Update index.html

index.html

@@ -183,11 +183,18 @@
           const [processor, model] = await SmolVLM.getInstance(modelId, dtypeSettings, device);
           const text = processor.apply_chat_template(messages, { add_generation_prompt: true });
           const inputs = await processor(text, [image], { do_image_splitting: doImageSplitting });
-          
-          const start = performance.now();
+                    
+          let numTokens = 0;
+          let startTime;
+          let tps = 0;
+          const token_callback_function = () => {
+            startTime = startTime || performance.now();
+            tps = (numTokens++ / (performance.now() - startTime)) * 1000;
+          };
           const streamer = new TextStreamer(processor.tokenizer, {
             skip_prompt: true,
             skip_special_tokens: true,
+            token_callback_function,
           });
           await model.generate({
             ...inputs,
@@ -195,10 +202,8 @@
             min_new_tokens: maxTokens,
             streamer,
           });
-          const end = performance.now();
-          const elapsed = end - start;
-          
-          const tps = maxTokens / (elapsed / 1000);  // accurate TPS
+          const elapsed = performance.now() - start;
+
 
           totalTime += elapsed;
           totalTps += tps;