mirror of
https://github.com/ParisNeo/lollms-webui.git
synced 2025-04-16 23:18:53 +00:00
tts_to_file now tts_to_audio
This commit is contained in:
Saifeddine ALOUI
parent
e80c380e9a
commit
72dc903842
@ -1,5 +1,5 @@
|
||||
# =================== Lord Of Large Language Multimodal Systems Configuration file ===========================
|
||||
version: 88
|
||||
version: 90
|
||||
binding_name: null
|
||||
model_name: null
|
||||
model_variant: null
|
||||
@ -86,8 +86,8 @@ xtts_base_url: http://localhost:8020
|
||||
xtts_use_deepspeed: false
|
||||
xtts_use_streaming_mode: true
|
||||
auto_read: false
|
||||
current_voice: null
|
||||
current_language: en
|
||||
xtts_current_voice: null
|
||||
xtts_current_language: en
|
||||
|
||||
# Image generation service
|
||||
enable_sd_service: false
|
||||
@ -145,11 +145,9 @@ audio_silenceTimer: 5000
|
||||
# Data vectorization
|
||||
activate_skills_lib: false # Activate vectorizing previous conversations
|
||||
skills_lib_database_name: "default" # Default skills database
|
||||
summerize_discussion: false # activate discussion summary (better but adds computation time)
|
||||
|
||||
max_summary_size: 512 # in tokens
|
||||
data_vectorization_visualize_on_vectorization: false
|
||||
use_files: true # Activate using files
|
||||
data_vectorization_activate: true # To activate/deactivate data vectorization
|
||||
data_vectorization_method: "tfidf_vectorizer" #"model_embedding" or "tfidf_vectorizer"
|
||||
data_visualization_method: "PCA" #"PCA" or "TSNE"
|
||||
|
||||
@ -1 +1 @@
|
||||
Subproject commit ec5428c68feaef544b582f862055d6204eecb7db
|
||||
Subproject commit 32c7f3ef9f330029e1958fbb964bb233a6d29e9d
|
||||
@ -444,6 +444,7 @@ class LOLLMSWebUI(LOLLMSElfServer):
|
||||
use_deep_speed=self.config.xtts_use_deepspeed,
|
||||
use_streaming_mode=self.config.xtts_use_streaming_mode
|
||||
)
|
||||
|
||||
except:
|
||||
self.warning(f"Personality {personality.name} request using custom voice but couldn't load XTTS")
|
||||
except Exception as ex:
|
||||
@ -1066,12 +1067,12 @@ class LOLLMSWebUI(LOLLMSElfServer):
|
||||
self.nb_received_tokens = 0
|
||||
self.start_time = datetime.now()
|
||||
if self.model is not None:
|
||||
if self.model.binding_type==BindingType.TEXT_IMAGE and len(self.personality.image_files)>0:
|
||||
if self.model.binding_type==BindingType.TEXT_IMAGE and len(client.discussion.image_files)>0:
|
||||
ASCIIColors.info(f"warmup for generating up to {n_predict} tokens")
|
||||
if self.config["override_personality_model_parameters"]:
|
||||
output = self.model.generate_with_images(
|
||||
prompt,
|
||||
self.personality.image_files,
|
||||
client.discussion.image_files,
|
||||
callback=callback,
|
||||
n_predict=n_predict,
|
||||
temperature=self.config['temperature'],
|
||||
@ -1097,7 +1098,7 @@ class LOLLMSWebUI(LOLLMSElfServer):
|
||||
])
|
||||
output = self.model.generate_with_images(
|
||||
prompt,
|
||||
self.personality.image_files,
|
||||
client.discussion.image_files,
|
||||
callback=callback,
|
||||
n_predict=min(n_predict,self.personality.model_n_predicts),
|
||||
temperature=self.personality.model_temperature,
|
||||
@ -1109,7 +1110,7 @@ class LOLLMSWebUI(LOLLMSElfServer):
|
||||
n_threads=self.config['n_threads']
|
||||
)
|
||||
try:
|
||||
post_processed_output = process_ai_output(output, self.personality.image_files, client.discussion.discussion_folder)
|
||||
post_processed_output = process_ai_output(output, client.discussion.image_files, client.discussion.discussion_folder)
|
||||
if len(post_processed_output)!=output:
|
||||
self.process_chunk(post_processed_output, MSG_TYPE.MSG_TYPE_FULL,client_id=client_id)
|
||||
except Exception as ex:
|
||||
@ -1198,48 +1199,23 @@ class LOLLMSWebUI(LOLLMSElfServer):
|
||||
use_deep_speed=self.config.xtts_use_deepspeed,
|
||||
use_streaming_mode=self.config.xtts_use_streaming_mode
|
||||
)
|
||||
language = convert_language_name(self.personality.language)
|
||||
self.tts.set_speaker_folder(Path(self.personality.audio_samples[0]).parent)
|
||||
fn = self.personality.name.lower().replace(' ',"_").replace('.','')
|
||||
fn = f"{fn}_{message_id}.wav"
|
||||
url = f"audio/{fn}"
|
||||
self.tts.tts_to_file(client.generated_text, Path(self.personality.audio_samples[0]).name, f"{fn}", language=language)
|
||||
fl = f"\n".join([
|
||||
f"<audio controls>",
|
||||
f' <source src="{url}" type="audio/wav">',
|
||||
f' Your browser does not support the audio element.',
|
||||
f'</audio>'
|
||||
])
|
||||
self.process_chunk("Generating voice output", MSG_TYPE.MSG_TYPE_STEP_END, {'status':True},client_id=client_id)
|
||||
self.process_chunk(fl,MSG_TYPE.MSG_TYPE_UI, client_id=client_id)
|
||||
|
||||
"""
|
||||
self.info("Creating audio output",10)
|
||||
self.personality.step_start("Creating audio output")
|
||||
if not PackageManager.check_package_installed("tortoise"):
|
||||
PackageManager.install_package("tortoise-tts")
|
||||
from tortoise import utils, api
|
||||
import sounddevice as sd
|
||||
if self.tts is None:
|
||||
self.tts = api.TextToSpeech( kv_cache=True, half=True)
|
||||
reference_clips = [utils.audio.load_audio(str(p), 22050) for p in self.personality.audio_samples]
|
||||
tk = self.model.tokenize(client.generated_text)
|
||||
if len(tk)>100:
|
||||
chunk_size = 100
|
||||
|
||||
for i in range(0, len(tk), chunk_size):
|
||||
chunk = self.model.detokenize(tk[i:i+chunk_size])
|
||||
if i==0:
|
||||
pcm_audio = self.tts.tts_with_preset(chunk, voice_samples=reference_clips, preset='fast').numpy().flatten()
|
||||
else:
|
||||
pcm_audio = np.concatenate([pcm_audio, self.tts.tts_with_preset(chunk, voice_samples=reference_clips, preset='ultra_fast').numpy().flatten()])
|
||||
if self.tts.ready:
|
||||
language = convert_language_name(self.personality.language)
|
||||
self.tts.set_speaker_folder(Path(self.personality.audio_samples[0]).parent)
|
||||
fn = self.personality.name.lower().replace(' ',"_").replace('.','')
|
||||
fn = f"{fn}_{message_id}.wav"
|
||||
url = f"audio/{fn}"
|
||||
self.tts.tts_to_file(client.generated_text, Path(self.personality.audio_samples[0]).name, f"{fn}", language=language)
|
||||
fl = f"\n".join([
|
||||
f"<audio controls>",
|
||||
f' <source src="{url}" type="audio/wav">',
|
||||
f' Your browser does not support the audio element.',
|
||||
f'</audio>'
|
||||
])
|
||||
self.process_chunk("Generating voice output", MSG_TYPE.MSG_TYPE_STEP_END, {'status':True},client_id=client_id)
|
||||
self.process_chunk(fl,MSG_TYPE.MSG_TYPE_UI, client_id=client_id)
|
||||
else:
|
||||
pcm_audio = self.tts.tts_with_preset(client.generated_text, voice_samples=reference_clips, preset='fast').numpy().flatten()
|
||||
sd.play(pcm_audio, 22050)
|
||||
self.personality.step_end("Creating audio output")
|
||||
"""
|
||||
|
||||
|
||||
self.InfoMessage("xtts is not up yet.\nPlease wait for it to load then try again. This may take some time.")
|
||||
|
||||
except Exception as ex:
|
||||
ASCIIColors.error("Couldn't read")
|
||||
|
||||
BIN
voice_sample_110.wav
Normal file
BIN
voice_sample_110.wav
Normal file
Binary file not shown.
File diff suppressed because one or more lines are too long
File diff suppressed because one or more lines are too long
4
web/dist/index.html
vendored
4
web/dist/index.html
vendored
@ -6,8 +6,8 @@
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
||||
<title>LoLLMS WebUI - Welcome</title>
|
||||
<script type="module" crossorigin src="/assets/index-a47cf3cb.js"></script>
|
||||
<link rel="stylesheet" href="/assets/index-ea2b1cac.css">
|
||||
<script type="module" crossorigin src="/assets/index-9dd9fa43.js"></script>
|
||||
<link rel="stylesheet" href="/assets/index-69152375.css">
|
||||
</head>
|
||||
<body>
|
||||
<div id="app"></div>
|
||||
|
||||
@ -865,23 +865,6 @@
|
||||
</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td style="min-width: 200px;">
|
||||
<label for="current_language" class="text-sm font-bold" style="margin-right: 1rem;">Force AI to answer in this language:</label>
|
||||
</td>
|
||||
<td>
|
||||
<div class="flex flex-row">
|
||||
<input
|
||||
type="text"
|
||||
id="current_language"
|
||||
required
|
||||
v-model="configFile.current_language"
|
||||
@change="settingsChanged=true"
|
||||
class="mt-1 px-2 py-1 border border-gray-300 rounded dark:bg-gray-600"
|
||||
>
|
||||
</div>
|
||||
</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td style="min-width: 200px;">
|
||||
<label for="fun_mode" class="text-sm font-bold" style="margin-right: 1rem;">Fun mode:</label>
|
||||
</td>
|
||||
@ -1101,23 +1084,6 @@
|
||||
</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td style="min-width: 200px;">
|
||||
<label for="summerize_discussion" class="text-sm font-bold" style="margin-right: 1rem;">Activate Continuous Learning from discussions:</label>
|
||||
</td>
|
||||
<td>
|
||||
<div class="flex flex-row">
|
||||
<input
|
||||
type="checkbox"
|
||||
id="summerize_discussion"
|
||||
required
|
||||
v-model="configFile.summerize_discussion"
|
||||
@change="settingsChanged=true"
|
||||
class="mt-1 px-2 py-1 border border-gray-300 rounded dark:bg-gray-600"
|
||||
>
|
||||
</div>
|
||||
</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td style="min-width: 200px;">
|
||||
<label for="data_vectorization_visualize_on_vectorization" class="text-sm font-bold" style="margin-right: 1rem;">show vectorized data:</label>
|
||||
</td>
|
||||
@ -1983,11 +1949,11 @@
|
||||
|
||||
<tr>
|
||||
<td style="min-width: 200px;">
|
||||
<label for="current_voice" class="text-sm font-bold" style="margin-right: 1rem;">Current voice:</label>
|
||||
<label for="xtts_current_voice" class="text-sm font-bold" style="margin-right: 1rem;">Current voice:</label>
|
||||
</td>
|
||||
<td>
|
||||
<div class="flex flex-row">
|
||||
<select v-model="current_voice" @change="settingsChanged=true" :disabled="!enable_voice_service">
|
||||
<select v-model="xtts_current_voice" @change="settingsChanged=true" :disabled="!enable_voice_service">
|
||||
<option v-for="voice in voices" :key="voice" :value="voice">
|
||||
{{ voice }}
|
||||
</option>
|
||||
@ -3460,10 +3426,10 @@ export default {
|
||||
uploadLogo(event){
|
||||
const file = event.target.files[0]; // Get the selected file
|
||||
const formData = new FormData(); // Create a FormData object
|
||||
formData.append('avatar', file); // Add the file to the form data with the key 'avatar'
|
||||
formData.append('logo', file); // Add the file to the form data with the key 'avatar'
|
||||
console.log("Uploading avatar")
|
||||
// Make an API request to upload the avatar
|
||||
axios.post('/upload_avatar', formData)
|
||||
axios.post('/upload_logo', formData)
|
||||
.then(response => {
|
||||
console.log("Logo uploaded successfully")
|
||||
|
||||
@ -3472,9 +3438,8 @@ export default {
|
||||
const fileName = response.data.fileName;
|
||||
console.log("response",response);
|
||||
this.app_custom_logo = fileName; // Update the user_avatar value with the file name
|
||||
this.configFile.app_custom_logo=fileName;
|
||||
|
||||
this.update_setting("app_custom_logo", fileName, ()=>{}).then(()=>{})
|
||||
this.$store.state.config.app_custom_logo=fileName;
|
||||
this.settingsChanged = true
|
||||
})
|
||||
.catch(error => {
|
||||
console.error('Error uploading avatar:', error);
|
||||
@ -3495,8 +3460,8 @@ export default {
|
||||
const fileName = response.data.fileName;
|
||||
console.log("response",response);
|
||||
this.user_avatar = fileName; // Update the user_avatar value with the file name
|
||||
this.configFile.user_avatar=fileName;
|
||||
this.update_setting("user_avatar", fileName, ()=>{}).then(()=>{})
|
||||
this.$store.state.config.user_avatar=fileName;
|
||||
this.settingsChanged = true
|
||||
})
|
||||
.catch(error => {
|
||||
console.error('Error uploading avatar:', error);
|
||||
@ -5532,23 +5497,23 @@ export default {
|
||||
this.$store.state.config.current_language = value
|
||||
},
|
||||
},
|
||||
current_voice:{
|
||||
xtts_current_voice:{
|
||||
get() {
|
||||
if (this.$store.state.config.current_voice===null || this.$store.state.config.current_voice===undefined){
|
||||
console.log("current voice", this.$store.state.config.current_voice)
|
||||
if (this.$store.state.config.xtts_current_voice===null || this.$store.state.config.xtts_current_voice===undefined){
|
||||
console.log("current voice", this.$store.state.config.xtts_current_voice)
|
||||
return "main_voice";
|
||||
}
|
||||
return this.$store.state.config.current_voice;
|
||||
return this.$store.state.config.xtts_current_voice;
|
||||
},
|
||||
set(value) {
|
||||
// You should not set the value directly here; use the updateSetting method instead
|
||||
if(value=="main_voice" || value===undefined){
|
||||
console.log("Current voice set to None")
|
||||
this.$store.state.config.current_voice = null
|
||||
this.$store.state.config.xtts_current_voice = null
|
||||
}
|
||||
else{
|
||||
console.log("Current voice set to ",value)
|
||||
this.$store.state.config.current_voice = value
|
||||
this.$store.state.config.xtts_current_voice = value
|
||||
}
|
||||
},
|
||||
},
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user