Architectures / NLP/LLM

🔁 T5 Small

T5 encoder-decoder — bidirectional encoder + masked decoder with cross-attention (512D, 8 heads)

From Raffel et al. (2019). Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer. JMLR 2020. This page is the graph, not the PDF: open it, edit it, verify it, export it.

Every number on this page is computed from the graph by the same functions the app runs, not written by hand.

When to pick it

Structure

23 layers. Output shapes are propagated from the input shape, batch dimension excluded.

What the verifier says

The same 43 single-graph checks that run on every edit in the app, on this graph.

The PyTorch it exports

Generated from the graph above. First 46 lines; the app exports the whole file, plus the training loop, the data contract and a deploy bundle.

# Architecture designed with Neurarch: https://neurarch.com
# PyTorch: compatible with Python 3.8+ and torch>=1.12
# Colab: pip install torch torchvision  (usually pre-installed)

import torch
import torch.nn as nn
import torch.nn.functional as F
from typing import Tuple

class T5Small(nn.Module):
    def __init__(self):
        super().__init__()

        self.embedding_1 = nn.Embedding(32128, 512)
        self.rmsNorm_1 = nn.RMSNorm(512)
        self.multiHeadAttention_1 = nn.MultiheadAttention(embed_dim=512, num_heads=8, batch_first=True)
        self.rmsNorm_2 = nn.RMSNorm(512)
        self.feedForward_1 = nn.Sequential(
            nn.Linear(512, 2048),
            nn.ReLU(),
            nn.Linear(2048, 512)
        )
        self.layerNorm_1 = nn.LayerNorm(512)
        self.embedding_2 = nn.Embedding(32128, 512)
        self.rmsNorm_3 = nn.RMSNorm(512)
        self.causalAttention_1 = nn.MultiheadAttention(embed_dim=512, num_heads=8, batch_first=True)
        self.rmsNorm_4 = nn.RMSNorm(512)
        self.multiHeadAttention_2 = nn.MultiheadAttention(embed_dim=512, num_heads=8, batch_first=True)
        self.rmsNorm_5 = nn.RMSNorm(512)
        self.feedForward_2 = nn.Sequential(
            nn.Linear(512, 2048),
            nn.ReLU(),
            nn.Linear(2048, 512)
        )
        self.layerNorm_2 = nn.LayerNorm(512)
        self.linear_1 = nn.Linear(512, 32128)

    def forward(self, src, tgt=None):
        # encoder_ids shape: [1,512]
        # decoder_ids shape: [1,128]
        embedding_ng_enc = self.embedding_1(src)
        rms_norm_rm_enc = self.rmsNorm_1(embedding_ng_enc)
        multi_head_attention_ha_enc = self.multiHeadAttention_1(rms_norm_rm_enc, rms_norm_rm_enc, rms_norm_rm_enc)[0]
        add_dd_enc = multi_head_attention_ha_enc
        rms_norm_m_enc2 = self.rmsNorm_2(add_dd_enc)
        feed_forward_rd_enc = self.feedForward_1(rms_norm_m_enc2)

For agents

Also in NLP/LLM