general.py

import torch
import torch.nn as nn
import torch.nn.functional as F
import math
 
class TransformerEncoder(nn.Module):
    def __init__(self, d_model, nhead, num_layers, dim_feedforward=2048, dropout=0.1):
        super(TransformerEncoder, self).__init__()
        self.d_model = d_model
        self.embedding = nn.Embedding(1000, d_model)  # Toy vocabulary size is 1000
        self.pos_encoder = PositionalEncoding(d_model, dropout)
        encoder_layers = nn.TransformerEncoderLayer(d_model, nhead, dim_feedforward, dropout)
        self.transformer_encoder = nn.TransformerEncoder(encoder_layers, num_layers)
        self.fc_out = nn.Linear(d_model, 1)  # Example output layer

    def forward(self, src):
        src = self.embedding(src) * math.sqrt(self.d_model)
        src = self.pos_encoder(src)
        output = self.transformer_encoder(src)
        output = output.mean(dim=1)  # Average over all tokens
        return self.fc_out(output)

class PositionalEncoding(nn.Module):
    def __init__(self, d_model, dropout=0.1, max_len=5000):
        super(PositionalEncoding, self).__init__()
        self.dropout = nn.Dropout(p=dropout)

        pe = torch.zeros(max_len, d_model)
        position = torch.arange(0, max_len, dtype=torch.float).unsqueeze(1)
        div_term = torch.exp(torch.arange(0, d_model, 2).float() * (-math.log(10000.0) / d_model))
        pe[:, 0::2] = torch.sin(position * div_term)
        pe[:, 1::2] = torch.cos(position * div_term)
        pe = pe.unsqueeze(0).transpose(0, 1)
        self.register_buffer('pe', pe)

    def forward(self, x):
        x = x + self.pe[:x.size(0), :]
        return self.dropout(x)

 
d_model = 512  # Embedding dimension
nhead = 8  # Number of heads
num_layers = 6  # Number of transformer layers

 
model = TransformerEncoder(d_model, nhead, num_layers)
 
example_input = torch.randint(0, 999, (10, 32))  # 10 tokens per sequence, batch size of 32

# Forward pass
output = model(example_input)
print(output.shape)