-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy path.env.example
More file actions
81 lines (71 loc) · 3.51 KB
/
Copy path.env.example
File metadata and controls
81 lines (71 loc) · 3.51 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
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
# CHEN - Environment Configuration Example
# Copy this file to .env and fill in the values you need.
# Only set the variables for the backends you actually use.
# ---------------------------------------------------------------------------
# Backend selection
# ---------------------------------------------------------------------------
# Which backend to use when no explicit backend is passed:
# mock | hf | vllm | llama_cpp
CHEN_DEFAULT_BACKEND=mock
# ---------------------------------------------------------------------------
# HuggingFace backend (Phase 2 KV-cache passing is implemented here)
# ---------------------------------------------------------------------------
# Any valid HuggingFace model id, or a local path.
#
# RECOMMENDED MODEL CONFIGURATIONS (all open, no token required):
#
# CPU-only tier (any modern laptop, ~5 GB RAM):
# CHEN_HF_ANALYST_MODEL=HuggingFaceTB/SmolLM2-1.7B-Instruct
# CHEN_HF_REASONER_MODEL=Qwen/Qwen2.5-3B-Instruct
# CHEN_HF_SYNTHESIZER_MODEL=HuggingFaceTB/SmolLM2-1.7B-Instruct
# CHEN_HF_CODER_MODEL=Qwen/CodeQwen1.5-7B-Chat
#
# GPU tier (8GB+ VRAM):
# CHEN_HF_ANALYST_MODEL=HuggingFaceTB/SmolLM2-1.7B-Instruct
# CHEN_HF_REASONER_MODEL=meta-llama/Meta-Llama-3-8B-Instruct (needs token)
# CHEN_HF_SYNTHESIZER_MODEL=HuggingFaceTB/SmolLM2-1.7B-Instruct
# CHEN_HF_CODER_MODEL=Qwen/CodeQwen1.5-7B-Chat
#
# Tiny tier (super fast, lower quality — good for first verification):
# CHEN_HF_ANALYST_MODEL=HuggingFaceTB/SmolLM2-360M-Instruct
# CHEN_HF_REASONER_MODEL=HuggingFaceTB/SmolLM2-1.7B-Instruct
# CHEN_HF_SYNTHESIZER_MODEL=HuggingFaceTB/SmolLM2-360M-Instruct
CHEN_HF_ANALYST_MODEL=HuggingFaceTB/SmolLM2-1.7B-Instruct
CHEN_HF_REASONER_MODEL=Qwen/Qwen2.5-3B-Instruct
CHEN_HF_SYNTHESIZER_MODEL=HuggingFaceTB/SmolLM2-1.7B-Instruct
CHEN_HF_ROUTER_MODEL=HuggingFaceTB/SmolLM2-360M-Instruct
# Device: auto | cpu | cuda | mps
CHEN_HF_DEVICE=auto
# Optional: HuggingFace access token (for gated models like Llama-3)
# Get one at https://huggingface.co/settings/tokens
HUGGING_FACE_HUB_TOKEN=
# ---------------------------------------------------------------------------
# vLLM backend (stub - requires CUDA GPU)
# ---------------------------------------------------------------------------
CHEN_VLLM_MODEL=meta-llama/Meta-Llama-3-8B-Instruct
CHEN_VLLM_TENSOR_PARALLEL_SIZE=1
CHEN_VLLM_GPU_MEMORY_UTILIZATION=0.85
# ---------------------------------------------------------------------------
# llama.cpp backend (stub - requires GGUF models)
# ---------------------------------------------------------------------------
CHEN_LLAMA_MODEL_PATH=./models/llama-3-8b-instruct.Q4_K_M.gguf
CHEN_LLAMA_N_CTX=4096
CHEN_LLAMA_N_GPU_LAYERS=0
# ---------------------------------------------------------------------------
# Shared External Memory (RAG+)
# ---------------------------------------------------------------------------
# Vector store backend: in_memory | chroma
CHEN_MEMORY_BACKEND=in_memory
CHEN_MEMORY_PERSIST_DIR=./chen_data/memory
# Embedding model (sentence-transformers id)
CHEN_MEMORY_EMBEDDING_MODEL=all-MiniLM-L6-v2
# ---------------------------------------------------------------------------
# Cost model (USD per 1M tokens) - override if you have negotiated rates
# ---------------------------------------------------------------------------
CHEN_COST_INPUT_PER_1M=0.15
CHEN_COST_OUTPUT_PER_1M=0.60
# ---------------------------------------------------------------------------
# Telemetry
# ---------------------------------------------------------------------------
CHEN_LOG_LEVEL=INFO
CHEN_ENABLE_TELEMETRY=false