15. Multi-Head Attention (MHA) & Causal Masking Implementation from Scratch in Pure PyTorch • Deep Learning & PyTorch Neural Networks Architecture • Rishabh Ki Kaksha

Deep Learning & PyTorch Neural Networks Architecture › 15. Multi-Head Attention (MHA) & Causal Masking Implementation from Scratch in Pure PyTorch

55 mins

← Prev Next →

Course Syllabus

Module 1: PyTorch Tensors, Autograd & Computational Graphs

01. PyTorch Tensor Architecture: Strides, Storage Buffers, Memory Layout (Contiguous vs Non-Contiguous) 02. Dynamic Computational Graphs & The PyTorch Autograd Engine (grad_fn, requires_grad, backward) 03. GPU Acceleration with CUDA: Tensor Device Placement (to("cuda")), CUDA Streams & Pinned Memory

Module 2: Multi-Layer Perceptrons (MLPs) & Custom Training Loops

04. Building Custom nn.Module Classes, Custom Parameters & Activation Functions (ReLU, GELU, Swish) 05. Custom Datasets & High-Performance DataLoaders (num_workers, collate_fn, pin_memory) 06. The Enterprise Training Loop: Forward Pass, Loss Computation, Zero Grad, Backward Pass & Optimizer Step 07. Neural Regularization Techniques: Dropout, Batch Normalization (BatchNorm1d) & Layer Normalization (LayerNorm)

Module 3: Convolutional Neural Networks (CNNs) & Computer Vision Architecture

08. Convolutional Operations: Kernels, Filters, Strides, Padding & Feature Map Dilation 09. Modern CNN Architectures: Residual Connections (ResNet), Squeeze-and-Excitation & ConvNeXt 10. Transfer Learning & Vision Fine-Tuning with Torchvision Pretrained Models

Module 4: Recurrent Architectures: RNNs, LSTMs & GRUs for Sequential Data

11. Sequential Modeling: Vanilla RNNs & The Vanishing / Exploding Gradient Problem 12. Long Short-Term Memory (LSTM) Networks: Forget Gate, Input Gate, Candidate State & Output Gate Math 13. Gated Recurrent Units (GRUs) & Bidirectional Sequence Processing in PyTorch

Module 5: Modern Transformer Architecture from Scratch in PyTorch

14. Scaled Dot-Product Attention: Query, Key, Value Matrices & Attention Softmax Calculation 15. Multi-Head Attention (MHA) & Causal Masking Implementation from Scratch in Pure PyTorch 16. Building a Complete GPT-Style Decoder-Only Transformer Language Model from Scratch

Module 6: PyTorch Distributed (DDP), Mixed Precision (AMP) & Model Checkpointing

17. Automatic Mixed Precision (torch.cuda.amp.autocast & GradScaler) for 2x Faster GPU Training 18. Distributed Data Parallel (DDP) Architecture: Multi-GPU Synchronization with torch.distributed 19. Robust Checkpointing: Saving State Dicts, Resuming Training & Early Stopping Callbacks 20. Model Export & Production Serialization: TorchScript (JIT Tracing), ONNX & TensorRT Compilation

Interactive Lesson Studio

# 15. Multi-Head Attention (MHA) & Causal Masking Implementation from Scratch in Pure PyTorch

> Deep Learning & PyTorch Notes by Rishabh - Topic 15

---

### 1. Multi-Head Attention (MHA) vs MQA vs GQA 👑

*   **Multi-Head Attention (MHA - Standard GPT):** Har head ka apna independent $Q, K, V$ projection ($H$ heads).
*   **Multi-Query Attention (MQA):** Multiple $Q$ heads, par sirf **1 single shared $K$ and $V$ head** (Saves 8x KV Cache memory!).
*   **Grouped-Query Attention (GQA - Llama 3 / Mistral Standard ⚡):** $Q$ heads ko groups me divide karke har group ko 1 shared $K, V$ head assign karna!

---

### 2. Multi-Head Attention Module in PyTorch

```python
import torch
import torch.nn as nn
import math

class MultiHeadAttention(nn.Module):
    def __init__(self, d_model: int = 512, n_heads: int = 8):
        super().__init__()
        self.d_model = d_model
        self.n_heads = n_heads
        self.d_k = d_model // n_heads

self.W_q = nn.Linear(d_model, d_model, bias=False)
        self.W_k = nn.Linear(d_model, d_model, bias=False)
        self.W_v = nn.Linear(d_model, d_model, bias=False)
        self.W_o = nn.Linear(d_model, d_model, bias=False)

def forward(self, x: torch.Tensor, is_causal: bool = True) -> torch.Tensor:
        B, S, D = x.shape
        # Project & Reshape to [Batch, n_heads, SeqLen, d_k]:
        q = self.W_q(x).view(B, S, self.n_heads, self.d_k).transpose(1, 2)
        k = self.W_k(x).view(B, S, self.n_heads, self.d_k).transpose(1, 2)
        v = self.W_v(x).view(B, S, self.n_heads, self.d_k).transpose(1, 2)

# Scaled Dot-Product with Causal Mask:
        scores = torch.matmul(q, k.transpose(-2, -1)) / math.sqrt(self.d_k)
        if is_causal:
            mask = torch.tril(torch.ones(S, S, device=x.device)).view(1, 1, S, S)
            scores = scores.masked_fill(mask == 0, float('-inf'))

attn = torch.softmax(scores, dim=-1)
        out = torch.matmul(attn, v) # [B, n_heads, S, d_k]

# Concat heads & final projection:
        out = out.transpose(1, 2).contiguous().view(B, S, D)
        return self.W_o(out)
```

---

### 3. Master Comparison Matrix

---

### 4. Interview Ke 3 Sawal Is Topic Se

1. **Causal Masking (`torch.tril`) lower-triangular matrix kyu hoti hai?** -> Autoregressive models me current token ko future tokens ko attend karne se rokkne ke liye ($-\infty$ softmax me $0$ probability banata hai).
2. **RoPE (Rotary Position Embeddings) Transformer me positional info kaise inject karta hai?** -> Complex numbers 2D plane rotation matrix se Query aur Key vectors ko token distance ke anusaar rotate karta hai.
3. **KV Caching inference time me prompt processing ko $O(N)$ se $O(1)$ kaise banata hai?** -> Purane tokens ke computed Key aur Value vectors ko RAM me cache karke sirf naye generated single token ka Query vector multiply karna.

---
**Next Topic:** 16. Building a Complete GPT-Style Decoder-Only Transformer Language Model from Scratch

Loading lesson content...

Python 3.12

Full IDE

Live Execution Terminal

Click "Execute Live" to run code.
                

Lesson Mastery AI Quiz

AI Doubt Solver

CERTIFICATE OF COMPLETION

Rishabh Gehlot

PHP 8+ & Laravel 11 Master Track