You can deploy this model and start earning money today!

Ameya-Kawade/
cmdcaliper

Name: cmdcaliper
Availability: InStock

sentence-transformers

text-embeddings-inference

endpoints_compatible

Views

No views yet

cmdcaliper — Linux Command Intent Engine

This model is a fine-tuned version of thenlper/gte-base optimized for Linux Command Intent and Semantic Similarity, following the methodology of the CmdCaliper research paper (Empirical Software Engineering & Cybersecurity).

Model Details

Architecture: Transformer + Masked Mean Pooling + L2 Normalization (SentenceTransformer)
Base Model: thenlper/gte-base
Training Objective: InfoNCE / Multiple Negatives Ranking Loss (MNRL) with strict collision-free batch sampling
Max Sequence Length: 128 tokens
Temperature (τ): 0.05
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity

Usage with Sentence-Transformers

python

from sentence_transformers import SentenceTransformer, util

# Load model
model = SentenceTransformer("Ameya-Kawade/cmdcaliper")

# Commands to compare
commands = [
    "find /var/log -type f -name '*.log'",
    "ls -la /var/log/*.log",
    "apt-get update && apt-get upgrade -y",
    "docker system prune -a --volumes",
]

# Compute embeddings
embeddings = model.encode(commands, normalize_embeddings=True)

# Compute similarity matrix
similarity = util.cos_sim(embeddings, embeddings)
print(similarity)

Intended Use

Detecting semantically equivalent bash / shell commands.
Linux command recommendation and autocomplete intent matching.
Security anomaly detection and command telemetry analysis.