server: public: move default parameters to its own file

Makefile

@@ -796,7 +796,7 @@ save-load-state: examples/save-load-state/save-load-state.cpp ggml.o llama.o $(C
 	$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
 
-server: examples/server/server.cpp examples/server/utils.hpp examples/server/httplib.h common/json.hpp examples/server/index.html.hpp examples/server/index.js.hpp examples/server/completion.js.hpp examples/server/json-schema-to-grammar.mjs.hpp common/stb_image.h ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
+server: examples/server/server.cpp examples/server/utils.hpp examples/server/httplib.h common/json.hpp examples/server/index.html.hpp examples/server/index.js.hpp examples/server/completion.js.hpp examples/server/params.js.hpp examples/server/json-schema-to-grammar.mjs.hpp common/stb_image.h ggml.o llama.o $(COMMON_DEPS) grammar-parser.o $(OBJS)
 	$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
 	$(CXX) $(CXXFLAGS) $(filter-out %.h %.hpp $<,$^) -Iexamples/server $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) $(LWINSOCK2)
 

build.zig

@@ -141,7 +141,7 @@ pub fn build(b: *std.build.Builder) !void {
         server.linkSystemLibrary("ws2_32");
     }
 
-    const server_assets = [_][]const u8{ "index.html", "index.js", "completion.js", "json-schema-to-grammar.mjs" };
+    const server_assets = [_][]const u8{ "index.html", "index.js", "completion.js", "params.js", "json-schema-to-grammar.mjs" };
     for (server_assets) |asset| {
         const input_path = b.fmt("examples/server/public/{s}", .{asset});
         const output_path = b.fmt("examples/server/{s}.hpp", .{asset});

examples/server/CMakeLists.txt

@@ -11,6 +11,7 @@ set(PUBLIC_ASSETS
     index.html
     index.js
     completion.js
+    params.js
     json-schema-to-grammar.mjs
 )
 foreach(asset ${PUBLIC_ASSETS})

examples/server/README.md

@@ -690,3 +690,12 @@ You can use html formatting if needed.
   </body>
 </html>
 ```
+
+### Update default parameters
+
+Update with desired values in `examples/server/public/params.js` and set
+`--path` flag to the server. For instance,
+
+```bash
+--path /path/to/llama.cpp/examples/server/public/
+```

examples/server/public/index.html

@@ -203,44 +203,13 @@
 
     import { llama } from './completion.js';
     import { SchemaConverter } from './json-schema-to-grammar.mjs';
+
+    // load default session values and parameters
+    import { session, params } from './params.js';
+
     let selected_image = false;
     var slot_id = -1;
 
-    const session = signal({
-      prompt: "This is a conversation between User and Llama, a friendly chatbot. Llama is helpful, kind, honest, good at writing, and never fails to answer any requests immediately and with precision.",
-      template: "{{prompt}}\n\n{{history}}\n{{char}}:",
-      historyTemplate: "{{name}}: {{message}}",
-      transcript: [],
-      type: "chat",  // "chat" | "completion"
-      char: "Llama",
-      user: "User",
-      image_selected: ''
-    })
-
-    const params = signal({
-      n_predict: 400,
-      temperature: 0.7,
-      repeat_last_n: 256, // 0 = disable penalty, -1 = context size
-      repeat_penalty: 1.18, // 1.0 = disabled
-      penalize_nl: false,
-      top_k: 40, // <= 0 to use vocab size
-      top_p: 0.95, // 1.0 = disabled
-      min_p: 0.05, // 0 = disabled
-      tfs_z: 1.0, // 1.0 = disabled
-      typical_p: 1.0, // 1.0 = disabled
-      presence_penalty: 0.0, // 0.0 = disabled
-      frequency_penalty: 0.0, // 0.0 = disabled
-      mirostat: 0, // 0/1/2
-      mirostat_tau: 5, // target entropy
-      mirostat_eta: 0.1, // learning rate
-      grammar: '',
-      n_probs: 0, // no completion_probabilities,
-      min_keep: 0, // min probs from each sampler,
-      image_data: [],
-      cache_prompt: true,
-      api_key: ''
-    })
-
     /* START: Support for storing prompt templates and parameters in browsers LocalStorage */
 
     const local_storage_storageKey = "llamacpp_server_local_storage";

examples/server/public/params.js

@@ -0,0 +1,36 @@
+import { signal } from './index.js';
+
+export const session = signal({
+      prompt: "This is a conversation between User and Llama, a friendly chatbot. Llama is helpful, kind, honest, good at writing, and never fails to answer any requests immediately and with precision.",
+      template: "{{prompt}}\n\n{{history}}\n{{char}}:",
+      historyTemplate: "{{name}}: {{message}}",
+      transcript: [],
+      type: "chat",  // "chat" | "completion"
+      char: "Llama",
+      user: "User",
+      image_selected: ''
+})
+
+export const params = signal({
+      n_predict: 400,
+      temperature: 0.7,
+      repeat_last_n: 256, // 0 = disable penalty, -1 = context size
+      repeat_penalty: 1.18, // 1.0 = disabled
+      penalize_nl: false,
+      top_k: 40, // <= 0 to use vocab size
+      top_p: 0.95, // 1.0 = disabled
+      min_p: 0.05, // 0 = disabled
+      tfs_z: 1.0, // 1.0 = disabled
+      typical_p: 1.0, // 1.0 = disabled
+      presence_penalty: 0.0, // 0.0 = disabled
+      frequency_penalty: 0.0, // 0.0 = disabled
+      mirostat: 0, // 0/1/2
+      mirostat_tau: 5, // target entropy
+      mirostat_eta: 0.1, // learning rate
+      grammar: '',
+      n_probs: 0, // no completion_probabilities,
+      min_keep: 0, // min probs from each sampler,
+      image_data: [],
+      cache_prompt: true,
+      api_key: ''
+})

examples/server/server.cpp

@@ -14,10 +14,11 @@
 #include "httplib.h"
 #include "json.hpp"
 
-// auto generated files (update with ./deps.sh)
+// auto generated files (update during build time (#6661))
 #include "index.html.hpp"
 #include "index.js.hpp"
 #include "completion.js.hpp"
+#include "params.js.hpp"
 #include "json-schema-to-grammar.mjs.hpp"
 
 #include <atomic>