support change Voices

app.py

@@ -8,8 +8,11 @@ with gr.Blocks(css=customCSS) as demo:
     exceed_flag = gr.State(value=False)
     tmp_string = gr.Textbox(value="", visible=False)
     character_area = gr.HTML(get_character_html("你好呀！我现在支持多语言了呢！"), elem_id="character_area")
+    with gr.Row(elem_id="tools_area"):
+        api_key_input = gr.Textbox(lines=1, max_lines=1, label="API Key:", placeholder="Comming Soon...", elem_classes="wonder-card", elem_id="api_key_input")
+        voice_list = gr.Dropdown(choices=["Chinese", "English", "Japanese"], value="Chinese", label="Choose voice:", elem_id="voice_list", elem_classes="wonder-card")
     with gr.Tab("Speak", elem_id="tab-speak"):
-        speak_input = gr.Textbox(lines=1, label="Talking Flower will say:", placeholder="Support Chinese, English and Japanese...", elem_classes="wonder-card input_text", elem_id="speak_input")
+        speak_input = gr.Textbox(lines=1, max_lines=4, label="Talking Flower will say:", placeholder="Support Chinese, English and Japanese...", elem_classes="wonder-card input_text", elem_id="speak_input")
         speak_button = gr.Button("Speak!", elem_id="speak_button", elem_classes="main-button wonder-card")
         example_category = gr.Examples(["夸夸你 | Praise", "游戏台词 | Scripts", "玩梗 | Meme"], inputs=[tmp_string], elem_id="examples")
     with gr.Tab("Chat", elem_id="tab-chat"):
@@ -27,12 +30,12 @@ with gr.Blocks(css=customCSS) as demo:
     )
     speak_input.submit(submit_lock_fn, show_progress=False).then(
         speak_fn,
-        inputs=[speak_input, exceed_flag],
+        inputs=[speak_input, exceed_flag, voice_list],
         outputs=[audio_output, character_area, exceed_flag, speak_button],
     )
     speak_button.click(submit_lock_fn, show_progress=False).then(
         speak_fn,
-        inputs=[speak_input, exceed_flag],
+        inputs=[speak_input, exceed_flag, voice_list],
         outputs=[audio_output, character_area, exceed_flag, speak_button],
     )
 

css/style.css

@@ -33,6 +33,7 @@
     --base-space: 8px;
 
     --card-width: 95%;
+    --voice-flex: 1;
 }
 
 @media screen and (min-width: 640px) {
@@ -40,6 +41,7 @@
         --gutter-width:24px;
 
         --card-width: 42%;
+        --voice-flex: 0.6;
     }
 }
 
@@ -63,6 +65,7 @@ gradio-app {
     background-position: center center !important;
 }
 
+/* Character Area */
 #character_area {
     width: var(--card-width) !important;
     align-self: center;
@@ -109,6 +112,60 @@ gradio-app {
     height: 10rem;
 }
 
+/* Tools Area */
+#tools_area {
+    width: var(--card-width) !important;
+    align-self: center;
+}
+#tools_area .form {
+    display: flex;
+    flex-direction: row;
+    flex-wrap: wrap;
+    align-items: center;
+    justify-content: center;
+    align-content: center;
+    align-items: center;
+}
+#api_key_input {
+    flex-grow: 1;
+}
+#voice_list {
+    flex-grow: var(--voice-flex);
+}
+#api_key_input input {
+    margin: 0 !important;
+    padding: 6px 12px 6px 12px;
+}
+#api_key_input label span {
+    font-family: var(--font-heading) !important;
+    margin: 0;
+    width: 5rem;
+}
+#api_key_input label,
+#voice_list label {
+    display: flex;
+    align-items: center;
+    flex-wrap: nowrap;
+    align-content: center;
+    justify-content: center;
+    flex-direction: row;
+}
+#voice_list label span {
+    font-family: var(--font-heading) !important;
+    margin: 0;
+    width: 7rem;
+}
+#voice_list input {
+    width: 4rem !important;
+    margin: 0 !important;
+}
+#voice_list .icon-wrap {
+    margin: 0 !important;
+}
+#voice_list ul {
+    width: 5rem !important;
+}
+
 /* Tabs */
 .tabs {
     width: var(--card-width) !important;

presets.py

@@ -43,6 +43,10 @@ def speak_fn(
         interval_between_para=0.2,      # 段间间隔
         interval_between_sent=1,        # 句间间隔
     ):
+    if speaker == "Chinese": speaker = "TalkFlower_CNzh"
+    elif speaker == "English": speaker = "TalkFlower_USen"
+    elif speaker == "Japanese": speaker = "TalkFlower_JPja"
+    else: speaker = "TalkFlower_CNzh"
     audio_list = []
     while text.find("\n\n") != -1:
         text = text.replace("\n\n", "\n")
@@ -86,7 +90,7 @@ def speak_fn(
                         break
                 skip_start = (idx != 0) and skip_start
                 skip_end = (idx != len(sentences_list) - 1) and skip_end
-                print(text_to_generate, lang_to_generate)
+                logging.info(f"{speaker[-4:]}: {text_to_generate}{lang_to_generate}")
                 
                 with torch.no_grad():
                     for i, piece in enumerate(text_to_generate):
@@ -176,9 +180,9 @@ def submit_lock_fn():
 
 
 def init_fn():
-    gr.Info("2023-11-28: 支持多语言啦！闲聊花花现在能说中、英、日语啦！")
+    gr.Info("2023-11-28: 支持多语言（中、英、日）！支持更换音色！")
     # gr.Info("2023-11-24: 优化长句生成效果；增加示例；更新了一些小彩蛋；画了一些大饼）")
-    gr.Info("Support languages: Chinese, English, Japanese. 欢迎在 Community 中提建议~")
+    # gr.Info("Support languages: Chinese, English, Japanese. 欢迎在 Community 中提建议~")
     
     index = random.randint(1,7)
     welcome_text = get_sentence("Welcome", index)