Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.
rfc: dynamic environ lookup
Michael Yang35b89b2eaba4
diff --git a/app/lifecycle/logging.go b/app/lifecycle/logging.goindex a8f1f7cdf..3672aad59 100644--- a/app/lifecycle/logging.go+++ b/app/lifecycle/logging.go@@ -14,7 +14,7 @@ import (func InitLogging() {level := slog.LevelInfo- if envconfig.Debug {+ if envconfig.Debug() {level = slog.LevelDebug}diff --git a/envconfig/config.go b/envconfig/config.goindex 0abc69686..426507be0 100644--- a/envconfig/config.go+++ b/envconfig/config.go@@ -26,11 +26,24 @@ func (o OllamaHost) String() string {var ErrInvalidHostPort = errors.New("invalid port specified in OLLAMA_HOST")+// Debug returns true if the OLLAMA_DEBUG environment variable is set to a truthy value.+func Debug() bool {+ if s := clean("OLLAMA_DEBUG"); s != "" {+ b, err := strconv.ParseBool(s)+ if err != nil {+ // non-empty value is truthy+ return true+ }++ return b+ }++ return false+}+var (// Set via OLLAMA_ORIGINS in the environmentAllowOrigins []string- // Set via OLLAMA_DEBUG in the environment- Debug bool// Experimental flash attentionFlashAttention bool// Set via OLLAMA_HOST in the environment@@ -80,7 +93,7 @@ type EnvVar struct {func AsMap() map[string]EnvVar {ret := map[string]EnvVar{- "OLLAMA_DEBUG": {"OLLAMA_DEBUG", Debug, "Show additional debug information (e.g. OLLAMA_DEBUG=1)"},+ "OLLAMA_DEBUG": {"OLLAMA_DEBUG", Debug(), "Show additional debug information (e.g. OLLAMA_DEBUG=1)"},"OLLAMA_FLASH_ATTENTION": {"OLLAMA_FLASH_ATTENTION", FlashAttention, "Enabled flash attention"},"OLLAMA_HOST": {"OLLAMA_HOST", Host, "IP Address for the ollama server (default 127.0.0.1:11434)"},"OLLAMA_KEEP_ALIVE": {"OLLAMA_KEEP_ALIVE", KeepAlive, "The duration that models stay loaded in memory (default \"5m\")"},@@ -137,15 +150,6 @@ func init() {}func LoadConfig() {- if debug := clean("OLLAMA_DEBUG"); debug != "" {- d, err := strconv.ParseBool(debug)- if err == nil {- Debug = d- } else {- Debug = true- }- }-if fa := clean("OLLAMA_FLASH_ATTENTION"); fa != "" {d, err := strconv.ParseBool(fa)if err == nil {diff --git a/envconfig/config_test.go b/envconfig/config_test.goindex a5d73fd7c..f083bb03c 100644--- a/envconfig/config_test.go+++ b/envconfig/config_test.go@@ -12,16 +12,15 @@ import ()func TestConfig(t *testing.T) {- Debug = false // Reset whatever was loaded in init()t.Setenv("OLLAMA_DEBUG", "")- LoadConfig()- require.False(t, Debug)+ require.False(t, Debug())+t.Setenv("OLLAMA_DEBUG", "false")- LoadConfig()- require.False(t, Debug)+ require.False(t, Debug())+t.Setenv("OLLAMA_DEBUG", "1")- LoadConfig()- require.True(t, Debug)+ require.True(t, Debug())+t.Setenv("OLLAMA_FLASH_ATTENTION", "1")LoadConfig()require.True(t, FlashAttention)diff --git a/gpu/gpu.go b/gpu/gpu.goindex 6e25cb46d..1815668fd 100644--- a/gpu/gpu.go+++ b/gpu/gpu.go@@ -611,7 +611,7 @@ func LoadOneapiMgmt(oneapiLibPaths []string) (int, *C.oneapi_handle_t, string) {}func getVerboseState() C.uint16_t {- if envconfig.Debug {+ if envconfig.Debug() {return C.uint16_t(1)}return C.uint16_t(0)diff --git a/llm/memory_test.go b/llm/memory_test.goindex f972f9275..06ae74387 100644--- a/llm/memory_test.go+++ b/llm/memory_test.go@@ -8,14 +8,14 @@ import ("testing""github.com/ollama/ollama/api"- "github.com/ollama/ollama/envconfig""github.com/ollama/ollama/gpu""github.com/stretchr/testify/assert""github.com/stretchr/testify/require")func TestEstimateGPULayers(t *testing.T) {- envconfig.Debug = true+ t.Setenv("OLLAMA_DEBUG", "1")+modelName := "dummy"f, err := os.CreateTemp(t.TempDir(), modelName)require.NoError(t, err)diff --git a/llm/server.go b/llm/server.goindex 08463ef01..eb9666506 100644--- a/llm/server.go+++ b/llm/server.go@@ -195,7 +195,7 @@ func NewLlamaServer(gpus gpu.GpuInfoList, model string, ggml *GGML, adapters, prparams = append(params, "--n-gpu-layers", fmt.Sprintf("%d", opts.NumGPU))}- if envconfig.Debug {+ if envconfig.Debug() {params = append(params, "--verbose")}@@ -381,7 +381,7 @@ func NewLlamaServer(gpus gpu.GpuInfoList, model string, ggml *GGML, adapters, pr}slog.Info("starting llama server", "cmd", s.cmd.String())- if envconfig.Debug {+ if envconfig.Debug() {filteredEnv := []string{}for _, ev := range s.cmd.Env {if strings.HasPrefix(ev, "CUDA_") ||diff --git a/server/routes.go b/server/routes.goindex 0d7ca003c..c049421b7 100644--- a/server/routes.go+++ b/server/routes.go@@ -1093,7 +1093,7 @@ func (s *Server) GenerateRoutes() http.Handler {func Serve(ln net.Listener) error {level := slog.LevelInfo- if envconfig.Debug {+ if envconfig.Debug() {level = slog.LevelDebug}