add basic ltx2.3 support

examples/cli/main.cpp

@@ -19,7 +19,6 @@
 #include "common/media_io.h"
 #include "common/resource_owners.hpp"
 #include "image_metadata.h"
-#include "llm.hpp"
 
 namespace fs = std::filesystem;
 
@@ -501,15 +500,6 @@ int main(int argc, const char* argv[]) {
     SDContextParams ctx_params;
     SDGenerationParams gen_params;
 
-    cli_params.verbose = true;
-    sd_set_log_callback(sd_log_cb, (void*)&cli_params);
-    GemmaTokenizer tokenizer;
-    auto tokens = tokenizer.tokenize("<html> 一只可爱的小猫");
-    for (auto token : tokens) {
-        LOG_INFO("%d", token);
-    }
-    return 0;
-
     parse_args(argc, argv, cli_params, ctx_params, gen_params);
     sd_set_log_callback(sd_log_cb, (void*)&cli_params);
     log_verbose = cli_params.verbose;

examples/common/common.cpp

@@ -340,6 +340,10 @@ ArgOptions SDContextParams::get_options() {
          "--high-noise-diffusion-model",
          "path to the standalone high noise diffusion model",
          &high_noise_diffusion_model_path},
+        {"",
+         "--embeddings-connectors",
+         "path to LTXAV embeddings connectors",
+         &embeddings_connectors_path},
         {"",
          "--vae",
          "path to standalone vae model",
@@ -656,6 +660,7 @@ std::string SDContextParams::to_string() const {
         << "  llm_vision_path: \"" << llm_vision_path << "\",\n"
         << "  diffusion_model_path: \"" << diffusion_model_path << "\",\n"
         << "  high_noise_diffusion_model_path: \"" << high_noise_diffusion_model_path << "\",\n"
+        << "  embeddings_connectors_path: \"" << embeddings_connectors_path << "\",\n"
         << "  vae_path: \"" << vae_path << "\",\n"
         << "  taesd_path: \"" << taesd_path << "\",\n"
         << "  esrgan_path: \"" << esrgan_path << "\",\n"
@@ -712,6 +717,7 @@ sd_ctx_params_t SDContextParams::to_sd_ctx_params_t(bool vae_decode_only, bool f
         llm_vision_path.c_str(),
         diffusion_model_path.c_str(),
         high_noise_diffusion_model_path.c_str(),
+        embeddings_connectors_path.c_str(),
         vae_path.c_str(),
         taesd_path.c_str(),
         control_net_path.c_str(),
@@ -2180,6 +2186,7 @@ sd_vid_gen_params_t SDGenerationParams::to_sd_vid_gen_params_t() {
     params.strength                 = strength;
     params.seed                     = seed;
     params.video_frames             = video_frames;
+    params.fps                      = fps;
     params.vace_strength            = vace_strength;
     params.vae_tiling_params        = vae_tiling_params;
     params.cache                    = cache_params;

examples/common/common.h

@@ -92,6 +92,7 @@ struct SDContextParams {
     std::string llm_vision_path;
     std::string diffusion_model_path;
     std::string high_noise_diffusion_model_path;
+    std::string embeddings_connectors_path;
     std::string vae_path;
     std::string taesd_path;
     std::string esrgan_path;