chore(defaults): enlarge defaults, drop gpu layers which is infered (#5308)

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

core/backend/options.go

@@ -99,7 +99,7 @@ func grpcModelOpts(c config.BackendConfig) *pb.ModelOptions {
 		mmap = *c.MMap
 	}
 
-	ctxSize := 1024
+	ctxSize := 4096
 	if c.ContextSize != nil {
 		ctxSize = *c.ContextSize
 	}

core/config/backend_config.go

@@ -304,9 +304,6 @@ func (cfg *BackendConfig) SetDefaults(opts ...ConfigLoaderOption) {
 	defaultTFZ := 1.0
 	defaultZero := 0
 
-	// Try to offload all GPU layers (if GPU is found)
-	defaultHigh := 99999999
-
 	trueV := true
 	falseV := false
 
@@ -366,9 +363,6 @@ func (cfg *BackendConfig) SetDefaults(opts ...ConfigLoaderOption) {
 	if cfg.MirostatTAU == nil {
 		cfg.MirostatTAU = &defaultMirostatTAU
 	}
-	if cfg.NGPULayers == nil {
-		cfg.NGPULayers = &defaultHigh
-	}
 
 	if cfg.LowVRAM == nil {
 		cfg.LowVRAM = &falseV