ImageGPT
Last updated
Last updated
The ImageGPT model was proposed in Generative Pretraining from Pixels by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever. ImageGPT (iGPT) is a GPT-2-like model trained to predict the next pixel value, allowing for both unconditional and conditional image generation.
The abstract from the paper is the following:
Inspired by progress in unsupervised representation learning for natural language, we examine whether similar models can learn useful representations for images. We train a sequence Transformer to auto-regressively predict pixels, without incorporating knowledge of the 2D input structure. Despite training on low-resolution ImageNet without labels, we find that a GPT-2 scale model learns strong image representations as measured by linear probing, fine-tuning, and low-data classification. On CIFAR-10, we achieve 96.3% accuracy with a linear probe, outperforming a supervised Wide ResNet, and 99.0% accuracy with full fine-tuning, matching the top supervised pre-trained models. We are also competitive with self-supervised benchmarks on ImageNet when substituting pixels for a VQVAE encoding, achieving 69.0% top-1 accuracy on a linear probe of our features.
Summary of the approach. Taken from the [original paper](https://cdn.openai.com/papers/Generative_Pretraining_from_Pixels_V2.pdf).
This model was contributed by nielsr, based on this issue. The original code can be found here.
Tips:
ImageGPT is almost exactly the same as GPT-2, with the exception that a different activation function is used (namely “quick gelu”), and the layer normalization layers don’t mean center the inputs. ImageGPT also doesn’t have tied input- and output embeddings.
As the time- and memory requirements of the attention mechanism of Transformers scales quadratically in the sequence length, the authors pre-trained ImageGPT on smaller input resolutions, such as 32x32 and 64x64. However, feeding a sequence of 32x32x3=3072 tokens from 0..255 into a Transformer is still prohibitively large. Therefore, the authors applied k-means clustering to the (R,G,B) pixel values with k=512. This way, we only have a 32*32 = 1024-long sequence, but now of integers in the range 0..511. So we are shrinking the sequence length at the cost of a bigger embedding matrix. In other words, the vocabulary size of ImageGPT is 512, + 1 for a special “start of sentence” (SOS) token, used at the beginning of every sequence. One can use ImageGPTImageProcessor to prepare images for the model.
Despite being pre-trained entirely unsupervised (i.e. without the use of any labels), ImageGPT produces fairly performant image features useful for downstream tasks, such as image classification. The authors showed that the features in the middle of the network are the most performant, and can be used as-is to train a linear model (such as a sklearn logistic regression model for example). This is also referred to as “linear probing”. Features can be easily obtained by first forwarding the image through the model, then specifying output_hidden_states=True, and then average-pool the hidden states at whatever layer you like.
Alternatively, one can further fine-tune the entire model on a downstream dataset, similar to BERT. For this, you can use ImageGPTForImageClassification.
ImageGPT comes in different sizes: there’s ImageGPT-small, ImageGPT-medium and ImageGPT-large. The authors did also train an XL variant, which they didn’t release. The differences in size are summarized in the following table:
Model variant | Depths | Hidden sizes | Decoder hidden size | Params (M) | ImageNet-1k Top 1 |
MiT-b0 | [2, 2, 2, 2] | [32, 64, 160, 256] | 256 | 3.7 | 70.5 |
MiT-b1 | [2, 2, 2, 2] | [64, 128, 320, 512] | 256 | 14.0 | 78.7 |
MiT-b2 | [3, 4, 6, 3] | [64, 128, 320, 512] | 768 | 25.4 | 81.6 |
MiT-b3 | [3, 4, 18, 3] | [64, 128, 320, 512] | 768 | 45.2 | 83.1 |
MiT-b4 | [3, 8, 27, 3] | [64, 128, 320, 512] | 768 | 62.6 | 83.6 |
MiT-b5 | [3, 6, 40, 3] | [64, 128, 320, 512] | 768 | 82.0 | 83.8 |
A list of official BOINC AI and community (indicated by 🌎) resources to help you get started with ImageGPT.
Image Classification
Demo notebooks for ImageGPT can be found here.
ImageGPTForImageClassification is supported by this example script and notebook.
See also: Image classification task guide
If you’re interested in submitting a resource to be included here, please feel free to open a Pull Request and we’ll review it! The resource should ideally demonstrate something new instead of duplicating an existing resource.
( vocab_size = 513n_positions = 1024n_embd = 512n_layer = 24n_head = 8n_inner = Noneactivation_function = 'quick_gelu'resid_pdrop = 0.1embd_pdrop = 0.1attn_pdrop = 0.1layer_norm_epsilon = 1e-05initializer_range = 0.02scale_attn_weights = Trueuse_cache = Truetie_word_embeddings = Falsescale_attn_by_inverse_layer_idx = Falsereorder_and_upcast_attn = False**kwargs )
Parameters
vocab_size (int, optional, defaults to 512) — Vocabulary size of the GPT-2 model. Defines the number of different tokens that can be represented by the inputs_ids passed when calling ImageGPTModel or TFImageGPTModel.
n_positions (int, optional, defaults to 32*32) — The maximum sequence length that this model might ever be used with. Typically set this to something large just in case (e.g., 512 or 1024 or 2048).
n_embd (int, optional, defaults to 512) — Dimensionality of the embeddings and hidden states.
n_layer (int, optional, defaults to 24) — Number of hidden layers in the Transformer encoder.
n_head (int, optional, defaults to 8) — Number of attention heads for each attention layer in the Transformer encoder.
n_inner (int, optional, defaults to None) — Dimensionality of the inner feed-forward layers. None will set it to 4 times n_embd
activation_function (str, optional, defaults to "quick_gelu") — Activation function (can be one of the activation functions defined in src/transformers/activations.py). Defaults to “quick_gelu”.
resid_pdrop (float, optional, defaults to 0.1) — The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.
embd_pdrop (int, optional, defaults to 0.1) — The dropout ratio for the embeddings.
attn_pdrop (float, optional, defaults to 0.1) — The dropout ratio for the attention.
layer_norm_epsilon (float, optional, defaults to 1e-5) — The epsilon to use in the layer normalization layers.
initializer_range (float, optional, defaults to 0.02) — The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
scale_attn_weights (bool, optional, defaults to True) — Scale attention weights by dividing by sqrt(hidden_size)..
use_cache (bool, optional, defaults to True) — Whether or not the model should return the last key/values attentions (not used by all models).
scale_attn_by_inverse_layer_idx (bool, optional, defaults to False) — Whether to additionally scale attention weights by 1 / layer_idx + 1.
reorder_and_upcast_attn (bool, optional, defaults to False) — Whether to scale keys (K) prior to computing attention (dot-product) and upcast attention dot-product/softmax to float() when training with mixed precision.
This is the configuration class to store the configuration of a ImageGPTModel or a TFImageGPTModel. It is used to instantiate a GPT-2 model according to the specified arguments, defining the model architecture. Instantiating a configuration with the defaults will yield a similar configuration to that of the ImageGPT openai/imagegpt-small architecture.
Configuration objects inherit from PretrainedConfig and can be used to control the model outputs. Read the documentation from PretrainedConfig for more information.
Example:
Copied
( *args**kwargs )
__call__
( images**kwargs )
Preprocess an image or a batch of images.
( clusters: typing.Union[typing.List[typing.List[int]], numpy.ndarray, NoneType] = Nonedo_resize: bool = Truesize: typing.Dict[str, int] = Noneresample: Resampling = <Resampling.BILINEAR: 2>do_normalize: bool = Truedo_color_quantize: bool = True**kwargs )
Parameters
clusters (np.ndarray or List[List[int]], optional) — The color clusters to use, of shape (n_clusters, 3) when color quantizing. Can be overriden by clusters in preprocess.
do_resize (bool, optional, defaults to True) — Whether to resize the image’s dimensions to (size["height"], size["width"]). Can be overridden by do_resize in preprocess.
size (Dict[str, int] optional, defaults to {"height" -- 256, "width": 256}): Size of the image after resizing. Can be overridden by size in preprocess.
resample (PILImageResampling, optional, defaults to PILImageResampling.BICUBIC) — Resampling filter to use if resizing the image. Can be overridden by resample in preprocess.
do_normalize (bool, optional, defaults to True) — Whether to normalize the image pixel value to between [-1, 1]. Can be overridden by do_normalize in preprocess.
do_color_quantize (bool, optional, defaults to True) — Whether to color quantize the image. Can be overridden by do_color_quantize in preprocess.
Constructs a ImageGPT image processor. This image processor can be used to resize images to a smaller resolution (such as 32x32 or 64x64), normalize them and finally color quantize them to obtain sequences of “pixel values” (color clusters).
preprocess
( images: typing.Union[ForwardRef('PIL.Image.Image'), numpy.ndarray, ForwardRef('torch.Tensor'), typing.List[ForwardRef('PIL.Image.Image')], typing.List[numpy.ndarray], typing.List[ForwardRef('torch.Tensor')]]do_resize: bool = Nonesize: typing.Dict[str, int] = Noneresample: Resampling = Nonedo_normalize: bool = Nonedo_color_quantize: typing.Optional[bool] = Noneclusters: typing.Union[typing.List[typing.List[int]], numpy.ndarray, NoneType] = Nonereturn_tensors: typing.Union[str, transformers.utils.generic.TensorType, NoneType] = Nonedata_format: typing.Union[str, transformers.image_utils.ChannelDimension, NoneType] = <ChannelDimension.FIRST: 'channels_first'>input_data_format: typing.Union[str, transformers.image_utils.ChannelDimension, NoneType] = None**kwargs )
Parameters
images (ImageInput) — Image to preprocess. Expects a single or batch of images with pixel values ranging from 0 to 255. If passing in images with pixel values between 0 and 1, set do_normalize=False.
do_resize (bool, optional, defaults to self.do_resize) — Whether to resize the image.
size (Dict[str, int], optional, defaults to self.size) — Size of the image after resizing.
resample (int, optional, defaults to self.resample) — Resampling filter to use if resizing the image. This can be one of the enum PILImageResampling, Only has an effect if do_resize is set to True.
do_normalize (bool, optional, defaults to self.do_normalize) — Whether to normalize the image
do_color_quantize (bool, optional, defaults to self.do_color_quantize) — Whether to color quantize the image.
clusters (np.ndarray or List[List[int]], optional, defaults to self.clusters) — Clusters used to quantize the image of shape (n_clusters, 3). Only has an effect if do_color_quantize is set to True.
return_tensors (str or TensorType, optional) — The type of tensors to return. Can be one of:
Unset: Return a list of np.ndarray.
TensorType.TENSORFLOW or 'tf': Return a batch of type tf.Tensor.
TensorType.PYTORCH or 'pt': Return a batch of type torch.Tensor.
TensorType.NUMPY or 'np': Return a batch of type np.ndarray.
TensorType.JAX or 'jax': Return a batch of type jax.numpy.ndarray.
data_format (ChannelDimension or str, optional, defaults to ChannelDimension.FIRST) — The channel dimension format for the output image. Can be one of:
ChannelDimension.FIRST: image in (num_channels, height, width) format.
ChannelDimension.LAST: image in (height, width, num_channels) format. Only has an effect if do_color_quantize is set to False.
input_data_format (ChannelDimension or str, optional) — The channel dimension format for the input image. If unset, the channel dimension format is inferred from the input image. Can be one of:
"channels_first" or ChannelDimension.FIRST: image in (num_channels, height, width) format.
"channels_last" or ChannelDimension.LAST: image in (height, width, num_channels) format.
"none" or ChannelDimension.NONE: image in (height, width) format.
Preprocess an image or batch of images.
( config: ImageGPTConfig )
Parameters
config (ImageGPTConfig) — Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the from_pretrained() method to load the model weights.
The bare ImageGPT Model transformer outputting raw hidden-states without any specific head on top.
This model inherits from PreTrainedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)
This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
forward
( input_ids: typing.Optional[torch.Tensor] = Nonepast_key_values: typing.Optional[typing.Tuple[typing.Tuple[torch.Tensor]]] = Noneattention_mask: typing.Optional[torch.Tensor] = Nonetoken_type_ids: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.Tensor] = Nonehead_mask: typing.Optional[torch.Tensor] = Noneinputs_embeds: typing.Optional[torch.Tensor] = Noneencoder_hidden_states: typing.Optional[torch.Tensor] = Noneencoder_attention_mask: typing.Optional[torch.Tensor] = Noneuse_cache: typing.Optional[bool] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = None**kwargs: typing.Any ) → transformers.modeling_outputs.BaseModelOutputWithPastAndCrossAttentions or tuple(torch.FloatTensor)
Parameters
input_ids (torch.LongTensor of shape (batch_size, sequence_length)) — input_ids_length = sequence_length if past_key_values is None else past_key_values[0][0].shape[-2] (sequence_length of input past key value states). Indices of input sequence tokens in the vocabulary.
If past_key_values is used, only input_ids that do not have their past calculated should be passed as input_ids.
Indices can be obtained using AutoImageProcessor. See ImageGPTImageProcessor.call() for details.
past_key_values (Tuple[Tuple[torch.Tensor]] of length config.n_layers) — Contains precomputed hidden-states (key and values in the attention blocks) as computed by the model (see past_key_values output below). Can be used to speed up sequential decoding. The input_ids which have their past given to this model should not be passed as input_ids as they have already been computed.
attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional) — Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
1 for tokens that are not masked,
0 for tokens that are masked.
token_type_ids (torch.LongTensor of shape (batch_size, sequence_length), optional) — Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:
0 corresponds to a sentence A token,
1 corresponds to a sentence B token.
position_ids (torch.LongTensor of shape (batch_size, sequence_length), optional) — Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].
head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional) — Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:
1 indicates the head is not masked,
0 indicates the head is masked.
inputs_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) — Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.
If past_key_values is used, optionally only the last inputs_embeds have to be input (see past_key_values).
use_cache (bool, optional) — If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values).
output_attentions (bool, optional) — Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.
output_hidden_states (bool, optional) — Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.
return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.
labels (torch.LongTensor of shape (batch_size, sequence_length), optional) — Labels for language modeling. Note that the labels are shifted inside the model, i.e. you can set labels = input_ids Indices are selected in [-100, 0, ..., config.vocab_size] All labels set to -100 are ignored (masked), the loss is only computed for labels in [0, ..., config.vocab_size]
Returns
transformers.modeling_outputs.BaseModelOutputWithPastAndCrossAttentions or tuple(torch.FloatTensor)
A transformers.modeling_outputs.BaseModelOutputWithPastAndCrossAttentions or a tuple of torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (ImageGPTConfig) and inputs.
last_hidden_state (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size)) — Sequence of hidden-states at the output of the last layer of the model.
If past_key_values is used only the last hidden-state of the sequences of shape (batch_size, 1, hidden_size) is output.
past_key_values (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) — Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)) and optionally if config.is_encoder_decoder=True 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).
Contains pre-computed hidden-states (key and values in the self-attention blocks and optionally if config.is_encoder_decoder=True in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.
hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.FloatTensor (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape (batch_size, sequence_length, hidden_size).
Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
attentions (tuple(torch.FloatTensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).
Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
cross_attentions (tuple(torch.FloatTensor), optional, returned when output_attentions=True and config.add_cross_attention=True is passed or when config.output_attentions=True) — Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).
Attentions weights of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.
The ImageGPTModel forward method, overrides the __call__ special method.
Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.
Examples:
Copied
( config: ImageGPTConfig )
Parameters
config (ImageGPTConfig) — Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the from_pretrained() method to load the model weights.
The ImageGPT Model transformer with a language modeling head on top (linear layer with weights tied to the input embeddings).
This model inherits from PreTrainedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)
This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
forward
( input_ids: typing.Optional[torch.Tensor] = Nonepast_key_values: typing.Optional[typing.Tuple[typing.Tuple[torch.Tensor]]] = Noneattention_mask: typing.Optional[torch.Tensor] = Nonetoken_type_ids: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.Tensor] = Nonehead_mask: typing.Optional[torch.Tensor] = Noneinputs_embeds: typing.Optional[torch.Tensor] = Noneencoder_hidden_states: typing.Optional[torch.Tensor] = Noneencoder_attention_mask: typing.Optional[torch.Tensor] = Nonelabels: typing.Optional[torch.Tensor] = Noneuse_cache: typing.Optional[bool] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = None**kwargs: typing.Any ) → transformers.modeling_outputs.CausalLMOutputWithCrossAttentions or tuple(torch.FloatTensor)
Parameters
input_ids (torch.LongTensor of shape (batch_size, sequence_length)) — input_ids_length = sequence_length if past_key_values is None else past_key_values[0][0].shape[-2] (sequence_length of input past key value states). Indices of input sequence tokens in the vocabulary.
If past_key_values is used, only input_ids that do not have their past calculated should be passed as input_ids.
Indices can be obtained using AutoImageProcessor. See ImageGPTImageProcessor.call() for details.
past_key_values (Tuple[Tuple[torch.Tensor]] of length config.n_layers) — Contains precomputed hidden-states (key and values in the attention blocks) as computed by the model (see past_key_values output below). Can be used to speed up sequential decoding. The input_ids which have their past given to this model should not be passed as input_ids as they have already been computed.
attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional) — Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
1 for tokens that are not masked,
0 for tokens that are masked.
token_type_ids (torch.LongTensor of shape (batch_size, sequence_length), optional) — Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:
0 corresponds to a sentence A token,
1 corresponds to a sentence B token.
position_ids (torch.LongTensor of shape (batch_size, sequence_length), optional) — Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].
head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional) — Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:
1 indicates the head is not masked,
0 indicates the head is masked.
inputs_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) — Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.
If past_key_values is used, optionally only the last inputs_embeds have to be input (see past_key_values).
use_cache (bool, optional) — If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values).
output_attentions (bool, optional) — Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.
output_hidden_states (bool, optional) — Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.
return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.
labels (torch.LongTensor of shape (batch_size, sequence_length), optional) — Labels for language modeling. Note that the labels are shifted inside the model, i.e. you can set labels = input_ids Indices are selected in [-100, 0, ..., config.vocab_size] All labels set to -100 are ignored (masked), the loss is only computed for labels in [0, ..., config.vocab_size]
Returns
transformers.modeling_outputs.CausalLMOutputWithCrossAttentions or tuple(torch.FloatTensor)
A transformers.modeling_outputs.CausalLMOutputWithCrossAttentions or a tuple of torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (ImageGPTConfig) and inputs.
loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) — Language modeling loss (for next-token prediction).
logits (torch.FloatTensor of shape (batch_size, sequence_length, config.vocab_size)) — Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.FloatTensor (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape (batch_size, sequence_length, hidden_size).
Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
attentions (tuple(torch.FloatTensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).
Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
cross_attentions (tuple(torch.FloatTensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).
Cross attentions weights after the attention softmax, used to compute the weighted average in the cross-attention heads.
past_key_values (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) — Tuple of torch.FloatTensor tuples of length config.n_layers, with each tuple containing the cached key, value states of the self-attention and the cross-attention layers if model is used in encoder-decoder setting. Only relevant if config.is_decoder = True.
Contains pre-computed hidden-states (key and values in the attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.
The ImageGPTForCausalImageModeling forward method, overrides the __call__ special method.
Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.
Examples:
Copied
( config: ImageGPTConfig )
Parameters
config (ImageGPTConfig) — Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out the from_pretrained() method to load the model weights.
The ImageGPT Model transformer with an image classification head on top (linear layer). ImageGPTForImageClassification average-pools the hidden states in order to do the classification.
This model inherits from PreTrainedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)
This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
forward
( input_ids: typing.Optional[torch.Tensor] = Nonepast_key_values: typing.Optional[typing.Tuple[typing.Tuple[torch.Tensor]]] = Noneattention_mask: typing.Optional[torch.Tensor] = Nonetoken_type_ids: typing.Optional[torch.Tensor] = Noneposition_ids: typing.Optional[torch.Tensor] = Nonehead_mask: typing.Optional[torch.Tensor] = Noneinputs_embeds: typing.Optional[torch.Tensor] = Nonelabels: typing.Optional[torch.Tensor] = Noneuse_cache: typing.Optional[bool] = Noneoutput_attentions: typing.Optional[bool] = Noneoutput_hidden_states: typing.Optional[bool] = Nonereturn_dict: typing.Optional[bool] = None**kwargs: typing.Any ) → transformers.modeling_outputs.SequenceClassifierOutputWithPast or tuple(torch.FloatTensor)
Parameters
input_ids (torch.LongTensor of shape (batch_size, sequence_length)) — input_ids_length = sequence_length if past_key_values is None else past_key_values[0][0].shape[-2] (sequence_length of input past key value states). Indices of input sequence tokens in the vocabulary.
If past_key_values is used, only input_ids that do not have their past calculated should be passed as input_ids.
Indices can be obtained using AutoImageProcessor. See ImageGPTImageProcessor.call() for details.
past_key_values (Tuple[Tuple[torch.Tensor]] of length config.n_layers) — Contains precomputed hidden-states (key and values in the attention blocks) as computed by the model (see past_key_values output below). Can be used to speed up sequential decoding. The input_ids which have their past given to this model should not be passed as input_ids as they have already been computed.
attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional) — Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:
1 for tokens that are not masked,
0 for tokens that are masked.
token_type_ids (torch.LongTensor of shape (batch_size, sequence_length), optional) — Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:
0 corresponds to a sentence A token,
1 corresponds to a sentence B token.
position_ids (torch.LongTensor of shape (batch_size, sequence_length), optional) — Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].
head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional) — Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:
1 indicates the head is not masked,
0 indicates the head is masked.
inputs_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) — Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.
If past_key_values is used, optionally only the last inputs_embeds have to be input (see past_key_values).
use_cache (bool, optional) — If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values).
output_attentions (bool, optional) — Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.
output_hidden_states (bool, optional) — Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.
return_dict (bool, optional) — Whether or not to return a ModelOutput instead of a plain tuple.
labels (torch.LongTensor of shape (batch_size,), optional) — Labels for computing the sequence classification/regression loss. Indices should be in [0, ..., config.num_labels - 1]. If config.num_labels == 1 a regression loss is computed (Mean-Square loss), If config.num_labels > 1 a classification loss is computed (Cross-Entropy).
Returns
transformers.modeling_outputs.SequenceClassifierOutputWithPast or tuple(torch.FloatTensor)
A transformers.modeling_outputs.SequenceClassifierOutputWithPast or a tuple of torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (ImageGPTConfig) and inputs.
loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided) — Classification (or regression if config.num_labels==1) loss.
logits (torch.FloatTensor of shape (batch_size, config.num_labels)) — Classification (or regression if config.num_labels==1) scores (before SoftMax).
past_key_values (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) — Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head))
Contains pre-computed hidden-states (key and values in the self-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.
hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) — Tuple of torch.FloatTensor (one for the output of the embeddings, if the model has an embedding layer, + one for the output of each layer) of shape (batch_size, sequence_length, hidden_size).
Hidden-states of the model at the output of each layer plus the optional initial embedding outputs.
attentions (tuple(torch.FloatTensor), optional, returned when output_attentions=True is passed or when config.output_attentions=True) — Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).
Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
The ImageGPTForImageClassification forward method, overrides the __call__ special method.
Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.
Examples:
Copied
