AWS Trainium & Inferentia documentation

Models

AWS Trainium & Inferentia

Join the Hugging Face community

and get access to the augmented documentation experience

Collaborate on models, datasets and Spaces

Faster examples with accelerated inference

Switch between documentation themes

to get started

Models

Generic model classes

NeuronTracedModel

The NeuronTracedModel class is available for instantiating a base Neuron model without a specific head. It is used as the base class for all tasks but text generation.

class optimum.neuron.NeuronTracedModel

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Base class running compiled and optimized models on Neuron devices.

It implements generic methods for interacting with the Hugging Face Hub as well as compiling vanilla transformers models to neuron-optimized TorchScript module and export it using optimum.exporters.neuron toolchain.

Class attributes:

model_type (str, optional, defaults to "neuron_model") — The name of the model type to use when registering the NeuronTracedModel classes.
auto_model_class (Type, optional, defaults to AutoModel) — The AutoModel class to be represented by the current NeuronTracedModel class.

Common attributes:

model (torch.jit._script.ScriptModule) — The loaded ScriptModule compiled for neuron devices.
config (PretrainedConfig) — The configuration of the model.
model_save_dir (Path) — The directory where a neuron compiled model is saved. By default, if the loaded model is local, the directory where the original model will be used. Otherwise, the cache directory will be used.

can_generate

< source >

( )

Returns whether this model can generate sequences with .generate().

get_input_static_shapes

< source >

( neuron_config: NeuronDefaultConfig )

Gets a dictionary of inputs with their valid static shapes.

load_model

< source >

( path: Union to_neuron: bool = False device_id: int = 0 )

Parameters

path (Union[str, Path]) — Path of the compiled model.
to_neuron (bool, defaults to False) — Whether to move manually the traced model to NeuronCore. It’s only needed when inline_weights_to_neff=False, otherwise it is loaded automatically to a Neuron device.
device_id (int, defaults to 0) — Index of NeuronCore to load the traced model to.

Loads a TorchScript module compiled by neuron(x)-cc compiler. It will be first loaded onto CPU and then moved to one or multiple NeuronCore.

remove_padding

< source >

( outputs: List dims: List indices: List padding_side: Literal = 'right' )

Parameters

outputs (List[torch.Tensor]) — List of torch tensors which are inference output.
dims (List[int]) — List of dimensions in which we slice a tensor.
indices (List[int]) — List of indices in which we slice a tensor along an axis.
padding_side (Literal["right", "left"], defaults to “right”) — The side on which the padding has been applied.

Removes padding from output tensors.

NeuronDecoderModel

The NeuronDecoderModel class is the base class for text generation models.

class optimum.neuron.NeuronDecoderModel

< source >

( config: PretrainedConfig checkpoint_dir: Union compiled_dir: Union = None generation_config: Optional = None )

Base class to convert and run pre-trained transformers decoder models on Neuron devices.

It implements the methods to convert a pre-trained transformers decoder model into a Neuron transformer model by:

transferring the checkpoint weights of the original into an optimized neuron graph,
compiling the resulting graph using the Neuron compiler.

Common attributes:

model (torch.nn.Module) — The decoder model with a graph optimized for neuron devices.
config (PretrainedConfig) — The configuration of the original model.
generation_config (GenerationConfig) — The generation configuration used by default when calling generate().

Natural Language Processing

The following Neuron model classes are available for natural language processing tasks.

NeuronModelForFeatureExtraction

class optimum.neuron.NeuronModelForFeatureExtraction

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with a BaseModelOutput for feature-extraction tasks.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Feature Extraction model on Neuron devices.

forward

< source >

( input_ids: Tensor attention_mask: Tensor token_type_ids: Optional = None **kwargs )

Parameters

input_ids (torch.Tensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode and PreTrainedTokenizer.__call__ for details. What are input IDs?
attention_mask (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — 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. What are attention masks?
token_type_ids (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:
- 1 for tokens that are sentence A,
- 0 for tokens that are sentence B. What are token type IDs?

The NeuronModelForFeatureExtraction 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.

Example of feature extraction: (Following model is compiled with neuronx compiler and can only be run on INF2. Replace “neuronx” with “neuron” if you are using INF1.)

>>> from transformers import AutoTokenizer
>>> from optimum.neuron import NeuronModelForFeatureExtraction

>>> tokenizer = AutoTokenizer.from_pretrained("optimum/all-MiniLM-L6-v2-neuronx")
>>> model = NeuronModelForFeatureExtraction.from_pretrained("optimum/all-MiniLM-L6-v2-neuronx")

>>> inputs = tokenizer("Dear Evan Hansen is the winner of six Tony Awards.", return_tensors="pt")

>>> outputs = model(**inputs)
>>> last_hidden_state = outputs.last_hidden_state
>>> list(last_hidden_state.shape)
[1, 13, 384]

NeuronModelForSentenceTransformers

class optimum.neuron.NeuronModelForSentenceTransformers

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model for Sentence Transformers.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Sentence Transformers model on Neuron devices.

forward

< source >

( input_ids: Tensor attention_mask: Tensor pixel_values: Optional = None token_type_ids: Optional = None **kwargs )

Parameters

input_ids (torch.Tensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode and PreTrainedTokenizer.__call__ for details. What are input IDs?
attention_mask (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — 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. What are attention masks?
token_type_ids (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:
- 1 for tokens that are sentence A,
- 0 for tokens that are sentence B. What are token type IDs?

The NeuronModelForSentenceTransformers forward method, overrides the __call__ special method.

Example of TEXT Sentence Transformers:

>>> from transformers import AutoTokenizer
>>> from optimum.neuron import NeuronModelForSentenceTransformers

>>> tokenizer = AutoTokenizer.from_pretrained("optimum/bge-base-en-v1.5-neuronx")
>>> model = NeuronModelForSentenceTransformers.from_pretrained("optimum/bge-base-en-v1.5-neuronx")

>>> inputs = tokenizer("In the smouldering promise of the fall of Troy, a mythical world of gods and mortals rises from the ashes.", return_tensors="pt")

>>> outputs = model(**inputs)
>>> token_embeddings = outputs.token_embeddings
>>> sentence_embedding = = outputs.sentence_embedding

NeuronModelForMaskedLM

class optimum.neuron.NeuronModelForMaskedLM

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with a MaskedLMOutput for masked language modeling tasks.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Masked language model for on Neuron devices.

forward

< source >

( input_ids: Tensor attention_mask: Tensor token_type_ids: Optional = None **kwargs )

Parameters

input_ids (torch.Tensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode and PreTrainedTokenizer.__call__ for details. What are input IDs?
attention_mask (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — 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. What are attention masks?
token_type_ids (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:
- 1 for tokens that are sentence A,
- 0 for tokens that are sentence B. What are token type IDs?

The NeuronModelForMaskedLM forward method, overrides the __call__ special method.

Example of fill mask: (Following model is compiled with neuronx compiler and can only be run on INF2. Replace “neuronx” with “neuron” if you are using INF1.)

>>> from transformers import AutoTokenizer
>>> from optimum.neuron import NeuronModelForMaskedLM
>>> import torch

>>> tokenizer = AutoTokenizer.from_pretrained("optimum/legal-bert-base-uncased-neuronx")
>>> model = NeuronModelForMaskedLM.from_pretrained("optimum/legal-bert-base-uncased-neuronx")

>>> inputs = tokenizer("This [MASK] Agreement is between General Motors and John Murray.", return_tensors="pt")

>>> outputs = model(**inputs)
>>> logits = outputs.logits
>>> list(logits.shape)
[1, 13, 30522]

NeuronModelForSequenceClassification

class optimum.neuron.NeuronModelForSequenceClassification

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with a sequence classification/regression head on top (a linear layer on top of the pooled output) e.g. for GLUE tasks.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Sequence Classification model on Neuron devices.

forward

< source >

( input_ids: Tensor attention_mask: Tensor token_type_ids: Optional = None **kwargs )

Parameters

input_ids (torch.Tensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode and PreTrainedTokenizer.__call__ for details. What are input IDs?
attention_mask (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — 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. What are attention masks?
token_type_ids (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:
- 1 for tokens that are sentence A,
- 0 for tokens that are sentence B. What are token type IDs?

The NeuronModelForSequenceClassification forward method, overrides the __call__ special method.

Example of single-label classification: (Following model is compiled with neuronx compiler and can only be run on INF2.)

>>> from transformers import AutoTokenizer
>>> from optimum.neuron import NeuronModelForSequenceClassification

>>> tokenizer = AutoTokenizer.from_pretrained("optimum/distilbert-base-uncased-finetuned-sst-2-english-neuronx")
>>> model = NeuronModelForSequenceClassification.from_pretrained("optimum/distilbert-base-uncased-finetuned-sst-2-english-neuronx")

>>> inputs = tokenizer("Hamilton is considered to be the best musical of human history.", return_tensors="pt")

>>> outputs = model(**inputs)
>>> logits = outputs.logits
>>> list(logits.shape)
[1, 2]

NeuronModelForQuestionAnswering

class optimum.neuron.NeuronModelForQuestionAnswering

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with a QuestionAnsweringModelOutput for extractive question-answering tasks like SQuAD.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Question Answering model on Neuron devices.

forward

< source >

( input_ids: Tensor attention_mask: Tensor token_type_ids: Optional = None **kwargs )

Parameters

input_ids (torch.Tensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode and PreTrainedTokenizer.__call__ for details. What are input IDs?
attention_mask (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — 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. What are attention masks?
token_type_ids (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:
- 1 for tokens that are sentence A,
- 0 for tokens that are sentence B. What are token type IDs?

The NeuronModelForQuestionAnswering forward method, overrides the __call__ special method.

Example of question answering: (Following model is compiled with neuronx compiler and can only be run on INF2.)

>>> import torch
>>> from transformers import AutoTokenizer
>>> from optimum.neuron import NeuronModelForQuestionAnswering

>>> tokenizer = AutoTokenizer.from_pretrained("optimum/roberta-base-squad2-neuronx")
>>> model = NeuronModelForQuestionAnswering.from_pretrained("optimum/roberta-base-squad2-neuronx")

>>> question, text = "Are there wheelchair spaces in the theatres?", "Yes, we have reserved wheelchair spaces with a good view."
>>> inputs = tokenizer(question, text, return_tensors="pt")
>>> start_positions = torch.tensor([1])
>>> end_positions = torch.tensor([12])

>>> outputs = model(**inputs, start_positions=start_positions, end_positions=end_positions)
>>> start_scores = outputs.start_logits
>>> end_scores = outputs.end_logits

NeuronModelForTokenClassification

class optimum.neuron.NeuronModelForTokenClassification

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with a token classification head on top (a linear layer on top of the hidden-states output) e.g. for Named-Entity-Recognition (NER) tasks.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Token Classification model on Neuron devices.

forward

< source >

( input_ids: Tensor attention_mask: Tensor token_type_ids: Optional = None **kwargs )

Parameters

input_ids (torch.Tensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode and PreTrainedTokenizer.__call__ for details. What are input IDs?
attention_mask (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — 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. What are attention masks?
token_type_ids (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:
- 1 for tokens that are sentence A,
- 0 for tokens that are sentence B. What are token type IDs?

The NeuronModelForTokenClassification forward method, overrides the __call__ special method.

Example of token classification: (Following model is compiled with neuronx compiler and can only be run on INF2.)

>>> from transformers import AutoTokenizer
>>> from optimum.neuron import NeuronModelForTokenClassification

>>> tokenizer = AutoTokenizer.from_pretrained("optimum/bert-base-NER-neuronx")
>>> model = NeuronModelForTokenClassification.from_pretrained("optimum/bert-base-NER-neuronx")

>>> inputs = tokenizer("Lin-Manuel Miranda is an American songwriter, actor, singer, filmmaker, and playwright.", return_tensors="pt")

>>> outputs = model(**inputs)
>>> logits = outputs.logits
>>> list(logits.shape)
[1, 20, 9]

NeuronModelForMultipleChoice

class optimum.neuron.NeuronModelForMultipleChoice

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with a multiple choice classification head on top (a linear layer on top of the pooled output and a softmax) e.g. for RocStories/SWAG tasks.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Multiple choice model on Neuron devices.

forward

< source >

( input_ids: Tensor attention_mask: Tensor token_type_ids: Optional = None **kwargs )

Parameters

input_ids (torch.Tensor of shape (batch_size, num_choices, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode and PreTrainedTokenizer.__call__ for details. What are input IDs?
attention_mask (Union[torch.Tensor, None] of shape (batch_size, num_choices, sequence_length), defaults to None) — 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. What are attention masks?
token_type_ids (Union[torch.Tensor, None] of shape (batch_size, num_choices, sequence_length), defaults to None) — Segment token indices to indicate first and second portions of the inputs. Indices are selected in [0, 1]:
- 1 for tokens that are sentence A,
- 0 for tokens that are sentence B. What are token type IDs?

The NeuronModelForMultipleChoice forward method, overrides the __call__ special method.

Example of mutliple choice: (Following model is compiled with neuronx compiler and can only be run on INF2.)

>>> from transformers import AutoTokenizer
>>> from optimum.neuron import NeuronModelForMultipleChoice

>>> tokenizer = AutoTokenizer.from_pretrained("optimum/bert-base-uncased_SWAG-neuronx")
>>> model = NeuronModelForMultipleChoice.from_pretrained("optimum/bert-base-uncased_SWAG-neuronx")

>>> num_choices = 4
>>> first_sentence = ["Members of the procession walk down the street holding small horn brass instruments."] * num_choices
>>> second_sentence = [
...     "A drum line passes by walking down the street playing their instruments.",
...     "A drum line has heard approaching them.",
...     "A drum line arrives and they're outside dancing and asleep.",
...     "A drum line turns the lead singer watches the performance."
... ]
>>> inputs = tokenizer(first_sentence, second_sentence, truncation=True, padding=True)

# Unflatten the inputs values expanding it to the shape [batch_size, num_choices, seq_length]
>>> for k, v in inputs.items():
...     inputs[k] = [v[i: i + num_choices] for i in range(0, len(v), num_choices)]
>>> inputs = dict(inputs.convert_to_tensors(tensor_type="pt"))
>>> outputs = model(**inputs)
>>> logits = outputs.logits
>>> logits.shape
[1, 4]

NeuronModelForCausalLM

class optimum.neuron.NeuronModelForCausalLM

< source >

( config: PretrainedConfig checkpoint_dir: Union compiled_dir: Union = None generation_config: Optional = None )

Parameters

model (torch.nn.Module) — torch.nn.Module is the neuron decoder graph.
config (transformers.PretrainedConfig) — PretrainedConfig is the Model configuration class with all the parameters of the model.
model_path (Path) — The directory where the compiled artifacts for the model are stored. It can be a temporary directory if the model has never been saved locally before.
generation_config (transformers.GenerationConfig) — GenerationConfig holds the configuration for the model generation task.

Neuron model with a causal language modeling head for inference on Neuron devices.

This model inherits from ~neuron.modeling.NeuronDecoderModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

forward

< source >

( input_ids: Tensor cache_ids: Tensor start_ids: Tensor = None return_dict: bool = True )

Parameters

input_ids (torch.LongTensor) — Indices of decoder input sequence tokens in the vocabulary of shape (batch_size, sequence_length).
cache_ids (torch.LongTensor) — The indices at which the cached key and value for the current inputs need to be stored.
start_ids (torch.LongTensor) — The indices of the first tokens to be processed, deduced form the attention masks.

The NeuronModelForCausalLM forward method, overrides the __call__ special method.

Example of text generation:

>>> from transformers import AutoTokenizer
>>> from optimum.neuron import NeuronModelForCausalLM
>>> import torch

>>> tokenizer = AutoTokenizer.from_pretrained("gpt2")
>>> model = NeuronModelForCausalLM.from_pretrained("gpt2", export=True)

>>> inputs = tokenizer("My favorite moment of the day is", return_tensors="pt")

>>> gen_tokens = model.generate(**inputs, do_sample=True, temperature=0.9, min_length=20, max_length=20)
>>> tokenizer.batch_decode(gen_tokens)

NeuronModelForSeq2SeqLM

class optimum.neuron.NeuronModelForSeq2SeqLM

< source >

( encoder: ScriptModule decoder: ScriptModule config: PretrainedConfig model_save_dir: Union = None encoder_file_name: Optional = 'model.neuron' decoder_file_name: Optional = 'model.neuron' preprocessors: Optional = None neuron_configs: Optional = None configs: Optional = None generation_config: Optional = None **kwargs )

Parameters

encoder (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module of the encoder with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.
decoder (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module of the decoder with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.
config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.

Neuron Sequence-to-sequence model with a language modeling head for text2text-generation tasks.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

forward

< source >

( attention_mask: Optional = None decoder_input_ids: Optional = None decoder_attention_mask: Optional = None encoder_outputs: Optional = None beam_scores: Optional = None return_dict: bool = False output_attentions: bool = False output_hidden_states: bool = False )

Parameters

input_ids (torch.Tensor of shape (batch_size, sequence_length)) — Indices of input sequence tokens in the vocabulary. Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode and PreTrainedTokenizer.__call__ for details. What are input IDs?
attention_mask (Union[torch.Tensor, None] of shape (batch_size, sequence_length), defaults to None) — 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. What are attention masks?

The NeuronModelForSeq2SeqLM forward method, overrides the __call__ special method.

(Following models are compiled with neuronx compiler and can only be run on INF2.)

Example of text-to-text generation with small T5 model:

from transformers import AutoTokenizer
from optimum.neuron import NeuronModelForSeq2SeqLM

neuron_model = NeuronModelForSeq2SeqLM.from_pretrained(google-t5/t5-small, export=True, dynamic_batch_size=False, batch_size=1, sequence_length=64, num_beams=4)
neuron_model.save_pretrained("t5_small_neuronx")
del neuron_model

neuron_model = NeuronModelForSeq2SeqLM.from_pretrained("t5_small_neuronx")
tokenizer = AutoTokenizer.from_pretrained("t5_small_neuronx")
inputs = tokenizer("translate English to German: Lets eat good food.", return_tensors="pt")

output = neuron_model.generate(
    **inputs,
    num_return_sequences=1,
)
results = [tokenizer.decode(t, skip_special_tokens=True) for t in output]

(For large models, in order to fit into Neuron cores, we need to apply tensor parallelism. Here below is an example ran on inf2.24xlarge.)

Example of text-to-text generation with tensor parallelism:

from transformers import AutoTokenizer
from optimum.neuron import NeuronModelForSeq2SeqLM
# 1. compile
if __name__ == "__main__":  # compulsory for parallel tracing since the API will spawn multiple processes.
    neuron_model = NeuronModelForSeq2SeqLM.from_pretrained(
        google/flan-t5-xl, export=True, tensor_parallel_size=8, dynamic_batch_size=False, batch_size=1, sequence_length=128, num_beams=4,
    )
    neuron_model.save_pretrained("flan_t5_xl_neuronx_tp8/")
    del neuron_model

# 2. inference
neuron_model = NeuronModelForSeq2SeqLM.from_pretrained("flan_t5_xl_neuronx_tp8")
tokenizer = AutoTokenizer.from_pretrained("flan_t5_xl_neuronx_tp8")
inputs = tokenizer("translate English to German: Lets eat good food.", return_tensors="pt")

output = neuron_model.generate(
    **inputs,
    num_return_sequences=1,
)
results = [tokenizer.decode(t, skip_special_tokens=True) for t in output]

Computer Vision

The following Neuron model classes are available for computer vision tasks.

NeuronModelForImageClassification

class optimum.neuron.NeuronModelForImageClassification

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with an image classification head on top (a linear layer on top of the final hidden state of the [CLS] token) e.g. for ImageNet.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Neuron Model for image-classification tasks. This class officially supports beit, convnext, convnextv2, deit, levit, mobilenet_v2, mobilevit, vit, etc.

forward

< source >

( pixel_values: Tensor **kwargs )

Parameters

pixel_values (Union[torch.Tensor, None] of shape (batch_size, num_channels, height, width), defaults to None) — Pixel values corresponding to the images in the current batch. Pixel values can be obtained from encoded images using AutoImageProcessor.

The NeuronModelForImageClassification forward method, overrides the __call__ special method.

Example of image classification:

>>> import requests
>>> from PIL import Image
>>> from optimum.neuron import NeuronModelForImageClassification
>>> from transformers import AutoImageProcessor

>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)

>>> preprocessor = AutoImageProcessor.from_pretrained("optimum/vit-base-patch16-224-neuronx")
>>> model = NeuronModelForImageClassification.from_pretrained("optimum/vit-base-patch16-224-neuronx")

>>> inputs = preprocessor(images=image, return_tensors="pt")

>>> outputs = model(**inputs)
>>> logits = outputs.logits
>>> predicted_label = logits.argmax(-1).item()

Example using optimum.neuron.pipeline:

>>> import requests
>>> from PIL import Image
>>> from transformers import AutoImageProcessor
>>> from optimum.neuron import NeuronModelForImageClassification, pipeline

>>> preprocessor = AutoImageProcessor.from_pretrained("optimum/vit-base-patch16-224-neuronx")
>>> model = NeuronModelForImageClassification.from_pretrained("optimum/vit-base-patch16-224-neuronx")
>>> pipe = pipeline("image-classification", model=model, feature_extractor=preprocessor)

>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> pred = pipe(url)

NeuronModelForSemanticSegmentation

class optimum.neuron.NeuronModelForSemanticSegmentation

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with a semantic segmentation head on top, e.g. for Pascal VOC.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Neuron Model for semantic-segmentation, with an all-MLP decode head on top e.g. for ADE20k, CityScapes. This class officially supports mobilevit, mobilenet-v2, etc.

forward

< source >

( pixel_values: Tensor **kwargs )

Parameters

pixel_values (Union[torch.Tensor, None] of shape (batch_size, num_channels, height, width), defaults to None) — Pixel values corresponding to the images in the current batch. Pixel values can be obtained from encoded images using AutoImageProcessor.

The NeuronModelForSemanticSegmentation forward method, overrides the __call__ special method.

Example of semantic segmentation:

>>> import requests
>>> from PIL import Image
>>> from optimum.neuron import NeuronModelForSemanticSegmentation
>>> from transformers import AutoImageProcessor

>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)

>>> preprocessor = AutoImageProcessor.from_pretrained("optimum/deeplabv3-mobilevit-small-neuronx")
>>> model = NeuronModelForSemanticSegmentation.from_pretrained("optimum/deeplabv3-mobilevit-small-neuronx")

>>> inputs = preprocessor(images=image, return_tensors="pt")

>>> outputs = model(**inputs)
>>> logits = outputs.logits

Example using optimum.neuron.pipeline:

>>> import requests
>>> from PIL import Image
>>> from transformers import AutoImageProcessor
>>> from optimum.neuron import NeuronModelForSemanticSegmentation, pipeline

>>> preprocessor = AutoImageProcessor.from_pretrained("optimum/deeplabv3-mobilevit-small-neuronx")
>>> model = NeuronModelForSemanticSegmentation.from_pretrained("optimum/deeplabv3-mobilevit-small-neuronx")
>>> pipe = pipeline("image-segmentation", model=model, feature_extractor=preprocessor)

>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> pred = pipe(url)

NeuronModelForObjectDetection

class optimum.neuron.NeuronModelForObjectDetection

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with object detection heads on top, for tasks such as COCO detection.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Neuron Model for object-detection, with object detection heads on top, for tasks such as COCO detection.

forward

< source >

( pixel_values: Tensor **kwargs )

Parameters

pixel_values (Union[torch.Tensor, None] of shape (batch_size, num_channels, height, width), defaults to None) — Pixel values corresponding to the images in the current batch. Pixel values can be obtained from encoded images using AutoImageProcessor.

The NeuronModelForObjectDetection forward method, overrides the __call__ special method.

Example of object detection:

>>> import requests
>>> from PIL import Image
>>> from optimum.neuron import NeuronModelForObjectDetection
>>> from transformers import AutoImageProcessor

>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> image = Image.open(requests.get(url, stream=True).raw)

>>> preprocessor = AutoImageProcessor.from_pretrained("hustvl/yolos-tiny")
>>> model = NeuronModelForObjectDetection.from_pretrained("hustvl/yolos-tiny", export=True, batch_size=1)

>>> inputs = preprocessor(images=image, return_tensors="pt")

>>> outputs = model(**inputs)
>>> target_sizes = torch.tensor([image.size[::-1]])
>>> results = image_processor.post_process_object_detection(outputs, threshold=0.9, target_sizes=target_sizes)[0]

Example using optimum.neuron.pipeline:

>>> import requests
>>> from PIL import Image
>>> from transformers import AutoImageProcessor
>>> from optimum.neuron import NeuronModelForObjectDetection, pipeline

>>> preprocessor = AutoImageProcessor.from_pretrained("hustvl/yolos-tiny")
>>> model = NeuronModelForObjectDetection.from_pretrained("hustvl/yolos-tiny")
>>> pipe = pipeline("object-detection", model=model, feature_extractor=preprocessor)

>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
>>> pred = pipe(url)

Audio

The following auto classes are available for the following audio tasks.

NeuronModelForAudioClassification

class optimum.neuron.NeuronModelForAudioClassification

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with an audio classification head.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Neuron Model for audio-classification, with a sequence classification head on top (a linear layer over the pooled output) for tasks like SUPERB Keyword Spotting.

forward

< source >

( input_values: Tensor **kwargs )

Parameters

input_values (torch.Tensor of shape (batch_size, sequence_length)) — Float values of input raw speech waveform.. Input values can be obtained from audio file loaded into an array using AutoProcessor.

The NeuronModelForAudioClassification forward method, overrides the __call__ special method.

Example of audio classification:

>>> from transformers import AutoProcessor
>>> from optimum.neuron import NeuronModelForAudioClassification
>>> from datasets import load_dataset
>>> import torch

>>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
>>> dataset = dataset.sort("id")
>>> sampling_rate = dataset.features["audio"].sampling_rate

>>> feature_extractor = AutoProcessor.from_pretrained("Jingya/wav2vec2-large-960h-lv60-self-neuronx-audio-classification")
>>> model = NeuronModelForAudioClassification.from_pretrained("Jingya/wav2vec2-large-960h-lv60-self-neuronx-audio-classification")

>>> # audio file is decoded on the fly
>>> inputs = feature_extractor(dataset[0]["audio"]["array"], sampling_rate=sampling_rate, return_tensors="pt")

>>> logits = model(**inputs).logits
>>> predicted_class_ids = torch.argmax(logits, dim=-1).item()
>>> predicted_label = model.config.id2label[predicted_class_ids]

Example using optimum.neuron.pipeline:

>>> from transformers import AutoProcessor
>>> from optimum.neuron import NeuronModelForAudioClassification, pipeline

>>> feature_extractor = AutoProcessor.from_pretrained("Jingya/wav2vec2-large-960h-lv60-self-neuronx-audio-classification")
>>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
>>> dataset = dataset.sort("id")

>>> model = NeuronModelForAudioClassification.from_pretrained("Jingya/wav2vec2-large-960h-lv60-self-neuronx-audio-classification")
>>> ac = pipeline("audio-classification", model=model, feature_extractor=feature_extractor)

>>> pred = ac(dataset[0]["audio"]["array"])

NeuronModelForAudioFrameClassification

class optimum.neuron.NeuronModelForAudioFrameClassification

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with an audio frame classification head.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Neuron Model with a frame classification head on top for tasks like Speaker Diarization.

forward

< source >

( input_values: Tensor **kwargs )

Parameters

input_values (torch.Tensor of shape (batch_size, sequence_length)) — Float values of input raw speech waveform.. Input values can be obtained from audio file loaded into an array using AutoProcessor.

The NeuronModelForAudioFrameClassification forward method, overrides the __call__ special method.

Example of audio frame classification:

>>> from transformers import AutoProcessor
>>> from optimum.neuron import NeuronModelForAudioFrameClassification
>>> from datasets import load_dataset
>>> import torch

>>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
>>> dataset = dataset.sort("id")
>>> sampling_rate = dataset.features["audio"].sampling_rate

>>> feature_extractor = AutoProcessor.from_pretrained("Jingya/wav2vec2-base-superb-sd-neuronx")
>>> model =  NeuronModelForAudioFrameClassification.from_pretrained("Jingya/wav2vec2-base-superb-sd-neuronx")

>>> inputs = feature_extractor(dataset[0]["audio"]["array"], return_tensors="pt", sampling_rate=sampling_rate)
>>> logits = model(**inputs).logits

>>> probabilities = torch.sigmoid(logits[0])
>>> labels = (probabilities > 0.5).long()
>>> labels[0].tolist()

NeuronModelForCTC

class optimum.neuron.NeuronModelForCTC

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with a connectionist temporal classification head.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Neuron Model with a language modeling head on top for Connectionist Temporal Classification (CTC).

forward

< source >

( input_values: Tensor **kwargs )

Parameters

input_values (torch.Tensor of shape (batch_size, sequence_length)) — Float values of input raw speech waveform.. Input values can be obtained from audio file loaded into an array using AutoProcessor.

The NeuronModelForCTC forward method, overrides the __call__ special method.

Example of CTC:

>>> from transformers import AutoProcessor, Wav2Vec2ForCTC
>>> from optimum.neuron import NeuronModelForCTC
>>> from datasets import load_dataset
>>> import torch

>>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
>>> dataset = dataset.sort("id")
>>> sampling_rate = dataset.features["audio"].sampling_rate

>>> processor = AutoProcessor.from_pretrained("Jingya/wav2vec2-large-960h-lv60-self-neuronx-ctc")
>>> model = NeuronModelForCTC.from_pretrained("Jingya/wav2vec2-large-960h-lv60-self-neuronx-ctc")

>>> # audio file is decoded on the fly
>>> inputs = processor(dataset[0]["audio"]["array"], sampling_rate=sampling_rate, return_tensors="pt")
>>> logits = model(**inputs).logits
>>> predicted_ids = torch.argmax(logits, dim=-1)

>>> transcription = processor.batch_decode(predicted_ids)

Example using optimum.neuron.pipeline:

>>> from transformers import AutoProcessor
>>> from optimum.neuron import NeuronModelForCTC, pipeline

>>> processor = AutoProcessor.from_pretrained("Jingya/wav2vec2-large-960h-lv60-self-neuronx-ctc")
>>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
>>> dataset = dataset.sort("id")

>>> model = NeuronModelForCTC.from_pretrained("Jingya/wav2vec2-large-960h-lv60-self-neuronx-ctc")
>>> asr = pipeline("automatic-speech-recognition", model=model, feature_extractor=processor.feature_extractor, tokenizer=processor.tokenizer)

NeuronModelForXVector

class optimum.neuron.NeuronModelForXVector

< source >

( model: ScriptModule config: PretrainedConfig model_save_dir: Union = None model_file_name: Optional = None preprocessors: Optional = None neuron_config: Optional = None **kwargs )

Parameters

config (transformers.PretrainedConfig) — PretrainedConfig is the 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 optimum.neuron.modeling.NeuronTracedModel.from_pretrained method to load the model weights.
model (torch.jit._script.ScriptModule) — torch.jit._script.ScriptModule is the TorchScript module with embedded NEFF(Neuron Executable File Format) compiled by neuron(x) compiler.

Neuron Model with an XVector feature extraction head on top for tasks like Speaker Verification.

This model inherits from ~neuron.modeling.NeuronTracedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving)

Neuron Model with an XVector feature extraction head on top for tasks like Speaker Verification.

forward

< source >

( input_values: Tensor **kwargs )

Parameters

input_values (torch.Tensor of shape (batch_size, sequence_length)) — Float values of input raw speech waveform.. Input values can be obtained from audio file loaded into an array using AutoProcessor.

The NeuronModelForXVector forward method, overrides the __call__ special method.

Example of Audio XVector:

>>> from transformers import AutoProcessor
>>> from optimum.neuron import NeuronModelForXVector
>>> from datasets import load_dataset
>>> import torch

>>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
>>> dataset = dataset.sort("id")
>>> sampling_rate = dataset.features["audio"].sampling_rate

>>> feature_extractor = AutoProcessor.from_pretrained("Jingya/wav2vec2-base-superb-sv-neuronx")
>>> model = NeuronModelForXVector.from_pretrained("Jingya/wav2vec2-base-superb-sv-neuronx")

>>> inputs = feature_extractor(
...     [d["array"] for d in dataset[:2]["audio"]], sampling_rate=sampling_rate, return_tensors="pt", padding=True
... )
>>> embeddings = model(**inputs).embeddings

>>> embeddings = torch.nn.functional.normalize(embeddings, dim=-1)

>>> cosine_sim = torch.nn.CosineSimilarity(dim=-1)
>>> similarity = cosine_sim(embeddings[0], embeddings[1])
>>> threshold = 0.7
>>> if similarity < threshold:
...     print("Speakers are not the same!")
>>> round(similarity.item(), 2)

Stable Diffusion

The following Neuron model classes are available for stable diffusion tasks.

NeuronStableDiffusionPipeline

Parameters

prompt (Optional[Union[str, List[str]]], defaults to None) — The prompt or prompts to guide image generation. If not defined, you need to pass prompt_embeds.
image (Optional["PipelineImageInput"], defaults to None) — The ControlNet input condition to provide guidance to the unet for generation. If the type is specified as torch.Tensor, it is passed to ControlNet as is. PIL.Image.Image can also be accepted as an image. The dimensions of the output image defaults to image’s dimensions. If height and/or width are passed, image is resized accordingly. If multiple ControlNets are specified in init, images must be passed as a list such that each element of the list can be correctly batched for input to a single ControlNet. When prompt is a list, and if a list of images is passed for a single ControlNet, each will be paired with each prompt in the prompt list. This also applies to multiple ControlNets, where a list of image lists can be passed to batch for each prompt and each ControlNet.
num_inference_steps (int, defaults to 50) — The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference.
timesteps (Optional[List[int]], defaults to None) — Custom timesteps to use for the denoising process with schedulers which support a timesteps argument in their set_timesteps method. If not defined, the default behavior when num_inference_steps is passed will be used. Must be in descending order.
sigmas (Optional[List[int]], defaults to None) — Custom sigmas to use for the denoising process with schedulers which support a sigmas argument in their set_timesteps method. If not defined, the default behavior when num_inference_steps is passed will be used.
guidance_scale (float, defaults to 7.5) — A higher guidance scale value encourages the model to generate images closely linked to the text prompt at the expense of lower image quality. Guidance scale is enabled when guidance_scale > 1.
negative_prompt (Optional[Union[str, List[str]]], defaults to None) — The prompt or prompts to guide what to not include in image generation. If not defined, you need to pass negative_prompt_embeds instead. Ignored when not using guidance (guidance_scale < 1).
num_images_per_prompt (int, defaults to 1) — The number of images to generate per prompt. If it is different from the batch size used for the compiltaion, it will be overriden by the static batch size of neuron (except for dynamic batching).
eta (float, defaults to 0.0) — Corresponds to parameter eta (η) from the DDIM paper. Only applies to the diffusers.schedulers.DDIMScheduler, and is ignored in other schedulers.
generator (Optional[Union[torch.Generator, List[torch.Generator]]], defaults to None) — A torch.Generator to make generation deterministic.
latents (Optional[torch.Tensor], defaults to None) — Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image generation. Can be used to tweak the same generation with different prompts. If not provided, a latents tensor is generated by sampling using the supplied random generator.
prompt_embeds (Optional[torch.Tensor], defaults to None) — Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, text embeddings are generated from the prompt input argument.
negative_prompt_embeds (Optional[torch.Tensor], defaults to None) — Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, negative_prompt_embeds are generated from the negative_prompt input argument. ip_adapter_image — (Optional[PipelineImageInput], defaults to None): Optional image input to work with IP Adapters.
ip_adapter_image_embeds (Optional[List[torch.Tensor]], defaults to None) — Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of IP-adapters. Each element should be a tensor of shape (batch_size, num_images, emb_dim). It should contain the negative image embedding if do_classifier_free_guidance is set to True. If not provided, embeddings are computed from the ip_adapter_image input argument.
output_type (str, defaults to "pil") — The output format of the generated image. Choose between PIL.Image or np.array.
return_dict (bool, defaults to True) — Whether or not to return a diffusers.pipelines.stable_diffusion.StableDiffusionPipelineOutput instead of a plain tuple.
cross_attention_kwargs (Optional[Dict[str, Any]], defaults to None) — A kwargs dictionary that if specified is passed along to the AttentionProcessor as defined in self.processor.
controlnet_conditioning_scale (Union[float, List[float]], defaults to 1.0) — The outputs of the ControlNet are multiplied by controlnet_conditioning_scale before they are added to the residual in the original unet. If multiple ControlNets are specified in init, you can set the corresponding scale as a list.
guess_mode (bool, defaults to False) — The ControlNet encoder tries to recognize the content of the input image even if you remove all prompts. A guidance_scale value between 3.0 and 5.0 is recommended.
control_guidance_start (Union[float, List[float]], defaults to 0.0) — The percentage of total steps at which the ControlNet starts applying.
control_guidance_end (Union[float, List[float]], optional, defaults to 1.0) — The percentage of total steps at which the ControlNet stops applying.
clip_skip (Optional[int], defaults to None) — Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that the output of the pre-final layer will be used for computing the prompt embeddings.
callback_on_step_end (Optional[Union[Callable[[int, int, Dict], None], PipelineCallback, MultiPipelineCallbacks]], defaults to None) — A function or a subclass of PipelineCallback or MultiPipelineCallbacks that is called at the end of each denoising step during the inference. with the following arguments: callback_on_step_end(self: DiffusionPipeline, step: int, timestep: int, callback_kwargs: Dict). callback_kwargs will include a list of all tensors as specified by callback_on_step_end_tensor_inputs.
callback_on_step_end_tensor_inputs (List[str], defaults to ["latents"]) — The list of tensor inputs for the callback_on_step_end function. The tensors specified in the list will be passed as callback_kwargs argument. You will only be able to include variables listed in the ._callback_tensor_inputs attribute of your pipeline class.

< source >

call

< source >

( prompt: Union = None prompt_2: Union = None image: Union = None num_inference_steps: int = 50 timesteps: List = None sigmas: List = None denoising_end: Optional = None guidance_scale: float = 5.0 negative_prompt: Union = None negative_prompt_2: Union = None num_images_per_prompt: Optional = 1 eta: float = 0.0 generator: Union = None latents: Optional = None prompt_embeds: Optional = None negative_prompt_embeds: Optional = None pooled_prompt_embeds: Optional = None negative_pooled_prompt_embeds: Optional = None ip_adapter_image: Union = None ip_adapter_image_embeds: Optional = None output_type: Optional = 'pil' return_dict: bool = True cross_attention_kwargs: Optional = None controlnet_conditioning_scale: Union = 1.0 guess_mode: bool = False control_guidance_start: Union = 0.0 control_guidance_end: Union = 1.0 original_size: Optional = None crops_coords_top_left: Tuple = (0, 0) target_size: Optional = None negative_original_size: Optional = None negative_crops_coords_top_left: Tuple = (0, 0) negative_target_size: Optional = None clip_skip: Optional = None callback_on_step_end: Union = None callback_on_step_end_tensor_inputs: List = ['latents'] **kwargs ) → diffusers.pipelines.stable_diffusion.StableDiffusionPipelineOutput or tuple

Parameters

prompt (Optional[Union[str, List[str]]], defaults to None) — The prompt or prompts to guide image generation. If not defined, you need to pass prompt_embeds.
prompt_2 (Optional[Union[str, List[str]]], defaults to None) — The prompt or prompts to be sent to tokenizer_2 and text_encoder_2. If not defined, prompt is used in both text-encoders.
image (Optional["PipelineImageInput"], defaults to None) — The ControlNet input condition to provide guidance to the unet for generation. If the type is specified as torch.Tensor, it is passed to ControlNet as is. PIL.Image.Image can also be accepted as an image. The dimensions of the output image defaults to image’s dimensions. If height and/or width are passed, image is resized accordingly. If multiple ControlNets are specified in init, images must be passed as a list such that each element of the list can be correctly batched for input to a single ControlNet.
num_inference_steps (int, defaults to 50) — The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference.
timesteps (Optional[List[int]], defaults to None) — Custom timesteps to use for the denoising process with schedulers which support a timesteps argument in their set_timesteps method. If not defined, the default behavior when num_inference_steps is passed will be used. Must be in descending order.
sigmas (Optional[List[int]], defaults to None) — Custom sigmas to use for the denoising process with schedulers which support a sigmas argument in their set_timesteps method. If not defined, the default behavior when num_inference_steps is passed will be used.
denoising_end (Optional[float], defaults to None) — When specified, determines the fraction (between 0.0 and 1.0) of the total denoising process to be completed before it is intentionally prematurely terminated. As a result, the returned sample will still retain a substantial amount of noise as determined by the discrete timesteps selected by the scheduler. The denoising_end parameter should ideally be utilized when this pipeline forms a part of a “Mixture of Denoisers” multi-pipeline setup, as elaborated in Refining the Image Output
guidance_scale (float, defaults to 5.0) — A higher guidance scale value encourages the model to generate images closely linked to the text prompt at the expense of lower image quality. Guidance scale is enabled when guidance_scale > 1.
negative_prompt (Optional[Union[str, List[str]]], defaults to None) — The prompt or prompts to guide what to not include in image generation. If not defined, you need to pass negative_prompt_embeds instead. Ignored when not using guidance (guidance_scale < 1).
negative_prompt_2 (Optional[Union[str, List[str]]], defaults to None) — The prompt or prompts to guide what to not include in image generation. This is sent to tokenizer_2 and text_encoder_2. If not defined, negative_prompt is used in both text-encoders.
num_images_per_prompt (int, defaults to 1) — The number of images to generate per prompt.
eta (float, defaults to 0.0) — Corresponds to parameter eta (η) from the DDIM paper. Only applies to the diffusers.schedulers.DDIMScheduler, and is ignored in other schedulers.
generator (Optional[Union[torch.Generator, List[torch.Generator]]], defaults to None) — A torch.Generator to make generation deterministic.
latents (Optional[torch.Tensor], defaults to None) — Pre-generated noisy latents sampled from a Gaussian distribution, to be used as inputs for image generation. Can be used to tweak the same generation with different prompts. If not provided, a latents tensor is generated by sampling using the supplied random generator.
prompt_embeds (Optional[torch.Tensor], defaults to None) — Pre-generated text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, text embeddings are generated from the prompt input argument.
negative_prompt_embeds (Optional[torch.Tensor], defaults to None) — Pre-generated negative text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, negative_prompt_embeds are generated from the negative_prompt input argument.
pooled_prompt_embeds (Optional[torch.Tensor], defaults to None) — Pre-generated pooled text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, pooled text embeddings are generated from prompt input argument.
negative_pooled_prompt_embeds (Optional[torch.Tensor], defaults to None) — Pre-generated negative pooled text embeddings. Can be used to easily tweak text inputs (prompt weighting). If not provided, pooled negative_prompt_embeds are generated from negative_prompt input argument. ip_adapter_image — (Optional[PipelineImageInput], defaults to None): Optional image input to work with IP Adapters.
ip_adapter_image_embeds (Optional[List[torch.Tensor]], defaults to None) — Pre-generated image embeddings for IP-Adapter. It should be a list of length same as number of IP-adapters. Each element should be a tensor of shape (batch_size, num_images, emb_dim). It should contain the negative image embedding if do_classifier_free_guidance is set to True. If not provided, embeddings are computed from the ip_adapter_image input argument.
output_type (Optional[str], defaults to "pil") — The output format of the generated image. Choose between PIL.Image or np.array.
return_dict (bool, defaults to True) — Whether or not to return a ~pipelines.stable_diffusion.StableDiffusionPipelineOutput instead of a plain tuple.
cross_attention_kwargs (Optional[Dict[str, Any]], defaults to None) — A kwargs dictionary that if specified is passed along to the AttentionProcessor as defined in self.processor.
controlnet_conditioning_scale (Union[float, List[float]], defaults to 1.0) — The outputs of the ControlNet are multiplied by controlnet_conditioning_scale before they are added to the residual in the original unet. If multiple ControlNets are specified in init, you can set the corresponding scale as a list.
guess_mode (bool, defaults to False) — The ControlNet encoder tries to recognize the content of the input image even if you remove all prompts. A guidance_scale value between 3.0 and 5.0 is recommended.
control_guidance_start (Union[float, List[float]], defaults to 0.0) — The percentage of total steps at which the ControlNet starts applying.
control_guidance_end (Union[float, List[float]], defaults to 1.0) — The percentage of total steps at which the ControlNet stops applying.
original_size (Optional[Tuple[int, int]], defaults to (1024, 1024)) — If original_size is not the same as target_size the image will appear to be down- or upsampled. original_size defaults to (height, width) if not specified. Part of SDXL’s micro-conditioning as explained in section 2.2 of https://huggingface.co./papers/2307.01952.
crops_coords_top_left (Tuple[int, int], defaults to (0, 0)) — crops_coords_top_left can be used to generate an image that appears to be “cropped” from the position crops_coords_top_left downwards. Favorable, well-centered images are usually achieved by setting crops_coords_top_left to (0, 0). Part of SDXL’s micro-conditioning as explained in section 2.2 of https://huggingface.co./papers/2307.01952.
target_size (Optional[Tuple[int, int]], defaults to None) — For most cases, target_size should be set to the desired height and width of the generated image. If not specified it will default to (height, width). Part of SDXL’s micro-conditioning as explained in section 2.2 of https://huggingface.co./papers/2307.01952.
negative_original_size (Optional[Tuple[int, int]], defaults to None) — To negatively condition the generation process based on a specific image resolution. Part of SDXL’s micro-conditioning as explained in section 2.2 of https://huggingface.co./papers/2307.01952. For more information, refer to this issue thread: https://github.com/huggingface/diffusers/issues/4208.
negative_crops_coords_top_left (Tuple[int, int], defaults to (0, 0)) — To negatively condition the generation process based on a specific crop coordinates. Part of SDXL’s micro-conditioning as explained in section 2.2 of https://huggingface.co./papers/2307.01952. For more information, refer to this issue thread: https://github.com/huggingface/diffusers/issues/4208.
negative_target_size (Optional[Tuple[int, int]], defaults to None) — To negatively condition the generation process based on a target image resolution. It should be as same as the target_size for most cases. Part of SDXL’s micro-conditioning as explained in section 2.2 of https://huggingface.co./papers/2307.01952. For more information, refer to this issue thread: https://github.com/huggingface/diffusers/issues/4208.
clip_skip (Optional[int], defaults to None) — Number of layers to be skipped from CLIP while computing the prompt embeddings. A value of 1 means that the output of the pre-final layer will be used for computing the prompt embeddings.
callback_on_step_end (Optional[Union[Callable[[int, int, Dict], None], PipelineCallback, MultiPipelineCallbacks]], defaults to None) — A function or a subclass of PipelineCallback or MultiPipelineCallbacks that is called at the end of each denoising step during the inference. with the following arguments: callback_on_step_end(self: DiffusionPipeline, step: int, timestep: int, callback_kwargs: Dict). callback_kwargs will include a list of all tensors as specified by callback_on_step_end_tensor_inputs.
callback_on_step_end_tensor_inputs (List[str], defaults to ["latents"]) — The list of tensor inputs for the callback_on_step_end function. The tensors specified in the list will be passed as callback_kwargs argument. You will only be able to include variables listed in the ._callback_tensor_inputs attribute of your pipeline class.

Returns

diffusers.pipelines.stable_diffusion.StableDiffusionPipelineOutput or tuple

If return_dict is True, diffusers.pipelines.stable_diffusion.StableDiffusionPipelineOutput is returned, otherwise a tuple is returned containing the output images.

The call function to the pipeline for generation.

Examples:

←Neuron Exporter