llama: remove diffusion functions

Author: am17an

include/llama.h

@@ -203,15 +203,6 @@ extern "C" {
 
     typedef bool (*llama_progress_callback)(float progress, void * user_data);
 
-    // Diffusion step callback - called after each diffusion step
-    // step: current step number (0-based)
-    // total_steps: total number of diffusion steps
-    // tokens: current token array state
-    // n_tokens: number of tokens in the array
-    // user_data: user-provided context data
-    // Returns: true to continue, false to cancel
-    typedef bool (*llama_diffusion_step_callback)(int32_t step, int32_t total_steps, const llama_token * tokens, int32_t n_tokens, void * user_data);
-
     // Input data for llama_encode/llama_decode
     // A llama_batch object can contain input about one or many sequences
     // The provided arrays (i.e. token, embd, pos, etc.) must have size of n_tokens
@@ -925,40 +916,6 @@ extern "C" {
             struct llama_context * ctx,
               struct llama_batch   batch);
 
-    // Diffusion generation parameters for Dream7B models
-    typedef struct llama_diffusion_params {
-        int32_t steps;                  // number of diffusion steps (default: 512)
-        float eps;                      // epsilon for timesteps (default: 1e-3f)
-        float temperature;              // sampling temperature (default: 0.0f)
-        float top_p;                    // top-p sampling (default: 1.0f)
-        int32_t top_k;                  // top-k sampling (default: 0 = disabled)
-        llama_token mask_token_id;      // mask token id
-        enum {
-            LLAMA_DIFFUSION_ALG_ORIGIN = 0,
-            LLAMA_DIFFUSION_ALG_MASKGIT_PLUS = 1,
-            LLAMA_DIFFUSION_ALG_TOPK_MARGIN = 2,
-            LLAMA_DIFFUSION_ALG_ENTROPY = 3,
-        } algorithm;                    // diffusion algorithm (default: ORIGIN)
-        float alg_temp;                 // algorithm temperature (default: 0.0f)
-        int32_t seed;                   // seed for random number generator (default: 0)
-        llama_diffusion_step_callback step_callback;  // called after each diffusion step
-        void * step_callback_user_data;               // user data for step callback
-    } llama_diffusion_params;
-
-    // Get default diffusion parameters
-    LLAMA_API struct llama_diffusion_params llama_diffusion_default_params(void);
-
-    // Diffusion generation for Dream7B models
-    // Returns the generated tokens, or NULL on error
-    // The caller is responsible for freeing the returned array
-    LLAMA_API llama_token * llama_diffusion_generate(
-            struct llama_context * ctx,
-            const llama_token * input_tokens,
-            int32_t n_input,
-            int32_t max_length,
-            struct llama_diffusion_params params,
-            int32_t * n_generated);
-
     // Set the number of threads used for decoding
     // n_threads is the number of threads used for generation (single token)
     // n_threads_batch is the number of threads used for prompt and batch processing (multiple tokens)