Revert "[multimodal] Allow generate and prefill to take move sematics" (#15000)

Reverts #14643

Author: larryliu0820

extension/llm/runner/multimodal_runner.cpp

@@ -62,11 +62,6 @@ Error MultimodalRunner::load() {
     ET_LOG(Info, format, __VA_ARGS__);     \
   }
 
-Error MultimodalRunner::prefill(std::vector<MultimodalInput>&& inputs) {
-  // Forward to the const reference version
-  return prefill(inputs);
-}
-
 Error MultimodalRunner::prefill(const std::vector<MultimodalInput>& inputs) {
   if (!is_loaded()) {
     ET_CHECK_OK_OR_RETURN_ERROR(load());
@@ -77,16 +72,6 @@ Error MultimodalRunner::prefill(const std::vector<MultimodalInput>& inputs) {
   return Error::Ok;
 }
 
-Error MultimodalRunner::generate(
-    std::vector<MultimodalInput>&& inputs,
-    const GenerationConfig& config,
-    std::function<void(const std::string&)> token_callback,
-    std::function<void(const Stats&)> stats_callback) {
-  // Forward to the const reference version
-  return generate(
-      inputs, config, std::move(token_callback), std::move(stats_callback));
-}
-
 Error MultimodalRunner::generate(
     const std::vector<MultimodalInput>& inputs,
     const GenerationConfig& config,

extension/llm/runner/multimodal_runner.h

@@ -119,21 +119,6 @@ class ET_EXPERIMENTAL MultimodalRunner {
       std::function<void(const std::string&)> token_callback = {},
       std::function<void(const Stats&)> stats_callback = {});
 
-  /**
-   * Generate tokens from multimodal inputs with move semantics.
-   * This overload allows efficient transfer of temporary vectors.
-   * @param inputs A vector of MultimodalInput objects (moved).
-   * @param config Generation configuration parameters.
-   * @param token_callback Callback function called for each generated token.
-   * @param stats_callback Callback function for generation statistics.
-   * @return The error code. KV cache position is tracked internally in pos_.
-   */
-  virtual ::executorch::runtime::Error generate(
-      std::vector<MultimodalInput>&& inputs,
-      const GenerationConfig& config,
-      std::function<void(const std::string&)> token_callback = {},
-      std::function<void(const Stats&)> stats_callback = {});
-
   /**
    * Prefill multimodal inputs, for example to reload chat history.
    * @param inputs A vector of MultimodalInput objects containing images and
@@ -143,15 +128,6 @@ class ET_EXPERIMENTAL MultimodalRunner {
   virtual ::executorch::runtime::Error prefill(
       const std::vector<MultimodalInput>& inputs);
 
-  /**
-   * Prefill multimodal inputs with move semantics.
-   * This overload allows efficient transfer of temporary vectors.
-   * @param inputs A vector of MultimodalInput objects (moved).
-   * @return The error code. KV cache position is tracked internally in pos_.
-   */
-  virtual ::executorch::runtime::Error prefill(
-      std::vector<MultimodalInput>&& inputs);
-
   inline void stop() {
     text_token_generator_->stop();
   }