Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.cirron.com/llms.txt

Use this file to discover all available pages before exploring further.

​
cirron info

Display comprehensive information about your ML model, including architecture details, dependencies, and metadata. The command also provides diagnostic checks and hardware detection.

​
Usage

cirron info [options]

​
Options

OptionDescriptionDefault
--update <type>Update metadata (metadata)null
--dry-runPreview changes without applyingfalse
--diagnosticsRun diagnostic checks on configuration and connectivityfalse
--hardwareShow hardware informationfalse
--jsonOutput in JSON formatfalse
--detailedShow detailed informationfalse

​
Basic Information Display

# Display model information
cirron info
The command provides detailed information about:

​
Basic Information

  • Model Type: Classification, regression, etc.
  • Framework: PyTorch, TensorFlow, Scikit-learn
  • Python Version: Required Python version
  • GPU Required: Whether GPU is needed

​
Model Details

  • Model Class: Detected model class name
  • Architecture: CNN, LSTM, Transformer, etc.
  • Parameters: Estimated parameter count
  • Input Shape: Expected input dimensions
  • Output Shape: Model output dimensions
  • Training/Test Data Shapes: Data dimensions

​
Version Control

  • Git Commit: Current commit hash
  • Branch: Current git branch
  • Repository Status: Clean or uncommitted changes

​
Associated Endpoints

  • Deployment URLs: Configured endpoint URLs
  • Environment Status: Production, staging, etc.

​
Key Dependencies

  • Framework Dependencies: torch, tensorflow, sklearn
  • Data Processing: numpy, pandas
  • Visualization: matplotlib, seaborn

​
Model Analysis

The command performs static analysis of your src/model.py file to extract:

​
Framework Detection

# PyTorch models
class MyModel(nn.Module):
    pass

# TensorFlow models  
class MyModel(tf.keras.Model):
    pass

# Scikit-learn models
model = RandomForestClassifier()

​
Architecture Pattern Recognition

  • CNN: Conv2d, Conv1d layers
  • LSTM/GRU: Recurrent neural networks
  • Transformer: Attention mechanisms
  • ResNet: Residual connections
  • Regularization: Dropout, BatchNorm
  • Embeddings: Word embeddings

​
Input Shape Detection

# Detected from sample calls
torch.randn(1, 3, 224, 224)  # Input shape: (1, 3, 224, 224)
np.random.randn(100, 10)     # Input shape: (100, 10)
tf.random.normal([32, 784])  # Input shape: (32, 784)

​
Metadata Management

​
Update Metadata

# Update metadata based on current analysis
cirron info --update metadata

# Preview changes without applying
cirron info --update metadata --dry-run
Updates the following metadata fields:
  • modelClassName: Detected model class
  • architecture: Recognized architecture patterns
  • inputShape: Detected input dimensions
  • trainingDataShape: Training data shape
  • testDataShape: Test data shape
  • gitCommitHash: Current git commit
  • detectedPatterns: Architecture patterns found

​
Metadata Mismatch Detection

The command automatically detects mismatches between stored metadata and current model analysis:

​
Critical Mismatches

  • Model Class Changes: Different model class detected
  • Input Shape Changes: Modified input dimensions

​
Warning Mismatches

  • Architecture Changes: New/different architecture patterns
  • Git Commit Changes: Code has been modified

​
Examples

​
Basic Information Display

# Display all model information
cirron info
Example output: 
Model Information: sentiment-analysis
──────────────────────────────────────────────────────────────────

Basic Information
  Model Type: classification
  Framework: pytorch
  Python Version: 3.9
  GPU Required: No

Model Details
  Model Class: SentimentClassifier
  Architecture: LSTM + Attention
  Parameters: 1,234,567
  Input Shape: (batch_size, sequence_length)
  Output Shape: (batch_size, num_classes)

Version Control
  Git Commit: a1b2c3d
  Branch: main
  Repository: Clean working directory

Associated Endpoints
  • production: https://api.example.com/v1/predict
  • staging: https://staging.example.com/v1/predict

Key Dependencies
  • torch
  • numpy
  • pandas

​
Update Metadata

# Update metadata with current analysis
cirron info --update metadata
Example output: 
Metadata Update Preview
──────────────────────────────────────────────────────────────────

  modelClassName:
    - SentimentClassifier
    + TransformerClassifier

  architecture:
    - LSTM + Attention
    + Transformer + Attention

  gitCommitHash:
    - a1b2c3d
    + e4f5g6h

✓ Successfully updated 3 metadata field(s) in cirron.yaml

​
Preview Changes

# See what would be updated without making changes
cirron info --update metadata --dry-run

​
Metadata Mismatch Warnings

Metadata Mismatch Detected
──────────────────────────────────────────────────────────────────

  • [critical] Model class changed: SentimentClassifier → TransformerClassifier
  • [warning] Architecture pattern changed: LSTM + Attention → Transformer + Attention
  • [warning] Git commit changed: a1b2c3d → e4f5g6h

Run cirron info --update metadata to refresh metadata.

​
Framework-Specific Analysis

​
PyTorch Analysis

# Detects nn.Module subclasses
class MyModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(784, 10)  # Output shape: 10

# Parameter counting
total_params = sum(p.numel() for p in model.parameters())

​
TensorFlow Analysis

# Detects Keras models
class MyModel(tf.keras.Model):
    def __init__(self):
        super().__init__()
        self.dense = tf.keras.layers.Dense(10)  # Output shape: 10

# Model summary
model.summary()  # Parameter count available

​
Scikit-Learn Analysis

# Detects sklearn models
from sklearn.ensemble import RandomForestClassifier
model = RandomForestClassifier()

# Fit method calls
model.fit(X_train, y_train)  # Shape depends on training data

​
Configuration

​
cirron.yaml Metadata Section

{
  "name": "my-model",
  "framework": "pytorch",
  "metadata": {
    "modelClassName": "MyModel",
    "architecture": "CNN + BatchNorm",
    "inputShape": "(1, 3, 224, 224)",
    "trainingDataShape": "(1000, 3, 224, 224)",
    "testDataShape": "(200, 3, 224, 224)",
    "gitCommitHash": "a1b2c3d",
    "detectedPatterns": ["CNN", "BatchNorm"],
    "lastUpdated": "2024-01-15T10:30:00Z"
  }
}

​
Error Handling

​
Common Errors

​
Not a Cirron Project

Error: Not a Cirron project. Run this command in a directory with cirron.yaml

​
Unknown Update Type

Error: Unknown update type: invalid. Available options: metadata

​
Concurrent File Changes

Error: cirron.yaml has been modified by another process. Please retry the update.

​
Recovery Actions

  1. Missing Project Configuration 
    cirron init my-project
  2. Invalid Update Type 
    # Use correct update type
cirron info --update metadata
  3. Concurrent Changes 
    # Retry after other processes complete
cirron info --update metadata

​
Integration

​
CI/CD Pipeline

# GitHub Actions example
- name: Update Model Metadata
  run: |
    cirron info --update metadata
    git add cirron.yaml
    git commit -m "Update model metadata"

​
Pre-deployment Validation

# Check model information before deployment
cirron info
cirron test --model

​
Diagnostics

Run diagnostic checks on your configuration, settings, and connectivity. 
# Run diagnostics
cirron info --diagnostics

# Diagnostics with JSON output
cirron info --diagnostics --json

# Detailed diagnostics
cirron info --diagnostics --detailed
Diagnostics checks include:
  • Configuration: Validates cirron.yaml structure and required fields
  • Settings: Checks global and project settings for consistency
  • Connectivity: Tests API endpoint reachability and authentication
  • Environment: Verifies Python, CUDA, and dependency availability

​
Hardware Info

Quick hardware detection shortcut. This is equivalent to cirron config hardware --detect. 
# Show hardware information
cirron info --hardware

# Hardware info in JSON format
cirron info --hardware --json
Reports CPU, memory, GPU, CUDA availability, and framework compatibility. For full hardware configuration and profile management, use cirron config hardware.

​
Troubleshooting

​
Missing Model Information

  • Ensure src/model.py exists and is valid Python
  • Check that model class inherits from framework base classes
  • Verify import statements are correct

​
Inaccurate Parameter Counts

  • Run with model loaded for accurate counts
  • Check for dynamic parameter generation
  • Verify model instantiation works

​
Metadata Not Updating

  • Check file permissions on cirron.yaml
  • Ensure no other processes are modifying the file
  • Verify git repository is accessible

​
Architecture Detection Issues

  • Use standard framework patterns
  • Avoid custom layer implementations
  • Check for typos in layer names