Merge pull request #156 from zeerd:dev

PiperOrigin-RevId: 627706909

gemma/gemma.cc

@@ -1057,80 +1057,96 @@ void GenerateImpl(GemmaImpl<TConfig>& gemma, size_t max_tokens,
   // In single-turn (non-chat) usage, pos and pos_offset start at 0 and are
   // always equal.
   size_t pos_offset = 0;  // offset relative to pos
-  const double prefill_start = hwy::platform::Now();
 
-  // Prefill stops before prompt_size - 1 since the last prompt token is the
+  auto prefill_phase = [&]() HWY_ATTR {
-  // first input token for generation.
+    bool keep_on = true;
-  while (pos_offset < prompt_size - 1) {
+    const double prefill_start = hwy::platform::Now();
-    const size_t batch_size =
-        std::min(kPrefillBatchSize, prompt_size - 1 - pos_offset);
-    HWY_DASSERT(batch_size <= kPrefillBatchSize);
-    HWY_DASSERT(pos_offset + batch_size <= prompt_size - 1);
-    const int* batch_tokens = prompt.data() + pos_offset;
-    Prefill<kPrefillBatchSize>(batch_tokens, batch_size, pos, weights,
-                               prefill_activations, kv_cache, pool, inner_pool);
-    for (size_t idx = 0; idx < batch_size; ++idx) {
-      stream_token(batch_tokens[idx], 0.0f);
-    }
-    pos += batch_size;
-    pos_offset += batch_size;
-  }
 
-  if (verbosity >= 2) {
+    // Prefill stops before prompt_size - 1 since the last prompt token is the
-    // in the future this output should not occur in GenerateImpl but instead
+    // first input token for generation.
-    // should be available as observable state for frontend code to handle I/O.
+    while (pos_offset < prompt_size - 1 && keep_on) {
-    const double prefill_end = hwy::platform::Now();
+      const size_t batch_size =
-    const double prefill_tok_sec =
+          std::min(kPrefillBatchSize, prompt_size - 1 - pos_offset);
-        static_cast<double>(pos_offset) / (prefill_end - prefill_start);
+      HWY_DASSERT(batch_size <= kPrefillBatchSize);
-    std::cout << "\n[ Prefill tokens / sec = " << prefill_tok_sec << " ]";
+      HWY_DASSERT(pos_offset + batch_size <= prompt_size - 1);
-  }
+      const int* batch_tokens = prompt.data() + pos_offset;
-
+      Prefill<kPrefillBatchSize>(batch_tokens, batch_size, pos, weights,
-  const double gen_start = hwy::platform::Now();
+                                 prefill_activations, kv_cache, pool, inner_pool);
-
+      for (size_t idx = 0; idx < batch_size; ++idx) {
-  HWY_DASSERT(pos_offset == prompt_size - 1);
+        keep_on = stream_token(batch_tokens[idx], 0.0f);
-
+        if(!keep_on) {
-  size_t pos_gen_start = pos_offset;
+          break;
-  int token = prompt.at(pos_offset);
+        }
-  stream_token(token, 0);
-  for (size_t generate_pos = 0;
-       pos < max_tokens && generate_pos < max_generated_tokens;
-       ++pos, ++pos_offset, ++generate_pos) {
-    const bool is_generating_phase = pos_offset >= prompt_size - 1;
-    Transformer(token, pos, weights, activations, kv_cache, pool, inner_pool,
-                layers_output);
-    float* final_activation = activations.x.data();
-    // The condition below is always true if we are doing Prefill above.
-    // We keep it here for clarity so that the code is correct even if Prefill
-    // is disabled.
-    if (is_generating_phase) {
-      PROFILER_ZONE("Gen.Embedding");
-      // Generation phase
-      MatVec<kVocabSize, TConfig::kModelDim>(weights.embedder_input_embedding,
-                                             0, final_activation,
-                                             activations.logits.data(), pool);
-      // Barrier: must have all logits so we can subtract max.
-      Softmax(activations.logits.data(), kVocabSize);
-      token = SampleTopK<TConfig::kTopK>(activations.logits.data(), kVocabSize,
-                                         gen, temperature, accept_token);
-      if (!stream_token(token, activations.logits[token])) {
-        token = EOS_ID;
       }
-    } else {
+      pos += batch_size;
-      // We would take this branch if we were not doing Prefill but would
+      pos_offset += batch_size;
-      // process the tokens of the prompt one at a time.
-      token = prompt.at(pos_offset + 1);
-      stream_token(token, 0);
     }
-    if (token == EOS_ID) {
+
-      if (verbosity >= 2) {
+    if (verbosity >= 2) {
-        const double gen_end = hwy::platform::Now();
+      // in the future this output should not occur in GenerateImpl but instead
-        const double gen_tok_sec =
+      // should be available as observable state for frontend code to handle I/O.
-            static_cast<double>(pos_offset - pos_gen_start) /
+      const double prefill_end = hwy::platform::Now();
-            (gen_end - gen_start);
+      const double prefill_tok_sec =
-        std::cout << "\n[ Generation tokens / sec = " << gen_tok_sec << " ]\n";
+          static_cast<double>(pos_offset) / (prefill_end - prefill_start);
+      std::cout << "\n[ Prefill tokens / sec = " << prefill_tok_sec << " ]";
+    }
+
+    return keep_on;
+  };
+
+  auto transform_phase = [&]() HWY_ATTR {
+
+    const double gen_start = hwy::platform::Now();
+
+    HWY_DASSERT(pos_offset == prompt_size - 1);
+
+    size_t pos_gen_start = pos_offset;
+    int token = prompt.at(pos_offset);
+    stream_token(token, 0);
+    for (size_t generate_pos = 0;
+         pos < max_tokens && generate_pos < max_generated_tokens;
+         ++pos, ++pos_offset, ++generate_pos) {
+      const bool is_generating_phase = pos_offset >= prompt_size - 1;
+      Transformer(token, pos, weights, activations, kv_cache, pool, inner_pool,
+                  layers_output);
+      float* final_activation = activations.x.data();
+      // The condition below is always true if we are doing Prefill above.
+      // We keep it here for clarity so that the code is correct even if Prefill
+      // is disabled.
+      if (is_generating_phase) {
+        PROFILER_ZONE("Gen.Embedding");
+        // Generation phase
+        MatVec<kVocabSize, TConfig::kModelDim>(weights.embedder_input_embedding,
+                                               0, final_activation,
+                                               activations.logits.data(), pool);
+        // Barrier: must have all logits so we can subtract max.
+        Softmax(activations.logits.data(), kVocabSize);
+        token = SampleTopK<TConfig::kTopK>(activations.logits.data(), kVocabSize,
+                                           gen, temperature, accept_token);
+        if (!stream_token(token, activations.logits[token])) {
+          token = EOS_ID;
+        }
+      } else {
+        // We would take this branch if we were not doing Prefill but would
+        // process the tokens of the prompt one at a time.
+        token = prompt.at(pos_offset + 1);
+        stream_token(token, 0);
+      }
+      if (token == EOS_ID) {
+        if (verbosity >= 2) {
+          const double gen_end = hwy::platform::Now();
+          const double gen_tok_sec =
+              static_cast<double>(pos_offset - pos_gen_start) /
+              (gen_end - gen_start);
+          std::cout << "\n[ Generation tokens / sec = " << gen_tok_sec << " ]\n";
+        }
+        break;
       }
-      break;
     }
+  };
+
+  if(prefill_phase()) {
+    transform_phase();
   }
 }