Quantization: Running Models on Edge Devices

Quantization: Compressing Models for Mobile and Embedded Devices

Neural networks typically use 32-bit floating-point (FP32) arithmetic, where each weight and activation uses 4 bytes of memory. Quantization reduces this to lower precision (INT8 uses 1 byte, INT4 uses 4 bits), dramatically decreasing model size and computational requirements. This is essential for deploying AI on resource-constrained devices like smartphones, smartwatches, and IoT sensors.

Quantization Basics

Quantization maps float values to integers using a scale and zero-point: quantized_value = round((float_value - zero_point) / scale). For INT8 (8-bit signed integers), values range from -128 to 127. If activations in a layer range from -2.5 to 2.5, scale = 5.0 / 255 and zero_point = 128 (center of INT8 range). During inference, you perform integer arithmetic (add, multiply) in the integer domain, then dequantize back to floats if needed. This eliminates expensive floating-point operations, which consume significant power and latency on mobile devices.

Post-Training Quantization vs Quantization-Aware Training

Post-training quantization (PTQ) quantizes a pre-trained FP32 model without retraining. It's fast and simple but may lose accuracy, especially with extreme quantization (INT4). The key is choosing good scale and zero-point; common approaches include min-max (use observed min/max values) and KL-divergence (minimize information loss). Quantization-aware training (QAT) simulates quantization during training, allowing the model to learn weights robust to quantization. QAT typically recovers accuracy lost during PTQ at the cost of retraining time.

Per-Layer vs Per-Channel Quantization

Per-layer quantization uses a single scale/zero-point for all weights in a layer. Per-channel quantization uses separate scale/zero-point for each output channel in convolutional layers. Per-channel is more flexible (accounts for different weight distributions across channels) but requires more metadata. For INT8, per-channel adds negligible overhead. For INT4, per-channel becomes essential for maintaining accuracy.

Running Models on Raspberry Pi and Mobile

A Raspberry Pi 4 has 4GB RAM and a 1.5 GHz quad-core CPU. A typical INT8-quantized MobileNetV2 (3.5 MB, 100M multiply-add operations) runs inference in ~500ms, achieving 5-10 images per second. With INT4 quantization, model size drops to 2 MB. TensorFlow Lite (Google's mobile ML framework) and ONNX Runtime support quantized models. For iOS, Core ML and Metal Perform Shaders accelerate quantized inference. On Android, NNAPI (Android Neural Networks API) provides hardware acceleration for Snapdragon processors.

Practical Deployment Example

Consider deploying a plant disease classifier on farmer's phones in India. Original model: ResNet-50, 98 MB, 5 second inference on phone. After pruning (remove 80% weights): 20 MB. After INT8 quantization: 5 MB, 0.5 second inference. Farmers can now run the model offline (no internet needed) with 10x faster inference. This combination of pruning and quantization is standard industry practice for mobile ML.

Engineering Perspective: Quantization: Running Models on Edge Devices

When you sit for a technical interview at any top company — whether it is Google, Microsoft, Amazon, or an Indian unicorn like Zerodha, Razorpay, or Meesho — they are not just testing whether you know the definition of quantization: running models on edge devices. They are testing whether you can APPLY these concepts to solve novel problems, whether you understand the TRADEOFFS involved, and whether you can reason about system behaviour at scale.

This chapter approaches quantization: running models on edge devices with that depth. We will examine not just what it is, but why it works the way it does, what alternatives exist and when to choose each one, and how real systems use these ideas in production. ISRO's mission control systems, India's UPI payment network handling 10 billion transactions per month, Aadhaar's biometric authentication serving 1.4 billion identities — all rely on the principles we discuss here.

Transformer Architecture: The Engine Behind GPT and Modern AI

The Transformer architecture, introduced in the landmark 2017 paper "Attention Is All You Need," revolutionised NLP and eventually all of deep learning. Here is the core mechanism:

# Self-Attention Mechanism (simplified)
import numpy as np

def self_attention(Q, K, V, d_k):
    """
    Q (Query): What am I looking for?
    K (Key):   What do I contain?
    V (Value): What do I actually provide?
    d_k:       Dimension of keys (for scaling)
    """
    # Step 1: Compute attention scores
    scores = np.matmul(Q, K.T) / np.sqrt(d_k)

    # Step 2: Softmax to get probabilities
    attention_weights = softmax(scores)

    # Step 3: Weighted sum of values
    output = np.matmul(attention_weights, V)
    return output

# Multi-Head Attention: Run multiple attention heads in parallel
# Each head learns different relationships:
# Head 1: syntactic relationships (subject-verb agreement)
# Head 2: semantic relationships (word meanings)
# Head 3: positional relationships (word order)
# Head 4: coreference (pronoun → noun it refers to)

The key insight of self-attention is that every token can attend to every other token simultaneously (unlike RNNs which process sequentially). This parallelism enables efficient GPU training. The computational complexity is O(n²·d) where n is sequence length and d is dimension, which is why context windows are a major engineering challenge.

State-of-the-art developments include: sparse attention (reducing O(n²) to O(n·√n)), mixture of experts (MoE — activating only a subset of parameters per input), retrieval-augmented generation (RAG — grounding responses in external documents), and constitutional AI (alignment through principles rather than RLHF alone). Indian researchers at institutions like IIT Bombay, IISc Bangalore, and Microsoft Research India are actively contributing to these frontiers.

India's Scale Challenges: Engineering for 1.4 Billion

Building technology for India presents unique engineering challenges that make it one of the most interesting markets in the world. UPI handles 10 billion transactions per month — more than all credit card transactions in the US combined. Aadhaar authenticates 100 million identities daily. Jio's network serves 400 million subscribers across 22 telecom circles. Hotstar streamed IPL to 50 million concurrent viewers — a world record. Each of these systems must handle India's diversity: 22 official languages, 28 states with different regulations, massive urban-rural connectivity gaps, and price-sensitive users expecting everything to work on ₹7,000 smartphones over patchy 4G connections. This is why Indian engineers are globally respected — if you can build systems that work in India, they will work anywhere.

Advanced Algorithms: Dynamic Programming and Graph Theory

Dynamic Programming (DP) solves complex problems by breaking them into overlapping subproblems. This is a favourite in competitive programming and interviews:

# Longest Common Subsequence — classic DP problem
# Used in: diff tools, DNA sequence alignment, version control

def lcs(s1, s2):
    m, n = len(s1), len(s2)
    dp = [[0] * (n + 1) for _ in range(m + 1)]

    for i in range(1, m + 1):
        for j in range(1, n + 1):
            if s1[i-1] == s2[j-1]:
                dp[i][j] = dp[i-1][j-1] + 1
            else:
                dp[i][j] = max(dp[i-1][j], dp[i][j-1])

    return dp[m][n]

# Dijkstra's Shortest Path — used by Google Maps!
import heapq

def dijkstra(graph, start):
    dist = {node: float('inf') for node in graph}
    dist[start] = 0
    pq = [(0, start)]  # (distance, node)

    while pq:
        d, u = heapq.heappop(pq)
        if d > dist[u]:
            continue
        for v, weight in graph[u]:
            if dist[u] + weight < dist[v]:
                dist[v] = dist[u] + weight
                heapq.heappush(pq, (dist[v], v))

    return dist

# Real use: Google Maps finding shortest route from
# Connaught Place to India Gate, considering traffic weights

Dijkstra's algorithm is how mapping applications find optimal routes. When you ask Google Maps to navigate from Mumbai to Pune, it models the road network as a weighted graph (intersections are nodes, roads are edges, travel time is weight) and runs a variant of Dijkstra's algorithm. Indian highways, city roads, and even railway networks can all be modelled this way. IRCTC's route optimisation for trains across 13,000+ stations uses graph algorithms at its core.

Research Frontiers and Open Problems in Quantization: Running Models on Edge Devices

Beyond production engineering, quantization: running models on edge devices connects to active research frontiers where fundamental questions remain open. These are problems where your generation of computer scientists will make breakthroughs.

Quantum computing threatens to upend many of our assumptions. Shor's algorithm can factor large numbers efficiently on a quantum computer, which would break RSA encryption — the foundation of internet security. Post-quantum cryptography is an active research area, with NIST standardising new algorithms (CRYSTALS-Kyber, CRYSTALS-Dilithium) that resist quantum attacks. Indian researchers at IISER, IISc, and TIFR are contributing to both quantum computing hardware and post-quantum cryptographic algorithms.

AI safety and alignment is another frontier with direct connections to quantization: running models on edge devices. As AI systems become more capable, ensuring they behave as intended becomes critical. This involves formal verification (mathematically proving system properties), interpretability (understanding WHY a model makes certain decisions), and robustness (ensuring models do not fail catastrophically on edge cases). The Alignment Research Center and organisations like Anthropic are working on these problems, and Indian researchers are increasingly contributing.

Edge computing and the Internet of Things present new challenges: billions of devices with limited compute and connectivity. India's smart city initiatives and agricultural IoT deployments (soil sensors, weather stations, drone imaging) require algorithms that work with intermittent connectivity, limited battery, and constrained memory. This is fundamentally different from cloud computing and requires rethinking many assumptions.

Finally, the ethical dimensions: facial recognition in public spaces (deployed in several Indian cities), algorithmic bias in loan approvals and hiring, deepfakes in political campaigns, and data sovereignty questions about where Indian citizens' data should be stored. These are not just technical problems — they require CS expertise combined with ethics, law, and social science. The best engineers of the future will be those who understand both the technical implementation AND the societal implications. Your study of quantization: running models on edge devices is one step on that path.

Key Vocabulary

Here are important terms from this chapter that you should know:

🏗️ Architecture Challenge

Design the backend for India's election results system. Requirements: 10 lakh (1 million) polling booths reporting simultaneously, results must be accurate (no double-counting), real-time aggregation at constituency and state levels, public dashboard handling 100 million concurrent users, and complete audit trail. Consider: How do you ensure exactly-once delivery of results? (idempotency keys) How do you aggregate in real-time? (stream processing with Apache Flink) How do you serve 100M users? (CDN + read replicas + edge computing) How do you prevent tampering? (digital signatures + blockchain audit log) This is the kind of system design problem that separates senior engineers from staff engineers.

The Frontier

You now have a deep understanding of quantization: running models on edge devices — deep enough to apply it in production systems, discuss tradeoffs in system design interviews, and build upon it for research or entrepreneurship. But technology never stands still. The concepts in this chapter will evolve: quantum computing may change our assumptions about complexity, new architectures may replace current paradigms, and AI may automate parts of what engineers do today.

What will NOT change is the ability to think clearly about complex systems, to reason about tradeoffs, to learn quickly and adapt. These meta-skills are what truly matter. India's position in global technology is only growing stronger — from the India Stack to ISRO to the startup ecosystem to open-source contributions. You are part of this story. What you build next is up to you.

Crafted for Class 10–12 • Model Deployment • Aligned with NEP 2020 & CBSE Curriculum