Skip to content

DASHBOARD_FEATURES.md

Latest commit

 

History

History
437 lines (342 loc) · 12 KB

DASHBOARD_FEATURES.md

File metadata and controls

437 lines (342 loc) · 12 KB
Raw

Mohawk Inference Engine - Professional Dashboard Guide

🎯 Overview

The Mohawk Inference Engine GUI provides a comprehensive, LM Studio-style dashboard with enterprise-grade features for managing multi-device inference sessions securely.

📊 Dashboard Features (LM Studio + Mohawk Unique)

1. 📚 Model Library Manager

Similar to: LM Studio's model library
Mohawk Extensions: Multi-device layer splitting, PQC-secured workers

Features:

  • ✅ Model Browser - Browse available models with quantization options
  • ✅ Download Models - Download Safetensors/HF models directly
  • ✅ Upload Local Models - Load local model files
  • ✅ Quantization Selector - Q4_K_M, Q5_K_M, Q8_0, FP16 options
  • ✅ Device Split Configuration - Configure multi-device layer splitting: 
    Format: 'cpu_threads;gpu_ids'
Example: 'cpu;0,1,2,3;cuda:0,1'
  • ✅ Model Status Tracking - Ready/Loading/Failed states

Quick Actions:

# Download a model
download_btn.click()

# Upload local model
upload_btn.click()

# Load selected model
load_selected_model()

2. 💬 Chat Interface

Similar to: LM Studio's chat panel
Mohawk Extensions: Multi-session support, context management

Features:

  • ✅ Conversation History - Scrollable message history
  • ✅ Parameter Controls:
    • Temperature (0.0 - 2.0)
    • Top-p sampling
    • Max tokens generation
  • ✅ System Prompt Editor - Customizable system instructions
  • ✅ Context Management:
    • Context size tracking
    • Clear history button
    • Token usage monitoring
  • ✅ Multi-turn Conversations - Full context retention

Usage Examples:

# Send a message
message_input.setPlainText("What is quantum computing?")
send_message()

# Get response (automatically appends to chat)
# Response appears with latency metrics

3. 📊 Performance Metrics Dashboard

Similar to: LM Studio's stats panel
Mohawk Extensions: Real-time GPU/CPU/Multi-device tracking

Features:

  • ✅ Throughput Chart - Requests per second (real-time)
  • ✅ Latency Monitoring:
    • p50 latency (median)
    • p95 latency (95th percentile)
    • p99 latency (99th percentile)
  • ✅ Resource Usage Charts:
    • CPU utilization
    • Memory consumption
    • GPU utilization per device
  • ✅ Statistics Summary:
    • Total requests processed
    • Average latency
    • Success rate
    • Active sessions count
    • Peak throughput

Real-time Updates:

# Metrics update every second automatically
# Dashboard shows live charts and progress bars

4. 🔗 Session Manager

Similar to: LM Studio's session queue
Mohawk Extensions: Priority queuing, multi-worker support

Features:

  • ✅ Session Table - View all active sessions with:
    • Session ID
    • Model name
    • Status (Queued/Running/Completed)
    • Throughput metrics
    • Latency percentiles
    • Tokens per second
    • Start time
  • ✅ Queue Configuration:
    • Max queue size setting
    • Priority levels (High/Normal/Low)
  • ✅ Job Management:
    • Queue new jobs with priority
    • Cancel running sessions
    • Monitor session lifecycle

Quick Actions:

# Queue high-priority job
high_priority_btn.click()

# Queue normal job
normal_priority_btn.click()

# Cancel specific session
cancel_session(session_row)

5. ⚙️ Worker Configuration

Mohawk Unique Feature: Multi-device layer splitting management

Features:

  • ✅ Worker List - View all connected workers with:
    • Worker ID
    • Host:Port address
    • Connection status
    • Loaded model
    • GPU thread count
    • Current load percentage
  • ✅ Worker Actions:
    • Connect/Disconnect workers
    • Restart failed workers
    • Monitor worker health
  • ✅ Multi-device Configuration:
    • Configure layer splitting across devices
    • Set GPU thread counts per worker
    • Balance load across workers

Usage Example:

# Add new worker
add_worker(host="localhost", port=8004)

# Configure device splitting for worker
device_split_config.setText("cpu;0,1,2,3;cuda:0,1")

# Connect to worker
connect_to_worker("localhost", 8003)

6. 🔒 Security Center

Mohawk Unique Feature: Comprehensive security dashboard

Features:

  • ✅ JWT Authentication Status:
    • Token validity indicator
    • Expiry time display
    • Refresh token button
  • ✅ mTLS Configuration:
    • Certificate validity check
    • Key encryption status (Fernet)
    • Certificate path management
  • ✅ Post-Quantum Cryptography (PQC):
    • Hybrid KEM support toggle
    • liboqs integration status
    • X25519 key exchange configuration
  • ✅ Security Event Log - Immutable audit trail

Security Features:

# JWT Authentication (RS256)
# Token expiry: 24 hours
# Refresh window: 1 hour

# mTLS with client certificates
# Certificate encryption with Fernet

# PQC Hybrid Mode (Optional)
# X25519 + Kyber for quantum-resistant security

7. 📜 Conversation History

Similar to: LM Studio's history panel
Mohawk Extensions: Model usage tracking, analytics

Features:

  • ✅ History Table - View all conversations with:
    • Timestamp
    • Model used
    • Tokens consumed
    • Duration
    • Status (Completed/Failed)
  • ✅ Usage Statistics:
    • Total tokens processed
    • Average latency across sessions
    • Models usage count
  • ✅ Search & Filter - Find specific conversations

🎨 UI Layout

┌─────────────────────────────────────────────────────────────┐
│ 🦅 Mohawk Inference Engine v2.1.0                           │
├─────────────────────────────────────────────────────────────┤
│                                                              │
│  ┌──────────────────────────────────────────────────────┐  │
│  │ 📚 Model Library    | 💬 Chat Interface              │  │
│  ├──────────────────────────────────────────────────────┤  │
│  │                                                        │  │
│  │  [Tabs: Models | Chat | Metrics | Sessions           │  │
│  │          | Workers | Security | History]             │  │
│  │                                                        │  │
│  └──────────────────────────────────────────────────────┘  │
│                                                              │
│  ┌──────────────────────────────────────────────────────┐  │
│  │ Navigation Toolbar                                    │  │
│  │ [📚 Models] [💬 Chat] [📊 Metrics] ...                │  │
│  └──────────────────────────────────────────────────────┘  │
│                                                              │
│  ┌──────────────────────────────────────────────────────┐  │
│  | Status Bar: "Throughput: 1,250 req/s | Latency..."   │  │
│  └──────────────────────────────────────────────────────┘  │
└─────────────────────────────────────────────────────────────┘

🚀 Quick Start Guide

Installation

# Clone repository
cd C:\Users\rwill\Mohawk-Inference-Engine

# Create virtual environment
python -m venv venv
venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Generate authentication key (first run)
mkdir -p certs
python mohawk_gui/main.py --key-file certs/auth_key.pem

Running the Application

# Development mode
python mohawk_gui/main.py

# Production mode with SSL
python mohawk_gui/main.py \
    --host 0.0.0.0 \
    --port 8003 \
    --ssl-enabled \
    --key-file certs/auth_key.pem

Building Executable

# Windows build
build_windows.bat

# Or use PyInstaller directly
pyinstaller \
    --name=Mohawk-Inference-Engine \
    --onefile \
    --windowed \
    mohawk_gui/main.py

🎯 Key Shortcuts

Shortcut Action
Ctrl+Shift+M Show/Hide Main Window (tray icon)
Shift+Enter New line in chat input
Enter Send message
F1 Show help

🔧 Configuration Options

Command Line Arguments

--host <host>        # Bind address (default: localhost)
--port <port>        # Port number (default: 8003)
--key-file <path>    # Auth key file path
--ssl-enabled        # Enable SSL/TLS
--metrics-interval <ms>  # Update interval (default: 1000ms)

config.toml Settings

[mohawk]
host = "localhost"
port = 8003
ssl_enabled = false

[workers]
enabled = true
auto_discover = false
timeout_ms = 5000
max_connections = 100

[sessions]
max_concurrent = 10
default_batch_size = 32

[metrics]
sampling_rate = 0.1
buffer_window_size = 1000

[security]
jwt_expiry_hours = 24
refresh_window_hours = 1
audit_enabled = true

📊 Dashboard Screenshots (Feature Map)

Tab 1: Model Library

  • Top: Search bar with model type/quant filters
  • Middle: Model table with status indicators
  • Bottom: Model details panel (load config)

Tab 2: Chat Interface

  • Left: Conversation history scroll area
  • Right: Settings panel (temp, top-p, max tokens)

Tab 3: Performance Metrics

  • Top: Throughput chart (PyQtGraph)
  • Middle: Latency percentiles display
  • Bottom: Resource usage bars (CPU/Mem/GPU)

Tab 4: Session Manager

  • Top: Active sessions table
  • Bottom: Queue configuration panel

Tab 5: Worker Configuration

  • Top: Workers list with status
  • Bottom: Add/Configure worker settings

Tab 6: Security Center

  • Top: Security status cards (JWT, mTLS, PQC)
  • Bottom: Security event log

Tab 7: Conversation History

  • Top: Usage history table
  • Bottom: Statistics summary

🎓 Best Practices

Model Loading

  1. Start with quantized models (Q4_K_M or Q5_K_M) for better memory efficiency
  2. Configure device splitting for multi-GPU setups
  3. Monitor GPU utilization in Metrics tab

Chat Usage

  1. Set temperature lower (0.5-0.7) for deterministic responses
  2. Increase max tokens for longer outputs
  3. Use system prompt to customize behavior

Worker Management

  1. Connect workers before loading models
  2. Monitor worker load in Workers tab
  3. Restart workers if latency increases

Security Setup

  1. Always enable JWT authentication
  2. Configure mTLS for production deployments
  3. Enable PQC for quantum-resistant security (optional)

🐛 Troubleshooting

Common Issues

Issue: "Import error: PyQt6"
Solution: pip install PyQt6 pyqtgraph

Issue: "No models loaded"
Solution: Use Model Library → Download or Upload models

Issue: "Worker connection failed"
Solution: Check worker is running on correct port (default: 8003)

Issue: "SSL certificate error"
Solution: Generate certificates in certs/ directory

📞 Support

For issues and questions:

Mohawk Inference Engine v2.1.0 - Production Ready! 🦅