lollms/configs/config.yaml

# =================== Lord Of Large Language Models Configuration file =========================== 
version: 46
binding_name: null
model_name: null



# Host information
host: localhost
port: 9600

# Genreration parameters 
discussion_prompt_separator: "!@>"
seed: -1
n_predict: 1024
ctx_size: 4084
min_n_predict: 512
temperature: 0.9
top_k: 50
top_p: 0.95
repeat_last_n: 40
repeat_penalty: 1.2

n_threads: 8

#Personality parameters
personalities: ["generic/lollms"]
active_personality_id: 0
override_personality_model_parameters: false #if true the personality parameters are overriden by those of the configuration (may affect personality behaviour) 

extensions: []

user_name: user
user_description: ""
use_user_name_in_discussions: false
user_avatar: default_user.svg
use_user_informations_in_discussion: false

# UI parameters
db_path: database.db

# Automatic updates
debug: False
auto_update: true
auto_sync_personalities: true
auto_sync_extensions: true
auto_sync_bindings: true
auto_sync_models: true



auto_save: true
auto_title: false
# Install mode (cpu, cpu-noavx, nvidia-tensorcores, nvidia, amd-noavx, amd, apple-intel, apple-silicon)
hardware_mode: nvidia-tensorcores
# Automatically open the browser
auto_show_browser: true

# Voice service
enable_voice_service: false
xtts_base_url: http://127.0.0.1:8020
auto_read: false
current_voice: null
current_language: en

# Image generation service
enable_sd_service: false
sd_base_url: http://127.0.0.1:7860

# ollama service
enable_ollama_service: false
ollama_base_url: http://0.0.0.0:11434

# petals service
enable_petals_service: false
petals_base_url: http://0.0.0.0:8010

# lollms service
enable_lollms_service: false
lollms_base_url: http://0.0.0.0:1234

# Audio
media_on: false
audio_in_language: 'en-US'
auto_speak: false
audio_out_voice: null
audio_pitch: 1
audio_auto_send_input: true
audio_silenceTimer: 5000

# Data vectorization
use_discussions_history: false # Activate vectorizing previous conversations
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"
data_vectorization_save_db: False # For each new session, new files
data_vectorization_chunk_size: 512 # chunk size
data_vectorization_overlap_size: 128 # overlap between chunks size
data_vectorization_nb_chunks: 2 # number of chunks to use
data_vectorization_build_keys_words: false # If true, when querrying the database, we use keywords generated from the user prompt instead of the prompt itself.
data_vectorization_force_first_chunk: false # If true, the first chunk of the document will systematically be used
data_vectorization_make_persistance: false # If true, the data will be persistant webween runs


# Helpers
pdf_latex_path: null

# boosting information
positive_boost: null
negative_boost: null
force_output_language_to_be: null
fun_mode: False