rlm/examples/hermes_example.py at main · miroslavb/rlm · 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
91
"""
Example: Using RLM with Hermes-compatible endpoints (Nous / OpenRouter).

This example shows how to use the RLM framework with the Hermes ecosystem.
The "hermes" backend automatically resolves to Nous Research inference API
or OpenRouter based on configuration.

Requirements:
    pip install rlms[hermes]

Environment variables:
    NOUS_API_KEY          — API key for Nous Research inference API
    OPENROUTER_API_KEY    — API key for OpenRouter (fallback)

Usage:
    python examples/hermes_example.py
"""

from rlm import RLM


def main():
    # Example 1: Using Nous Research endpoint (default)
    print("=== Example 1: Nous Research (default) ===")
    rlm_nous = RLM(
        backend="hermes",
        backend_kwargs={
            "model_name": "xiaomi/mimo-v2-pro",
            # base_url defaults to Nous inference API
            # api_key resolved from NOUS_API_KEY env var
        },
        max_depth=3,
        max_iterations=15,
        verbose=True,
    )
    result = rlm_nous.completion(
        "Calculate the factorial of 20 recursively. Use rlm_query to break this into subproblems."
    )
    print(f"Response: {result.response}")
    print(f"Iterations: {result.metadata.total_iterations}")
    print()

    # Example 2: Using OpenRouter explicitly
    print("=== Example 2: OpenRouter ===")
    rlm_or = RLM(
        backend="hermes",
        backend_kwargs={
            "model_name": "xiaomi/mimo-v2-pro",
            "base_url": "openrouter",  # shorthand for OpenRouter URL
            # api_key resolved from OPENROUTER_API_KEY env var
        },
        max_depth=2,
        max_iterations=10,
        verbose=True,
    )
    result = rlm_or.completion(
        "Sort the list [38, 27, 43, 3, 9, 82, 10] using merge sort. "
        "Implement it recursively using rlm_query for each half."
    )
    print(f"Response: {result.response}")
    print()

    # Example 3: With context (data processing)
    print("=== Example 3: Context-based processing ===")
    large_text = """
    The quick brown fox jumps over the lazy dog.
    Machine learning is a subset of artificial intelligence.
    Python is a versatile programming language.
    Neural networks are inspired by biological neurons.
    Deep learning uses multiple layers of neural networks.
    Natural language processing enables computers to understand text.
    """
    rlm_ctx = RLM(
        backend="hermes",
        backend_kwargs={
            "model_name": "xiaomi/mimo-v2-pro",
            "base_url": "nous",
        },
        max_depth=2,
        max_iterations=10,
    )
    result = rlm_ctx.completion(
        "Analyze the text context. Find themes and summarize key points. "
        "Use rlm_query to analyze each sentence separately, then combine.",
        context=large_text,
    )
    print(f"Response: {result.response}")


if __name__ == "__main__":
    main()