-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathconfig.example.py
More file actions
60 lines (54 loc) · 2.25 KB
/
Copy pathconfig.example.py
File metadata and controls
60 lines (54 loc) · 2.25 KB
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
# config.example.py
# RENAMED from config.py -> config.example.py
# INSTRUCTIONS: Rename this file to 'config.py' and add your API keys.
# =====================================================
# 1. CLOUD SETTINGS (RunPod Integration)
# =====================================================
USE_RUNPOD = True # Set to False to force Offline Only Mode
API_KEY = "YOUR_RUNPOD_API_KEY_HERE"
POD_ID = "YOUR_POD_ID_HERE"
# Base URL constructed from Pod ID
RUNPOD_BASE_URL = f"https://{POD_ID}-8000.proxy.runpod.net/v1"
# =====================================================
# 2. THE HYBRID MAP (Model selection menu)
# =====================================================
# FORMAT: "Menu Display Name": "Local Filename"
# 1. The Key is what you see in the boot menu.
# 2. The Value is the filename in your 'models/' folder.
# NOTE: For Cloud Mode, the system will try to resolve a HuggingFace ID from this filename.
MODEL_MAP = {
"sophosympatheia/Midnight-Miqu-70B": "Midnight-Miqu-70B.awq",
"Qwen/Qwen2.5-72B-Instruct-AWQ": "Qwen2.5-72B-Instruct.awq",
"openai/gpt-oss-20b": "gpt-oss-20b-Q2_K.gguf",
"Qwen/Qwen3-VL-8B-Instruct": "Qwen3-VL-8B-Instruct.gguf",
"Qwen/Qwen2.5-14B-Instruct": "Qwen2.5-14B-Q4_K_M.gguf",
"dphn/Dolphin-Mistral-24B-Venice-Edition": "Dolphin-Mistral-24B-Venice-Edition-q4_k_m.gguf",
"meta-llama/Meta-Llama-3-8B-Instruct": "Llama-3-8B-Q5.gguf",
"mistralai/Mistral-7B-Instruct-v0.3": "Mistral-7B-v0.3.gguf"
}
# =====================================================
# 3. GPU HARDWARE PROFILES (Smart Cloud Selection)
# =====================================================
# Define which GPUs are suitable for which models.
GPU_TIERS = {
"tier_standard": [ # 48GB Cards (Good for <40B models)
"NVIDIA A40",
"NVIDIA RTX A6000",
"NVIDIA RTX 6000 Ada",
"NVIDIA A100 80GB PCIe" # Overflow option
],
"tier_ultra": [ # 80GB+ Cards (Required for 70B+ models)
"NVIDIA A100 80GB PCIe",
"NVIDIA A100-SXM4-80GB",
"NVIDIA H100 80GB HBM3"
]
}
MODEL_SPECIFIC_TIERS = {
"cecibas/Midnight-Miqu-70B-v1.5-4bit": "tier_ultra",
"Qwen/Qwen2.5-72B-Instruct-AWQ": "tier_ultra",
}
KEYWORD_TIERS = [
("70b", "tier_ultra"),
("72b", "tier_ultra"),
("miqu", "tier_ultra"),
]