replacing tts with piper + updates

Dockerfile

@@ -23,7 +23,7 @@ RUN apt-get update \
 RUN rosdep update && rosdep install --from-paths src --ignore-src -r -y
 RUN pip3 install -r src/llama_ros/requirements.txt --use-deprecated=legacy-resolver
 RUN pip3 install -r src/whisper_ros/requirements.txt --use-deprecated=legacy-resolver
-RUN pip3 install -r src/tts_ros/requirements.txt --use-deprecated=legacy-resolver
+RUN pip3 install -r src/piper_ros/requirements.txt --use-deprecated=legacy-resolver
 
 # colcon the ws
 FROM deps AS builder

README.md

@@ -18,11 +18,11 @@
 
 ## Dependencies
 
-- [yasmin](https://github.com/uleroboticsgroup/yasmin)[[3.0.1](https://github.com/uleroboticsgroup/yasmin/releases/tag/3.0.1)]
-- [audio_common](https://github.com/mgonzs13/audio_common)[[4.0.2](https://github.com/mgonzs13/audio_common/releases/tag/4.0.2)]
-- [llama_ros](https://github.com/mgonzs13/llama_ros)[[4.1.5](https://github.com/mgonzs13/llama_ros/releases/tag/4.1.5)]
-- [whisper_ros](https://github.com/mgonzs13/whisper_ros)[[2.7.1](https://github.com/mgonzs13/whisper_ros/releases/tag/2.7.1)]
-- [tts_ros](https://github.com/mgonzs13/tts_ros)[[2.4.0](https://github.com/mgonzs13/tts_ros/releases/tag/2.4.0)]
+- [yasmin](https://github.com/uleroboticsgroup/yasmin)[[3.0.3](https://github.com/uleroboticsgroup/yasmin/releases/tag/3.0.3)]
+- [audio_common](https://github.com/mgonzs13/audio_common)[[4.0.3](https://github.com/mgonzs13/audio_common/releases/tag/4.0.3)]
+- [llama_ros](https://github.com/mgonzs13/llama_ros)[[4.1.7](https://github.com/mgonzs13/llama_ros/releases/tag/4.1.7)]
+- [whisper_ros](https://github.com/mgonzs13/whisper_ros)[[3.0.2](https://github.com/mgonzs13/whisper_ros/releases/tag/3.0.2)]
+- [piper_ros](https://github.com/mgonzs13/piper_ros)[[1.0.1](https://github.com/mgonzs13/piper_ros/releases/tag/1.0.1)]
 
 ## Installation
 
@@ -36,7 +36,7 @@ $ cd ~/ros2_ws
 $ rosdep install --from-paths src --ignore-src -r -y
 $ pip3 install -r src/llama_ros/requirements.txt
 $ pip3 install -r src/whisper_ros/requirements.txt
-$ pip3 install -r src/tts_ros/requirements.txt
+$ pip3 install -r src/piper_ros/requirements.txt
 $ colcon build --cmake-args -DGGML_CUDA=ON
 ```
 

chatbot_bringup/launch/chatbot.launch.py

@@ -17,12 +17,11 @@
 import os
 from launch_ros.actions import Node
 from launch import LaunchDescription
+from launch.substitutions import LaunchConfiguration
 from launch.actions import SetEnvironmentVariable, IncludeLaunchDescription
 from launch.launch_description_sources import PythonLaunchDescriptionSource
 from ament_index_python.packages import get_package_share_directory
 from llama_bringup.utils import create_llama_launch
-from launch.conditions import IfCondition
-from launch.substitutions import LaunchConfiguration, PythonExpression
 
 
 def generate_launch_description():
@@ -42,39 +41,39 @@ def generate_launch_description():
         launch_arguments={
             "launch_audio_capturer": LaunchConfiguration(
                 "launch_audio_capturer", default=True
-            )
+            ),
+            "model_repo": "ggerganov/whisper.cpp",
+            "model_filename": "ggml-large-v3-turbo-q5_0.bin",
         }.items(),
     )
 
     llama_cmd = create_llama_launch(
         n_ctx=4096,
         n_batch=256,
-        n_gpu_layers=33,
+        n_gpu_layers=29,
         n_threads=-1,
         n_predict=-1,
-        model_repo="QuantFactory/Hermes-2-Theta-Llama-3-8B-GGUF",
-        model_filename="Hermes-2-Theta-Llama-3-8B.Q4_K_M.gguf",
+        model_repo="bartowski/Qwen2.5-1.5B-Instruct-GGUF",
+        model_filename="Qwen2.5-1.5B-Instruct-Q4_K_M.gguf",
         system_prompt_type="ChatML",
     )
 
-    audio_player_cmd = Node(
-        package="audio_common",
-        executable="audio_player_node",
-        name="player_node",
-        namespace="audio",
-        output="both",
-        remappings=[("audio", "out")],
-        condition=IfCondition(
-            PythonExpression([LaunchConfiguration("launch_audio_player", default=True)])
+    piper_node_cmd = IncludeLaunchDescription(
+        PythonLaunchDescriptionSource(
+            os.path.join(
+                get_package_share_directory("piper_bringup"),
+                "launch",
+                "piper.launch.py",
+            )
         ),
-    )
-
-    tts_node_cmd = Node(
-        package="tts_ros",
-        executable="tts_node",
-        output="both",
-        parameters=[{"device": "cuda"}],
-        remappings=[("audio", "/audio/out")],
+        launch_arguments={
+            "launch_audio_player": LaunchConfiguration(
+                "launch_audio_player", default=True
+            ),
+            "model_repo": "rhasspy/piper-voices",
+            "model_filename": "en/en_US/lessac/low/en_US-lessac-low.onnx",
+            "config_filename": "en/en_US/lessac/low/en_US-lessac-low.onnx.json",
+        }.items(),
     )
 
     chatbot_node_cmd = Node(
@@ -95,9 +94,8 @@ def generate_launch_description():
     ld.add_action(stdout_linebuf_envvar)
 
     ld.add_action(whisper_cmd)
+    ld.add_action(piper_node_cmd)
     ld.add_action(llama_cmd)
-    ld.add_action(audio_player_cmd)
-    ld.add_action(tts_node_cmd)
     ld.add_action(chatbot_node_cmd)
     ld.add_action(yasmin_viewer_cmd)
 

chatbot_bringup/package.xml

@@ -12,7 +12,7 @@
   <depend>audio_common</depend>
   <depend>llama_bringup</depend>
   <depend>whisper_bringup</depend>
-  <depend>tts_ros</depend>
+  <depend>piper_ros</depend>
 
   <depend>chatbot_ros</depend>
 

dependencies.repos

@@ -2,20 +2,20 @@ repositories:
   yasmin:
     type: git
     url: https://github.com/uleroboticsgroup/yasmin
-    version: tags/3.0.1
+    version: tags/3.0.3
   audio_common:
     type: git
     url: https://github.com/mgonzs13/audio_common
-    version: tags/4.0.2
+    version: tags/4.0.3
   llama_ros:
     type: git
     url: https://github.com/mgonzs13/llama_ros
-    version: tags/4.1.5
+    version: tags/4.1.7
   whisper_ros:
     type: git
     url: https://github.com/mgonzs13/whisper_ros
-    version: tags/2.7.1
-  tts_ros:
+    version: tags/3.0.2
+  piper_ros:
     type: git
-    url: https://github.com/mgonzs13/tts_ros
-    version: tags/2.4.0
+    url: https://github.com/mgonzs13/piper_ros
+    version: tags/1.0.1