Update LLMTests with new APIs

extension/benchmark/apple/Benchmark/Tests/LLaMA/LLaMATests.mm

@@ -87,9 +87,10 @@ @implementation LLaMATests
       for (NSUInteger index = 2; specialTokens.count < 256; ++index) {
         [specialTokens addObject:[NSString stringWithFormat:@"<|reserved_special_token_%zu|>", index]];
       }
-      auto __block runner = [[ExecuTorchLLMTextRunner alloc] initWithModelPath:modelPath
-                                                                 tokenizerPath:tokenizerPath
-                                                                 specialTokens:specialTokens];
+      ExecuTorchLLMTextRunner *__block runner =
+          [[ExecuTorchLLMTextRunner alloc] initWithModelPath:modelPath
+                                               tokenizerPath:tokenizerPath
+                                               specialTokens:specialTokens];
       NSError *error;
       BOOL status = [runner loadWithError:&error];
       if (!status) {
@@ -100,12 +101,14 @@ @implementation LLaMATests
       [testCase measureWithMetrics:@[ tokensPerSecondMetric, [XCTClockMetric new], [XCTMemoryMetric new] ]
                             block:^{
                               tokensPerSecondMetric.tokenCount = 0;
-                              BOOL status = [runner generate:@"Once upon a time"
-                                              sequenceLength:50
-                                           withTokenCallback:^(NSString *token) {
-                                tokensPerSecondMetric.tokenCount++;
+                              BOOL status = [runner generateWithPrompt:@"Once upon a time"
+                                                                config:[[ExecuTorchLLMGenerationConfig alloc] initWithBlock:^(ExecuTorchLLMGenerationConfig *config) {
+                                config.sequenceLength = 50;
+                              }]
+                                                         tokenCallback:^(NSString *token) {
+                                ++tokensPerSecondMetric.tokenCount;
                               }
-                                                       error:NULL];
+                                                                 error:NULL];
                               XCTAssertTrue(status);
                             }];
     },

extension/llm/apple/ExecuTorchLLM/Exported/ExecuTorchLLMMultimodalRunner.h

@@ -193,10 +193,10 @@ __attribute__((objc_subclassing_restricted))
  @param error      On failure, populated with an NSError explaining the issue.
  @return YES if generation completes successfully, NO if an error occurred.
 */
-- (BOOL)generate:(NSArray<ExecuTorchLLMMultimodalInput *> *)inputs
-           config:(ExecuTorchLLMConfig *)config
-withTokenCallback:(nullable void (^)(NSString *))callback
-            error:(NSError **)error
+- (BOOL)generateWithInputs:(NSArray<ExecuTorchLLMMultimodalInput *> *)inputs
+                    config:(ExecuTorchLLMConfig *)config
+             tokenCallback:(nullable void (^)(NSString *))callback
+                     error:(NSError **)error
     NS_SWIFT_NAME(generate(_:_:tokenCallback:));
 
 /**

extension/llm/apple/ExecuTorchLLM/Exported/ExecuTorchLLMMultimodalRunner.mm

@@ -162,10 +162,10 @@ - (BOOL)loadWithError:(NSError**)error {
   return YES;
 }
 
-- (BOOL)generate:(NSArray<ExecuTorchLLMMultimodalInput *> *)inputs
-           config:(ExecuTorchLLMConfig *)config
-withTokenCallback:(nullable void (^)(NSString *))callback
-            error:(NSError **)error {
+- (BOOL)generateWithInputs:(NSArray<ExecuTorchLLMMultimodalInput *> *)inputs
+                    config:(ExecuTorchLLMConfig *)config
+             tokenCallback:(nullable void (^)(NSString *))callback
+                     error:(NSError **)error {
   if (![self loadWithError:error]) {
     return NO;
   }

extension/llm/apple/ExecuTorchLLM/Exported/ExecuTorchLLMTextRunner.h

@@ -60,10 +60,10 @@ __attribute__((deprecated("This API is experimental.")))
  @param error      On failure, populated with an NSError explaining the issue.
  @return YES if generation completes successfully, NO if an error occurred.
 */
-- (BOOL)generate:(NSString *)prompt
-           config:(ExecuTorchLLMConfig *)config
-withTokenCallback:(nullable void (^)(NSString *token))callback
-            error:(NSError **)error
+- (BOOL)generateWithPrompt:(NSString *)prompt
+                    config:(ExecuTorchLLMConfig *)config
+             tokenCallback:(nullable void (^)(NSString *token))callback
+                     error:(NSError **)error
     NS_SWIFT_NAME(generate(_:_:tokenCallback:));
 
 /**

extension/llm/apple/ExecuTorchLLM/Exported/ExecuTorchLLMTextRunner.mm

@@ -74,10 +74,10 @@ - (BOOL)loadWithError:(NSError**)error {
   return YES;
 }
 
-- (BOOL)generate:(NSString*)prompt
-           config:(ExecuTorchLLMConfig *)config
-withTokenCallback:(nullable void (^)(NSString*))callback
-            error:(NSError**)error {
+- (BOOL)generateWithPrompt:(NSString*)prompt
+                    config:(ExecuTorchLLMConfig *)config
+             tokenCallback:(nullable void (^)(NSString*))callback
+                     error:(NSError**)error {
   if (![self loadWithError:error]) {
     return NO;
   }