SlurmGenie/.env.example at main · OriAlpha/SlurmGenie · GitHub

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
82
83
84
85
86
87
88
89
90
# SlurmGenie configuration template
#
# Copy this file to `.env` in your working directory and uncomment / edit
# the values you need. All variables can also be exported in your shell.
# Anything left commented out falls back to the defaults shown below.
#
# Reference: src/slurmgenie/models/config.py is the single source of truth
# for defaults and validation.

# ---------------------------------------------------------------------- #
# Slurm binaries
# ---------------------------------------------------------------------- #
# Path to each Slurm CLI tool. Defaults assume they are on $PATH.
#SLURMGENIE_SACCT_PATH=sacct
#SLURMGENIE_SQUEUE_PATH=squeue
#SLURMGENIE_SCONTROL_PATH=scontrol
#SLURMGENIE_SSTAT_PATH=sstat

# Where Slurm writes job logs by default. SlurmGenie searches here when
# `scontrol show job` doesn't return an explicit StdOut path.
#SLURMGENIE_SLURM_LOG_DIR=/var/log/slurm

# ---------------------------------------------------------------------- #
# NVIDIA / GPU monitoring
# ---------------------------------------------------------------------- #
#SLURMGENIE_NVIDIA_SMI_PATH=nvidia-smi

# GPU utilization (%) below which a GPU is considered idle (0.0 - 100.0)
#SLURMGENIE_GPU_IDLE_THRESHOLD=10.0

# ---------------------------------------------------------------------- #
# LLM backend selection
# ---------------------------------------------------------------------- #
# Which provider to use: local | ollama | openai | anthropic
#SLURMGENIE_LLM_BACKEND=local

# Shared LLM tuning (applies to every backend)
#SLURMGENIE_LLM_CONTEXT_LENGTH=2048
#SLURMGENIE_LLM_MAX_TOKENS=512

# ---------------------------------------------------------------------- #
# Local backend (llama-cpp-python + GGUF)
# Requires the `llm` extra:  pip install -e ".[llm]"
# ---------------------------------------------------------------------- #
#SLURMGENIE_LLM_REPO_ID=Qwen/Qwen2.5-0.5B-Instruct-GGUF
#SLURMGENIE_LLM_MODEL_FILE=qwen2.5-0.5b-instruct-q4_k_m.gguf

# ---------------------------------------------------------------------- #
# Ollama backend
# Run a local Ollama server and `ollama pull <model>` first.
# ---------------------------------------------------------------------- #
#SLURMGENIE_OLLAMA_BASE_URL=http://localhost:11434
#SLURMGENIE_OLLAMA_MODEL=llama3.1:8b

# ---------------------------------------------------------------------- #
# OpenAI backend
# Requires the `cloud` extra:  pip install -e ".[cloud]"
# ---------------------------------------------------------------------- #
#SLURMGENIE_OPENAI_API_KEY=sk-...
#SLURMGENIE_OPENAI_MODEL=gpt-4o-mini
# Optional override for Azure OpenAI / proxies (leave empty for default)
#SLURMGENIE_OPENAI_BASE_URL=

# ---------------------------------------------------------------------- #
# Anthropic backend
# Requires the `cloud` extra:  pip install -e ".[cloud]"
# ---------------------------------------------------------------------- #
#SLURMGENIE_ANTHROPIC_API_KEY=sk-ant-...
#SLURMGENIE_ANTHROPIC_MODEL=claude-sonnet-4-20250514

# ---------------------------------------------------------------------- #
# Slack notifications
# Requires the `slack` extra:  pip install -e ".[slack]"
# ---------------------------------------------------------------------- #
# Incoming webhook URL (leave unset to disable Slack)
#SLURMGENIE_SLACK_WEBHOOK_URL=https://hooks.slack.com/services/T.../B.../xxx
# Optional channel override (defaults to the webhook's configured channel)
#SLURMGENIE_SLACK_CHANNEL=#hpc-alerts

# ---------------------------------------------------------------------- #
# Cluster context (copilot / chat / generate)
# ---------------------------------------------------------------------- #
# Comma-separated partition names for topology hints. Leave empty to auto-discover.
#SLURMGENIE_DEFAULT_PARTITIONS=gpu,compute

# ---------------------------------------------------------------------- #
# General
# ---------------------------------------------------------------------- #
# Number of log lines per LLM request chunk
#SLURMGENIE_LOG_CHUNK_SIZE=500