common : reuse llama_embd_normalize

common/common.cpp

@@ -1852,3 +1852,19 @@ void dump_kv_cache_view_seqs(const llama_kv_cache_view & view, int row_size) {
 
     printf("\n=== Done dumping\n");
 }
+
+void llama_embd_normalize(const float * inp, float * out, int n) {
+    float norm = 0;
+    for (int i = 0; i < n; i++) {
+        norm += inp[i] * inp[i];
+    }
+    norm = sqrt(norm);
+    if (norm == 0) {
+        return;
+    }
+    norm = 1.0 / norm;
+    for (int i = 0; i < n; i++) {
+        out[i] = inp[i] * norm;
+    }
+}
+

common/common.h

@@ -260,3 +260,10 @@ void dump_kv_cache_view(const llama_kv_cache_view & view, int row_size = 80);
 
 // Dump the KV cache view showing individual sequences in each cell (long output).
 void dump_kv_cache_view_seqs(const llama_kv_cache_view & view, int row_size = 40);
+
+//
+// Embedding utils
+//
+
+void llama_embd_normalize(const float * inp, float * out, int n);
+

examples/embedding/embedding.cpp

@@ -23,17 +23,6 @@ static void batch_add_seq(llama_batch & batch, const std::vector<int32_t> & toke
     }
 }
 
-static void normalize(const float * vec, float * out, int n) {
-    float norm = 0;
-    for (int i = 0; i < n; i++) {
-        norm += vec[i] * vec[i];
-    }
-    norm = sqrt(norm);
-    for (int i = 0; i < n; i++) {
-        out[i] = vec[i] / norm;
-    }
-}
-
 static void batch_decode(llama_context * ctx, llama_batch & batch, float * output, int n_seq, int n_embd) {
     // clear previous kv_cache values (irrelevant for embeddings)
     llama_kv_cache_clear(ctx);
@@ -44,7 +33,6 @@ static void batch_decode(llama_context * ctx, llama_batch & batch, float * outpu
         fprintf(stderr, "%s : failed to decode\n", __func__);
     }
 
-    // normalize on copy
     for (int i = 0; i < batch.n_tokens; i++) {
         if (!batch.logits[i]) {
             continue;
@@ -61,7 +49,7 @@ static void batch_decode(llama_context * ctx, llama_batch & batch, float * outpu
         }
 
         float * out = output + batch.seq_id[i][0] * n_embd;
-        normalize(embd, out, n_embd);
+        llama_embd_normalize(embd, out, n_embd);
     }
 }
 

examples/server/server.cpp

@@ -1318,6 +1318,8 @@ struct server_context {
 
         const int n_embd = llama_n_embd(model);
 
+        std::vector<float> embd_res(n_embd, 0.0f);
+
         for (int i = 0; i < batch.n_tokens; ++i) {
             if (!batch.logits[i] || batch.seq_id[i][0] != slot.id + 1) {
                 continue;
@@ -1341,8 +1343,10 @@ struct server_context {
                 continue;
             }
 
+            llama_embd_normalize(embd, embd_res.data(), n_embd);
+
             res.data = json {
-                {"embedding", std::vector<float>(embd, embd + n_embd)},
+                {"embedding", embd_res},
             };
         }
 
@@ -2651,17 +2655,6 @@ inline void signal_handler(int signal) {
     shutdown_handler(signal);
 }
 
-static void normalize(std::vector<float>& vec) {
-    float norm = 0;
-    for (float val : vec) {
-        norm += val * val;
-    }
-    norm = sqrt(norm);
-    for (float& val : vec) {
-        val /= norm;
-    }
-}
-
 int main(int argc, char ** argv) {
 #if SERVER_VERBOSE != 1
     log_disable();
@@ -3357,11 +3350,6 @@ int main(int argc, char ** argv) {
             server_task_result result = ctx_server.queue_results.recv(id_task);
             ctx_server.queue_results.remove_waiting_task_id(id_task);
 
-            // normalize the embedding
-            std::vector<float> embedding = json_value(result.data, "embedding", json::array());
-            normalize(embedding);
-            result.data["embedding"] = embedding;
-
             // append to the responses
             responses.push_back(result.data);
         }