class

GPTForSequenceClassification

extendsPretrainedModel

GPTForSequenceClassification(config: GPTConfig)

source edit

GPT-1 with a last-token sequence-classification head.

Wraps GPTModel with a dropout-regularised linear classifier operating on the rightmost non-pad token's hidden state — the canonical decoder-style pooling. Unlike BERT (CLS at position 0), decoder LMs only have a meaningful summary at the final attended position because every preceding position has attended to a strict prefix only. When attention_mask is provided we honour it; otherwise the last position of every row is used.

Parameters

configGPTConfig

Hyperparameters. config.num_labels sets the output dimension; config.classifier_dropout (falling back to hidden_dropout) sets the dropout applied before the linear.

Attributes

transformerGPTModel

Underlying decoder trunk.

dropoutnn.Dropout

Dropout layer applied to the pooled last-token embedding.

classifiernn.Linear

Final linear of shape (hidden_size, num_labels) producing per-class logits.

Notes

Reference: Radford et al., "Improving Language Understanding by Generative Pre-Training", OpenAI Technical Report, 2018, §3.3.

Pooling picks index $t^\star = \sum_t \mathbf{1}[\mathrm{mask}_t] - 1$ (the rightmost real token) per row.

Examples

>>> import lucid
>>> from lucid.models.text.gpt import GPTConfig, GPTForSequenceClassification
>>> cfg = GPTConfig(num_labels=3, num_hidden_layers=2, hidden_size=128,
...                 num_attention_heads=2, intermediate_size=512)
>>> model = GPTForSequenceClassification(cfg).eval()
>>> input_ids = lucid.tensor([[101, 7592, 102]])
>>> out = model(input_ids)
>>> out.logits.shape   # (B=1, num_labels=3)
(1, 3)

Used by 2

Constructors

dunder

init

→None

__init__(config: GPTConfig)

source edit

Instance methods

forward

→MaskedLMOutput

forward(input_ids: Tensor, attention_mask: Tensor | None = None, labels: Tensor | None = None)

source edit

>>> import lucid >>> from lucid.models.text.gpt import GPTConfig, GPTForSequenceClassification >>> cfg = GPTConfig(num_labels=3, num_hidden_layers=2, hidden_size=128, ... num_attention_heads=2, intermediate_size=512) >>> model = GPTForSequenceClassification(cfg).eval() >>> input_ids = lucid.tensor([[101, 7592, 102]]) >>> out = model(input_ids) >>> out.logits.shape # (B=1, num_labels=3) (1, 3)