File size: 3,395 Bytes
40b123c
ac35688
 
 
 
 
 
 
 
 
 
 
 
 
 
40b123c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ac35688
 
40b123c
ac35688
 
 
0990d08
40b123c
 
 
 
 
 
 
 
 
 
 
 
 
 
ac35688
dfce113
40b123c
ac35688
 
 
 
 
40b123c
 
 
 
 
ac35688
 
40b123c
 
 
ac35688
 
40b123c
 
44b65e5
40b123c
 
44b65e5
40b123c
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
# FILE: modules/config_settings_public.py (Final Verified Version)

import os
import sys
import logging
from huggingface_hub import hf_hub_download

logger = logging.getLogger("ZOTHEOS_Config")
if not logger.handlers:
    handler = logging.StreamHandler(sys.stdout)
    formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - [%(funcName)s] - %(message)s')
    handler.setFormatter(formatter)
    logger.addHandler(handler)
    logger.setLevel(logging.INFO)

IS_WEB_MODE = os.path.exists("/home/user/app")

# --- βœ…βœ…βœ… VERIFIED AND CORRECTED REPO IDS βœ…βœ…βœ… ---
MODEL_DEFINITIONS = {
    "mistral": {
        "repo_id": "TheBloke/Mistral-7B-Instruct-v0.2-GGUF",
        "filename": "mistral-7b-instruct-v0.2.Q4_K_M.gguf"
    },
    "gemma": {
        "repo_id": "TheBloke/Gemma-2B-IT-GGUF", # Corrected: Capital 'G' and 'IT'
        "filename": "gemma-2b-it.Q4_K_M.gguf"
    },
    "qwen": {
        "repo_id": "TheBloke/Qwen1.5-1.8B-Chat-GGUF", # Corrected to use TheBloke's repo
        "filename": "qwen1.5-1.8b-chat.Q4_K_M.gguf"
    }
}

MODEL_PATHS = {}

if IS_WEB_MODE:
    logger.info("βœ…βœ…βœ… RUNNING IN WEB MODE (Hugging Face Space) βœ…βœ…βœ…")
    N_GPU_LAYERS_FALLBACK = 0
    for name, model_info in MODEL_DEFINITIONS.items():
        logger.info(f"Downloading model: {name} from repo: {model_info['repo_id']}")
        try:
            MODEL_PATHS[name] = hf_hub_download(repo_id=model_info["repo_id"], filename=model_info["filename"])
            logger.info(f"βœ… Successfully downloaded {name}.")
        except Exception as e:
            logger.error(f"❌ FAILED to download model {name}. Error: {e}")
            raise e
else: # LOCAL MODE
    logger.info("βœ…βœ…βœ… RUNNING IN LOCAL MODE (Desktop/PC) βœ…βœ…βœ…")
    APP_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
    BASE_MODELS_DIR = os.path.join(APP_DIR, "models")
    for name, model_info in MODEL_DEFINITIONS.items():
        MODEL_PATHS[name] = os.path.join(BASE_MODELS_DIR, model_info["filename"])
    N_GPU_LAYERS_FALLBACK = -1

# --- Shared Configurations ---
MAX_CONCURRENT_MODELS = 3
N_CTX_FALLBACK = 2048
VERBOSE_LLAMA_CPP = True
MODEL_SPECIFIC_PARAMS = {
    "_default": {
        "n_gpu_layers": N_GPU_LAYERS_FALLBACK, "n_ctx": N_CTX_FALLBACK, "f16_kv": True,
        "use_mmap": True, "verbose": VERBOSE_LLAMA_CPP
    },
    "mistral": {"chat_format": "mistral-instruct"}, "gemma": {"chat_format": "gemma"},
    "qwen": {"chat_format": "chatml"}
}
INFERENCE_PRESETS = {
    "balanced": {"temperature": 0.7, "top_p": 0.9, "top_k": 40, "repeat_penalty": 1.1, "max_tokens": 1024},
    "precise": {"temperature": 0.2, "top_p": 0.7, "top_k": 20, "repeat_penalty": 1.05, "max_tokens": 1536},
    "creative": {"temperature": 0.9, "top_p": 0.95, "top_k": 60, "repeat_penalty": 1.15, "max_tokens": 1024}
}
DEFAULT_INFERENCE_PRESET = "balanced"
DEFAULT_SYSTEM_PROMPT = "You are ZOTHEOS, an ethical AI developed to help humanity."
MODEL_ROLES = {"mistral": "analyst", "gemma": "humanist", "qwen": "skeptic"}
MODEL_ROLE_SYSTEM_PROMPTS = {
    "analyst": "You are an impartial analyst.", "humanist": "You are a human-centered assistant.",
    "skeptic": "You are a critical evaluator and a respectful skeptic.", "general": DEFAULT_SYSTEM_PROMPT
}
ZOTHEOS_VERSION = "2.2 (Live and Corrected)"
logger.info(f"Config loaded. Version: {ZOTHEOS_VERSION}, Web Mode: {IS_WEB_MODE}")