feat: add LCM scheduler (#983)

denoiser.hpp

@@ -253,6 +253,23 @@ struct SGMUniformScheduler : SigmaScheduler {
     }
 };
 
+struct LCMScheduler : SigmaScheduler {
+    std::vector<float> get_sigmas(uint32_t n, float sigma_min, float sigma_max, t_to_sigma_t t_to_sigma) override {
+        std::vector<float> result;
+        result.reserve(n + 1);
+        const int original_steps = 50;
+        const int k              = TIMESTEPS / original_steps;
+        for (int i = 0; i < n; i++) {
+            // the rounding ensures we match the training schedule of the LCM model
+            int index    = (i * original_steps) / n;
+            int timestep = (original_steps - index) * k - 1;
+            result.push_back(t_to_sigma(timestep));
+        }
+        result.push_back(0.0f);
+        return result;
+    }
+};
+
 struct KarrasScheduler : SigmaScheduler {
     std::vector<float> get_sigmas(uint32_t n, float sigma_min, float sigma_max, t_to_sigma_t t_to_sigma) override {
         // These *COULD* be function arguments here,
@@ -375,6 +392,10 @@ struct Denoiser {
                 LOG_INFO("get_sigmas with SmoothStep scheduler");
                 scheduler = std::make_shared<SmoothStepScheduler>();
                 break;
+            case LCM_SCHEDULER:
+                LOG_INFO("get_sigmas with LCM scheduler");
+                scheduler = std::make_shared<LCMScheduler>();
+                break;
             default:
                 LOG_INFO("get_sigmas with discrete scheduler (default)");
                 scheduler = std::make_shared<DiscreteScheduler>();

examples/cli/README.md

@@ -107,8 +107,8 @@ Options:
                                            compatibility issues with quantized parameters, but it usually offers faster inference
                                            speed and, in some cases, lower memory usage. The at_runtime mode, on the other
                                            hand, is exactly the opposite.
-  --scheduler                              denoiser sigma scheduler, one of [discrete, karras, exponential, ays, gits, smoothstep, sgm_uniform, simple], default:
+  --scheduler                              denoiser sigma scheduler, one of [discrete, karras, exponential, ays, gits, smoothstep, sgm_uniform, simple, lcm],
-                                           discrete
+                                           default: discrete
   --skip-layers                            layers to skip for SLG steps (default: [7,8,9])
   --high-noise-sampling-method             (high noise) sampling method, one of [euler, euler_a, heun, dpm2, dpm++2s_a, dpm++2m, dpm++2mv2, ipndm, ipndm_v, lcm,
                                            ddim_trailing, tcd] default: euler for Flux/SD3/Wan, euler_a otherwise

examples/cli/main.cpp

@@ -1197,7 +1197,7 @@ void parse_args(int argc, const char** argv, SDParams& params) {
          on_lora_apply_mode_arg},
         {"",
          "--scheduler",
-         "denoiser sigma scheduler, one of [discrete, karras, exponential, ays, gits, smoothstep, sgm_uniform, simple], default: discrete",
+         "denoiser sigma scheduler, one of [discrete, karras, exponential, ays, gits, smoothstep, sgm_uniform, simple, lcm], default: discrete",
          on_scheduler_arg},
         {"",
          "--skip-layers",

stable-diffusion.cpp

@@ -2268,6 +2268,7 @@ const char* scheduler_to_str[] = {
     "sgm_uniform",
     "simple",
     "smoothstep",
+    "lcm",
 };
 
 const char* sd_scheduler_name(enum scheduler_t scheduler) {

stable-diffusion.h

@@ -61,6 +61,7 @@ enum scheduler_t {
     SGM_UNIFORM_SCHEDULER,
     SIMPLE_SCHEDULER,
     SMOOTHSTEP_SCHEDULER,
+    LCM_SCHEDULER,
     SCHEDULER_COUNT
 };