
mirror of https://github.com/ollama/ollama
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Enable Ollama engine by default (#13443) 

This changes the default behavior to use the Ollama engine for supported
models, while retaining the ability to disable the Ollama engine and
fall back to the Llama engine.  Models in the OllamaEngineRequired list
will always run on the Ollama engine.
This commit is contained in:
Daniel Hiltgen 2025-12-12 11:48:43 -08:00 committed by GitHub
parent de9ecfd01c
commit 7730895158
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
envconfig/config.go
View File

@ -199,7 +199,7 @@ var (
// MultiUserCache optimizes prompt caching for multi-user scenarios
MultiUserCache = Bool("OLLAMA_MULTIUSER_CACHE")
// Enable the new Ollama engine
NewEngine = Bool("OLLAMA_NEW_ENGINE")
NewEngine = BoolWithDefault("OLLAMA_NEW_ENGINE")
// ContextLength sets the default context length
ContextLength = Uint("OLLAMA_CONTEXT_LENGTH", 4096)
// Auth enables authentication between the Ollama client and server
@ -291,7 +291,7 @@ func AsMap() map[string]EnvVar {
"OLLAMA_SCHED_SPREAD": {"OLLAMA_SCHED_SPREAD", SchedSpread(), "Always schedule model across all GPUs"},
"OLLAMA_MULTIUSER_CACHE": {"OLLAMA_MULTIUSER_CACHE", MultiUserCache(), "Optimize prompt caching for multi-user scenarios"},
"OLLAMA_CONTEXT_LENGTH": {"OLLAMA_CONTEXT_LENGTH", ContextLength(), "Context length to use unless otherwise specified (default: 4096)"},
"OLLAMA_NEW_ENGINE": {"OLLAMA_NEW_ENGINE", NewEngine(), "Enable the new Ollama engine"},
"OLLAMA_NEW_ENGINE": {"OLLAMA_NEW_ENGINE", NewEngine(true), "Enable the new Ollama engine"},
"OLLAMA_REMOTES": {"OLLAMA_REMOTES", Remotes(), "Allowed hosts for remote models (default \"ollama.com\")"},
// Informational

2
llm/server.go
View File

@ -143,7 +143,7 @@ func NewLlamaServer(systemInfo ml.SystemInfo, gpus []ml.DeviceInfo, modelPath st
var llamaModel *llama.Model
var textProcessor model.TextProcessor
var err error
if envconfig.NewEngine() || f.KV().OllamaEngineRequired() {
if envconfig.NewEngine(true) || f.KV().OllamaEngineRequired() {
if len(projectors) == 0 {
textProcessor, err = model.NewTextProcessor(modelPath)
} else {