Google

The encoding part consists of six identical parts - encoders. Identical but they do not share any weights. The number six is just what they used in the paper. Let’s consider each part in more detail. Also in the context of the machine translation before going through the encoders we need to pass words into the embedding layer. 


Transformer Encoding:

Here on the image you can see what each Transformer encoder part itself consists from three parts:

Self - Attention layer
Feed Forward layer
Add normalize 

Regarding the feed-forward layer, we already know how this works. The main mystery here is the self - Attention layer. In order to understand self -attention layer better I will divide self-layer understanding into several parts where I will modify the usual seq2seq model encoder:


Transformer Encoder part:

The standard procedure for processing an input sequence with a Transformer encoder begins by representing each input token, $$x_i$$, as its corresponding embedding, $$\mathbf{e}_i$$. This sequence of embeddings, $$\mathbf{e}_{0},...,\mathbf{e}_m$$, is then fed into the encoder. The encoder processes this input to produce a sequence of contextualized output vectors, or hidden states, $$\mathbf{h}_{0},...,\mathbf{h}_m$$.

Standard Transformer Encoding Procedure

In models like the Transformer, token embeddings such as $x_i$ are typically position-independent, meaning they do not inherently contain information about their location in a sequence. This makes the model's processing order-insensitive. To provide this crucial positional context, positional embeddings, $PE(i)$, are introduced. These are added to their corresponding token embeddings, allowing the model to distinguish between identical tokens at different positions.

Role of Positional Embeddings in Order-Insensitive Models

The architecture of a Transformer encoder is defined by several essential hyperparameters. These include the vocabulary size ($$|V|$$) and the embedding size ($$d_e$$) used for token representations. Additionally, the hidden size ($$d$$) specifies the input and output dimensionality for both the self-attention and the Feed-Forward Network (FFN) sub-layers. Other crucial hyperparameters are the number of attention heads ($$n_{\mathrm{head}}$$) for the multi-head self-attention mechanism, the internal hidden layer size of the FFN ($$d_{\textrm{ffn}}$$), and the model depth ($$L$$), which indicates the number of stacked layers.

Key Hyperparameters of a Transformer Encoder

This example illustrates the encoding process for a masked bilingual sentence pair. The input sequence, `[CLS] [MASK]是 [MASK]动物。 [SEP] Whales [MASK] [MASK] . [SEP]`, is first converted into a series of token embeddings, denoted as `e0` through `e11`. This embedding sequence is then processed by a Transformer Encoder, which outputs a corresponding sequence of contextualized hidden states, `h0` through `h11`. These hidden states serve as the basis for predicting the original masked tokens: '鲸鱼', '哺乳', 'are', and 'mammals'.

Transformer Encoding of a Masked Bilingual Sentence Pair

Prefix tuning is a parameter-efficient fine-tuning (PEFT) technique for large language models. Instead of fine-tuning all the model's parameters, it keeps the original model frozen and introduces a small number of trainable vectors, called a prefix. This prefix is prepended to the sequence of hidden states at each transformer layer, and only these prefix parameters are optimized during training for a specific task. This approach allows the model to be adapted to new tasks by learning a small, task-specific prefix that steers the behavior of the larger frozen model.

Prefix Tuning

In a sequence-to-sequence model, the input is processed by a stack of six encoder layers that have identical structures. A proposal is made to modify this architecture so that all six encoder layers share the exact same set of weights, with the goal of reducing the total number of model parameters. Which statement best analyzes the primary consequence of this change on the model's ability to process information?

A sentence is fed into the encoder side of a Transformer model. Arrange the following steps in the correct sequence to describe how the initial input is processed by the stack of encoders.

Given the following scenario, identify the most probable limitation of the current encoder design and propose a specific, justifiable change to the encoder stack to improve its ability to capture long-range dependencies.

Learn Before

Related