The MusicGen Melody model was proposed in Simple and Controllable Music Generation by Jade Copet, Felix Kreuk, Itai Gat, Tal Remez, David Kant, Gabriel Synnaeve, Yossi Adi and Alexandre Défossez.
MusicGen Melody is a single stage auto-regressive Transformer model capable of generating high-quality music samples conditioned on text descriptions or audio prompts. The text descriptions are passed through a frozen text encoder model to obtain a sequence of hidden-state representations. MusicGen is then trained to predict discrete audio tokens, or audio codes, conditioned on these hidden-states. These audio tokens are then decoded using an audio compression model, such as EnCodec, to recover the audio waveform.
Through an efficient token interleaving pattern, MusicGen does not require a self-supervised semantic representation of the text/audio prompts, thus eliminating the need to cascade multiple models to predict a set of codebooks (e.g. hierarchically or upsampling). Instead, it is able to generate all the codebooks in a single forward pass.
The abstract from the paper is the following:
We tackle the task of conditional music generation. We introduce MusicGen, a single Language Model (LM) that operates over several streams of compressed discrete music representation, i.e., tokens. Unlike prior work, MusicGen is comprised of a single-stage transformer LM together with efficient token interleaving patterns, which eliminates the need for cascading several models, e.g., hierarchically or upsampling. Following this approach, we demonstrate how MusicGen can generate high-quality samples, while being conditioned on textual description or melodic features, allowing better controls over the generated output. We conduct extensive empirical evaluation, considering both automatic and human studies, showing the proposed approach is superior to the evaluated baselines on a standard text-to-music benchmark. Through ablation studies, we shed light over the importance of each of the components comprising MusicGen.
This model was contributed by ylacombe. The original code can be found here. The pre-trained checkpoints can be found on the Hugging Face Hub.
There are two key differences with MusicGen:
MusicGen Melody is compatible with two generation modes: greedy and sampling. In practice, sampling leads to significantly better results than greedy, thus we encourage sampling mode to be used where possible. Sampling is enabled by default, and can be explicitly specified by setting do_sample=True
in the call to MusicgenMelodyForConditionalGeneration.generate()
, or by overriding the model’s generation config (see below).
Transformers supports both mono (1-channel) and stereo (2-channel) variants of MusicGen Melody. The mono channel versions generate a single set of codebooks. The stereo versions generate 2 sets of codebooks, 1 for each channel (left/right), and each set of codebooks is decoded independently through the audio compression model. The audio streams for each channel are combined to give the final stereo output.
The model can generate an audio sample conditioned on a text and an audio prompt through use of the MusicgenMelodyProcessor to pre-process the inputs.
In the following examples, we load an audio file using the 🤗 Datasets library, which can be pip installed through the command below:
pip install --upgrade pip
pip install datasets[audio]
The audio file we are about to use is loaded as follows:
>>> from datasets import load_dataset
>>> dataset = load_dataset("sanchit-gandhi/gtzan", split="train", streaming=True)
>>> sample = next(iter(dataset))["audio"]
The audio prompt should ideally be free of the low-frequency signals usually produced by instruments such as drums and bass. The Demucs model can be used to separate vocals and other signals from the drums and bass components.
If you wish to use Demucs, you first need to follow the installation steps here before using the following snippet:
from demucs import pretrained
from demucs.apply import apply_model
from demucs.audio import convert_audio
import torch
wav = torch.tensor(sample["array"]).to(torch.float32)
demucs = pretrained.get_model('htdemucs')
wav = convert_audio(wav[None], sample["sampling_rate"], demucs.samplerate, demucs.audio_channels)
wav = apply_model(demucs, wav[None])
You can then use the following snippet to generate music:
>>> from transformers import AutoProcessor, MusicgenMelodyForConditionalGeneration
>>> processor = AutoProcessor.from_pretrained("facebook/musicgen-melody")
>>> model = MusicgenMelodyForConditionalGeneration.from_pretrained("facebook/musicgen-melody")
>>> inputs = processor(
... audio=wav,
... sampling_rate=demucs.samplerate,
... text=["80s blues track with groovy saxophone"],
... padding=True,
... return_tensors="pt",
... )
>>> audio_values = model.generate(**inputs, do_sample=True, guidance_scale=3, max_new_tokens=256)
You can also pass the audio signal directly without using Demucs, although the quality of the generation will probably be degraded:
>>> from transformers import AutoProcessor, MusicgenMelodyForConditionalGeneration
>>> processor = AutoProcessor.from_pretrained("facebook/musicgen-melody")
>>> model = MusicgenMelodyForConditionalGeneration.from_pretrained("facebook/musicgen-melody")
>>> inputs = processor(
... audio=sample["array"],
... sampling_rate=sample["sampling_rate"],
... text=["80s blues track with groovy saxophone"],
... padding=True,
... return_tensors="pt",
... )
>>> audio_values = model.generate(**inputs, do_sample=True, guidance_scale=3, max_new_tokens=256)
The audio outputs are a three-dimensional Torch tensor of shape (batch_size, num_channels, sequence_length)
. To listen to the generated audio samples, you can either play them in an ipynb notebook:
from IPython.display import Audio
sampling_rate = model.config.audio_encoder.sampling_rate
Audio(audio_values[0].numpy(), rate=sampling_rate)
Or save them as a .wav
file using a third-party library, e.g. soundfile
:
>>> import soundfile as sf
>>> sampling_rate = model.config.audio_encoder.sampling_rate
>>> sf.write("musicgen_out.wav", audio_values[0].T.numpy(), sampling_rate)
The same MusicgenMelodyProcessor can be used to pre-process a text-only prompt.
>>> from transformers import AutoProcessor, MusicgenMelodyForConditionalGeneration
>>> processor = AutoProcessor.from_pretrained("facebook/musicgen-melody")
>>> model = MusicgenMelodyForConditionalGeneration.from_pretrained("facebook/musicgen-melody")
>>> inputs = processor(
... text=["80s pop track with bassy drums and synth", "90s rock song with loud guitars and heavy drums"],
... padding=True,
... return_tensors="pt",
... )
>>> audio_values = model.generate(**inputs, do_sample=True, guidance_scale=3, max_new_tokens=256)
The guidance_scale
is used in classifier free guidance (CFG), setting the weighting between the conditional logits (which are predicted from the text prompts) and the unconditional logits (which are predicted from an unconditional or ‘null’ prompt). Higher guidance scale encourages the model to generate samples that are more closely linked to the input prompt, usually at the expense of poorer audio quality. CFG is enabled by setting guidance_scale > 1
. For best results, use guidance_scale=3
(default).
You can also generate in batch:
>>> from transformers import AutoProcessor, MusicgenMelodyForConditionalGeneration
>>> from datasets import load_dataset
>>> processor = AutoProcessor.from_pretrained("facebook/musicgen-melody")
>>> model = MusicgenMelodyForConditionalGeneration.from_pretrained("facebook/musicgen-melody")
>>> # take the first quarter of the audio sample
>>> sample_1 = sample["array"][: len(sample["array"]) // 4]
>>> # take the first half of the audio sample
>>> sample_2 = sample["array"][: len(sample["array"]) // 2]
>>> inputs = processor(
... audio=[sample_1, sample_2],
... sampling_rate=sample["sampling_rate"],
... text=["80s blues track with groovy saxophone", "90s rock song with loud guitars and heavy drums"],
... padding=True,
... return_tensors="pt",
... )
>>> audio_values = model.generate(**inputs, do_sample=True, guidance_scale=3, max_new_tokens=256)
The inputs for unconditional (or ‘null’) generation can be obtained through the method MusicgenMelodyProcessor.get_unconditional_inputs():
>>> from transformers import MusicgenMelodyForConditionalGeneration, MusicgenMelodyProcessor
>>> model = MusicgenMelodyForConditionalGeneration.from_pretrained("facebook/musicgen-melody")
>>> unconditional_inputs = MusicgenMelodyProcessor.from_pretrained("facebook/musicgen-melody").get_unconditional_inputs(num_samples=1)
>>> audio_values = model.generate(**unconditional_inputs, do_sample=True, max_new_tokens=256)
The default parameters that control the generation process, such as sampling, guidance scale and number of generated tokens, can be found in the model’s generation config, and updated as desired:
>>> from transformers import MusicgenMelodyForConditionalGeneration
>>> model = MusicgenMelodyForConditionalGeneration.from_pretrained("facebook/musicgen-melody")
>>> # inspect the default generation config
>>> model.generation_config
>>> # increase the guidance scale to 4.0
>>> model.generation_config.guidance_scale = 4.0
>>> # decrease the max length to 256 tokens
>>> model.generation_config.max_length = 256
Note that any arguments passed to the generate method will supersede those in the generation config, so setting do_sample=False
in the call to generate will supersede the setting of model.generation_config.do_sample
in the generation config.
The MusicGen model can be de-composed into three distinct stages:
Thus, the MusicGen model can either be used as a standalone decoder model, corresponding to the class MusicgenMelodyForCausalLM, or as a composite model that includes the text encoder and audio encoder, corresponding to the class MusicgenMelodyForConditionalGeneration. If only the decoder needs to be loaded from the pre-trained checkpoint, it can be loaded by first specifying the correct config, or be accessed through the .decoder
attribute of the composite model:
>>> from transformers import AutoConfig, MusicgenMelodyForCausalLM, MusicgenMelodyForConditionalGeneration
>>> # Option 1: get decoder config and pass to `.from_pretrained`
>>> decoder_config = AutoConfig.from_pretrained("facebook/musicgen-melody").decoder
>>> decoder = MusicgenMelodyForCausalLM.from_pretrained("facebook/musicgen-melody", **decoder_config.to_dict())
>>> # Option 2: load the entire composite model, but only return the decoder
>>> decoder = MusicgenMelodyForConditionalGeneration.from_pretrained("facebook/musicgen-melody").decoder
Since the text encoder and audio encoder models are frozen during training, the MusicGen decoder MusicgenMelodyForCausalLM can be trained standalone on a dataset of encoder hidden-states and audio codes. For inference, the trained decoder can be combined with the frozen text encoder and audio encoder to recover the composite MusicgenMelodyForConditionalGeneration model.
src/transformers/models/musicgen_melody/convert_musicgen_melody_transformers.py
with the following command:python src/transformers/models/musicgen_melody/convert_musicgen_melody_transformers.py \
--checkpoint="facebook/musicgen-melody" --pytorch_dump_folder /output/path
Tips:
do_sample
in the call to MusicgenMelodyForConditionalGeneration.generate()
( vocab_size = 2048 max_position_embeddings = 2048 num_hidden_layers = 24 ffn_dim = 4096 num_attention_heads = 16 layerdrop = 0.0 use_cache = True activation_function = 'gelu' hidden_size = 1024 dropout = 0.1 attention_dropout = 0.0 activation_dropout = 0.0 initializer_factor = 0.02 scale_embedding = False num_codebooks = 4 audio_channels = 1 pad_token_id = 2048 bos_token_id = 2048 eos_token_id = None tie_word_embeddings = False **kwargs )
Parameters
int
, optional, defaults to 2048) —
Vocabulary size of the MusicgenMelodyDecoder model. Defines the number of different tokens that can be
represented by the inputs_ids
passed when calling MusicgenMelodyDecoder
. int
, optional, defaults to 2048) —
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). int
, optional, defaults to 24) —
Number of decoder layers. int
, optional, defaults to 4096) —
Dimensionality of the “intermediate” (often named feed-forward) layer in the Transformer block. int
, optional, defaults to 16) —
Number of attention heads for each attention layer in the Transformer block. float
, optional, defaults to 0.0) —
The LayerDrop probability for the decoder. See the [LayerDrop paper](see https://arxiv.org/abs/1909.11556)
for more details. bool
, optional, defaults to True
) —
Whether the model should return the last key/values attentions (not used by all models) str
or function
, optional, defaults to "gelu"
) —
The non-linear activation function (function or string) in the decoder and pooler. If string, "gelu"
,
"relu"
, "silu"
and "gelu_new"
are supported. int
, optional, defaults to 1024) —
Dimensionality of the layers and the pooler layer. float
, optional, defaults to 0.1) —
The dropout probability for all fully connected layers in the embeddings, text_encoder, and pooler. float
, optional, defaults to 0.0) —
The dropout ratio for the attention probabilities. float
, optional, defaults to 0.0) —
The dropout ratio for activations inside the fully connected layer. float
, optional, defaults to 0.02) —
The standard deviation of the truncated_normal_initializer for initializing all weight matrices. bool
, optional, defaults to False
) —
Scale embeddings by diving by sqrt(hidden_size). int
, optional, defaults to 4) —
The number of parallel codebooks forwarded to the model. int
, optional, defaults to 1) —
Number of audio channels used by the model (either mono or stereo). Stereo models generate a separate
audio stream for the left/right output channels. Mono models generate a single audio stream output. int
, optional, defaults to 2048) — The id of the padding token. int
, optional, defaults to 2048) — The id of the beginning-of-sequence token. int
, optional) — The id of the end-of-sequence token. bool
, optional, defaults to False
) — Whether to tie word embeddings with the text encoder. This is the configuration class to store the configuration of an MusicgenMelodyDecoder
. It is used to instantiate a
Musicgen Melody decoder according to the specified arguments, defining the model architecture. Instantiating a
configuration with the defaults will yield a similar configuration to that of the Musicgen Melody
facebook/musicgen-melody architecture.
Configuration objects inherit from PretrainedConfig and can be used to control the model outputs. Read the documentation from PretrainedConfig for more information.
( feature_extractor tokenizer )
Parameters
MusicgenMelodyFeatureExtractor
) —
An instance of MusicgenMelodyFeatureExtractor. The feature extractor is a required input. T5Tokenizer
) —
An instance of T5Tokenizer. The tokenizer is a required input. Constructs a MusicGen Melody processor which wraps a Wav2Vec2 feature extractor - for raw audio waveform processing - and a T5 tokenizer into a single processor class.
MusicgenProcessor offers all the functionalities of MusicgenMelodyFeatureExtractor and T5Tokenizer. See
__call__()
and decode() for more information.
( num_samples = 1 return_tensors = 'pt' )
Helper function to get null inputs for unconditional generation, enabling the model to be used without the feature extractor or tokenizer.
Example:
>>> from transformers import MusicgenMelodyForConditionalGeneration, MusicgenMelodyProcessor
>>> model = MusicgenMelodyForConditionalGeneration.from_pretrained("facebook/musicgen-melody")
>>> # get the unconditional (or 'null') inputs for the model
>>> processor = MusicgenMelodyProcessor.from_pretrained("facebook/musicgen-melody")
>>> unconditional_inputs = processor.get_unconditional_inputs(num_samples=1)
>>> audio_samples = model.generate(**unconditional_inputs, max_new_tokens=256)
( feature_size = 12 sampling_rate = 32000 hop_length = 4096 chunk_length = 30 n_fft = 16384 num_chroma = 12 padding_value = 0.0 return_attention_mask = False stem_indices = [3, 2] **kwargs )
Parameters
int
, optional, defaults to 12) —
The feature dimension of the extracted features. int
, optional, defaults to 32000) —
The sampling rate at which the audio files should be digitalized expressed in hertz (Hz). int
, optional, defaults to 4096) —
Length of the overlaping windows for the STFT used to obtain the Mel Frequency coefficients. int
, optional, defaults to 30) —
The maximum number of chunks of sampling_rate
samples used to trim and pad longer or shorter audio
sequences. int
, optional, defaults to 16384) —
Size of the Fourier transform. int
, optional, defaults to 12) —
Number of chroma bins to use. float
, optional, defaults to 0.0) —
Padding value used to pad the audio. bool
, optional, defaults to False
) —
Whether to return the attention mask. Can be overwritten when calling the feature extractor.
For Whisper models, attention_mask
should always be passed for batched inference, to avoid subtle
bugs.
List[int]
, optional, defaults to [3, 2]
) —
Stem channels to extract if demucs outputs are passed. Constructs a MusicgenMelody feature extractor.
This feature extractor inherits from SequenceFeatureExtractor which contains most of the main methods. Users should refer to this superclass for more information regarding those methods.
This class extracts chroma features from audio processed by Demucs or directly from raw audio waveform.
( audio sampling_rate = None )
Extracts stems from the output of the Demucs audio separation model, then converts to mono-channel and resample to the feature extractor sampling rate.
( num_chroma = 12 chroma_length = 235 **kwargs )
Parameters
int
, optional, defaults to 12) — Number of chroma bins to use. int
, optional, defaults to 235) —
Maximum chroma duration if audio is used to condition the model. Corresponds to the maximum duration used during training. This is the configuration class to store the configuration of a MusicgenMelodyModel. It is used to instantiate a Musicgen Melody model according to the specified arguments, defining the text encoder, audio encoder and Musicgen Melody decoder configs. Instantiating a configuration with the defaults will yield a similar configuration to that of the Musicgen Melody facebook/musicgen-melody architecture.
Configuration objects inherit from PretrainedConfig and can be used to control the model outputs. Read the documentation from PretrainedConfig for more information.
Example:
>>> from transformers import (
... MusicgenMelodyConfig,
... MusicgenMelodyDecoderConfig,
... T5Config,
... EncodecConfig,
... MusicgenMelodyForConditionalGeneration,
... )
>>> # Initializing text encoder, audio encoder, and decoder model configurations
>>> text_encoder_config = T5Config()
>>> audio_encoder_config = EncodecConfig()
>>> decoder_config = MusicgenMelodyDecoderConfig()
>>> configuration = MusicgenMelodyConfig.from_sub_models_config(
... text_encoder_config, audio_encoder_config, decoder_config
... )
>>> # Initializing a MusicgenMelodyForConditionalGeneration (with random weights) from the facebook/musicgen-melody style configuration
>>> model = MusicgenMelodyForConditionalGeneration(configuration)
>>> # Accessing the model configuration
>>> configuration = model.config
>>> config_text_encoder = model.config.text_encoder
>>> config_audio_encoder = model.config.audio_encoder
>>> config_decoder = model.config.decoder
>>> # Saving the model, including its configuration
>>> model.save_pretrained("musicgen_melody-model")
>>> # loading model and config from pretrained folder
>>> musicgen_melody_config = MusicgenMelodyConfig.from_pretrained("musicgen_melody-model")
>>> model = MusicgenMelodyForConditionalGeneration.from_pretrained("musicgen_melody-model", config=musicgen_melody_config)
( text_encoder_config: PretrainedConfig audio_encoder_config: PretrainedConfig decoder_config: MusicgenMelodyDecoderConfig **kwargs ) → MusicgenMelodyConfig
Instantiate a MusicgenMelodyConfig (or a derived class) from text encoder, audio encoder and decoder configurations.
( config: MusicgenMelodyDecoderConfig )
Parameters
The bare MusicgenMelody decoder model outputting raw hidden-states without any specific head on top.
The Musicgen Melody model was proposed in Simple and Controllable Music Generation by Jade Copet, Felix Kreuk, Itai Gat, Tal Remez, David Kant, Gabriel Synnaeve, Yossi Adi, Alexandre Défossez. It is a decoder-only transformer trained on the task of conditional music generation.
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.
( input_ids: LongTensor = None attention_mask: Optional = None encoder_hidden_states: Optional = None encoder_attention_mask: Optional = None head_mask: Optional = None past_key_values: Optional = None inputs_embeds: Optional = None use_cache: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None )
Parameters
torch.LongTensor
of shape (batch_size * num_codebooks, sequence_length)
) —
Indices of input sequence tokens in the vocabulary, corresponding to the sequence of audio codes.
Indices can be obtained by encoding an audio prompt with an audio encoder model to predict audio codes, such as with the EncodecModel. See EncodecModel.encode() for details.
The input_ids
will automatically be converted from shape (batch_size * num_codebooks, target_sequence_length)
to (batch_size, num_codebooks, target_sequence_length)
in the forward pass. If
you obtain audio codes from an audio encoding model, such as EncodecModel, ensure that the number of
frames is equal to 1, and that you reshape the audio codes from (frames, batch_size, num_codebooks, target_sequence_length)
to (batch_size * num_codebooks, target_sequence_length)
prior to passing them as
input_ids
.
torch.Tensor
of shape (batch_size, sequence_length)
, optional) —
Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]
:
torch.FloatTensor
of shape (batch_size, encoder_sequence_length, hidden_size)
, optional) —
Sequence of hidden-states representing the concatenation of the text encoder output and the processed audio encoder output.
Used as a conditional signal and will thus be concatenated to the projeted decoder_input_ids
. torch.LongTensor
of shape (batch_size, encoder_sequence_length)
, optional) —
Mask to avoid performing attention on conditional hidden states. Mask values
selected in [0, 1]
:
torch.Tensor
of shape (decoder_layers, decoder_attention_heads)
, optional) —
Mask to nullify selected heads of the attention modules. Mask values selected in [0, 1]
:
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.
If past_key_values
are used, the user can optionally input only the last decoder_input_ids
(those that
don’t have their past key value states given to this model) of shape (batch_size, 1)
instead of all
decoder_input_ids
of shape (batch_size, sequence_length)
.
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. bool
, optional) —
Whether or not to return the attentions tensors of all attention layers. See attentions
under returned
tensors for more detail. bool
, optional) —
Whether or not to return the hidden states of all layers. See hidden_states
under returned tensors for
more detail. bool
, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. The MusicgenMelodyModel 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.
( config: MusicgenMelodyDecoderConfig )
Parameters
The Musicgen Melody decoder model with a language modelling head on top.
The Musicgen Melody model was proposed in Simple and Controllable Music Generation by Jade Copet, Felix Kreuk, Itai Gat, Tal Remez, David Kant, Gabriel Synnaeve, Yossi Adi, Alexandre Défossez. It is a decoder-only transformer trained on the task of conditional music generation.
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.
( input_ids: LongTensor = None attention_mask: Optional = None encoder_hidden_states: Optional = None encoder_attention_mask: Optional = None head_mask: Optional = None past_key_values: Optional = None inputs_embeds: Optional = None use_cache: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None labels: Optional = None ) → transformers.models.musicgen_melody.modeling_musicgen_melody.MusicgenMelodyOutputWithPast
or tuple(torch.FloatTensor)
Parameters
torch.LongTensor
of shape (batch_size * num_codebooks, sequence_length)
) —
Indices of input sequence tokens in the vocabulary, corresponding to the sequence of audio codes.
Indices can be obtained by encoding an audio prompt with an audio encoder model to predict audio codes, such as with the EncodecModel. See EncodecModel.encode() for details.
The input_ids
will automatically be converted from shape (batch_size * num_codebooks, target_sequence_length)
to (batch_size, num_codebooks, target_sequence_length)
in the forward pass. If
you obtain audio codes from an audio encoding model, such as EncodecModel, ensure that the number of
frames is equal to 1, and that you reshape the audio codes from (frames, batch_size, num_codebooks, target_sequence_length)
to (batch_size * num_codebooks, target_sequence_length)
prior to passing them as
input_ids
.
torch.Tensor
of shape (batch_size, sequence_length)
, optional) —
Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]
:
torch.FloatTensor
of shape (batch_size, encoder_sequence_length, hidden_size)
, optional) —
Sequence of hidden-states representing the concatenation of the text encoder output and the processed audio encoder output.
Used as a conditional signal and will thus be concatenated to the projeted decoder_input_ids
. torch.LongTensor
of shape (batch_size, encoder_sequence_length)
, optional) —
Mask to avoid performing attention on conditional hidden states. Mask values
selected in [0, 1]
:
torch.Tensor
of shape (decoder_layers, decoder_attention_heads)
, optional) —
Mask to nullify selected heads of the attention modules. Mask values selected in [0, 1]
:
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.
If past_key_values
are used, the user can optionally input only the last decoder_input_ids
(those that
don’t have their past key value states given to this model) of shape (batch_size, 1)
instead of all
decoder_input_ids
of shape (batch_size, sequence_length)
.
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. bool
, optional) —
Whether or not to return the attentions tensors of all attention layers. See attentions
under returned
tensors for more detail. bool
, optional) —
Whether or not to return the hidden states of all layers. See hidden_states
under returned tensors for
more detail. bool
, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. torch.LongTensor
of shape (batch_size, sequence_length, num_codebooks)
, 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.models.musicgen_melody.modeling_musicgen_melody.MusicgenMelodyOutputWithPast
or tuple(torch.FloatTensor)
A transformers.models.musicgen_melody.modeling_musicgen_melody.MusicgenMelodyOutputWithPast
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 (MusicgenMelodyConfig) 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).
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.
encoder_hidden_states (torch.FloatTensor
of shape (batch_size, encoder_sequence_length, hidden_size)
, optional) — Sequence of conditional hidden-states representing the concatenation of the projeted text encoder output and the projeted audio encoder output.
Used as a conditional signal.
The MusicgenMelodyForCausalLM 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.
( config: MusicgenMelodyConfig = None text_encoder: Optional = None audio_encoder: Optional = None decoder: Optional = None )
Parameters
Optional[PreTrainedModel]
, optional) — Text encoder. Optional[PreTrainedModel]
, optional) — Audio code decoder. Optional[MusicgenMelodyForCausalLM]
, optional) — MusicGen Melody decoder used to generate audio codes. The composite Musicgen Melody model with a text and audio conditional models, a MusicgenMelody decoder and an audio encoder, for music generation tasks with one or both of text and audio prompts.
The Musicgen Melody model was proposed in Simple and Controllable Music Generation by Jade Copet, Felix Kreuk, Itai Gat, Tal Remez, David Kant, Gabriel Synnaeve, Yossi Adi, Alexandre Défossez. It is a decoder-only transformer trained on the task of conditional music generation.
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.
( input_ids: Optional = None attention_mask: Optional = None input_features: Optional = None decoder_input_ids: Optional = None decoder_attention_mask: Optional = None past_key_values: Tuple = None encoder_hidden_states: Optional = None inputs_embeds: Optional = None decoder_inputs_embeds: Optional = None labels: Optional = None use_cache: Optional = None output_attentions: Optional = None output_hidden_states: Optional = None return_dict: Optional = None **kwargs ) → transformers.models.musicgen_melody.modeling_musicgen_melody.MusicgenMelodyOutputWithPast
or tuple(torch.FloatTensor)
Parameters
torch.LongTensor
of shape (batch_size, sequence_length)
) —
Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
it.
Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.
torch.Tensor
of shape (batch_size, sequence_length)
, optional) —
Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]
:
torch.FloatTensor
of shape (batch_size, audio_sequence_length, num_chroma)
) —
Input audio features.
This should be returned by the MusicgenMelodyFeatureExtractor class that you can also
retrieve from AutoFeatureExtractor. See MusicgenMelodyFeatureExtractor.__call__()
for details. torch.LongTensor
of shape (batch_size * num_codebooks, target_sequence_length)
, optional) —
Indices of decoder input sequence tokens in the vocabulary, corresponding to the sequence of audio codes.
Indices can be obtained by encoding an audio prompt with an audio encoder model to predict audio codes, such as with the EncodecModel. See EncodecModel.encode() for details.
The decoder_input_ids
will automatically be converted from shape (batch_size * num_codebooks, target_sequence_length)
to (batch_size, num_codebooks, target_sequence_length)
in the forward pass. If
you obtain audio codes from an audio encoding model, such as EncodecModel, ensure that the number of
frames is equal to 1, and that you reshape the audio codes from (frames, batch_size, num_codebooks, target_sequence_length)
to (batch_size * num_codebooks, target_sequence_length)
prior to passing them as
decoder_input_ids
.
torch.LongTensor
of shape (batch_size, target_sequence_length)
, optional) —
Default behavior: generate a tensor that ignores pad tokens in decoder_input_ids
. Causal mask will also
be used by default. 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, encoder_sequence_length + 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.
If past_key_values
are used, the user can optionally input only the last decoder_input_ids
(those that
don’t have their past key value states given to this model) of shape (batch_size, 1)
instead of all
decoder_input_ids
of shape (batch_size, sequence_length)
.
torch.FloatTensor
of shape (batch_size, encoder_sequence_length, hidden_size)
, optional) —
Sequence of conditional hidden-states representing the concatenation of the projeted text encoder output and the projeted audio encoder output.
Used as a conditional signal and will thus be concatenated to the projeted decoder_input_ids
. 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. torch.FloatTensor
of shape (batch_size, target_sequence_length, hidden_size)
, optional) —
Optionally, instead of passing decoder_input_ids
you can choose to directly pass an embedded
representation. If past_key_values
is used, optionally only the last decoder_inputs_embeds
have to be
input (see past_key_values
). This is useful if you want more control over how to convert
decoder_input_ids
indices into associated vectors than the model’s internal embedding lookup matrix.
If decoder_input_ids
and decoder_inputs_embeds
are both unset, decoder_inputs_embeds
takes the value
of inputs_embeds
.
torch.LongTensor
of shape (batch_size, sequence_length, num_codebooks)
, 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]
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
). bool
, optional) —
Whether or not to return the attentions tensors of all attention layers. See attentions
under returned
tensors for more detail. bool
, optional) —
Whether or not to return the hidden states of all layers. See hidden_states
under returned tensors for
more detail. bool
, optional) —
Whether or not to return a ModelOutput instead of a plain tuple. Returns
transformers.models.musicgen_melody.modeling_musicgen_melody.MusicgenMelodyOutputWithPast
or tuple(torch.FloatTensor)
A transformers.models.musicgen_melody.modeling_musicgen_melody.MusicgenMelodyOutputWithPast
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 (MusicgenMelodyConfig) 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).
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.
encoder_hidden_states (torch.FloatTensor
of shape (batch_size, encoder_sequence_length, hidden_size)
, optional) — Sequence of conditional hidden-states representing the concatenation of the projeted text encoder output and the projeted audio encoder output.
Used as a conditional signal.
The MusicgenMelodyForConditionalGeneration 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:
>>> from transformers import AutoProcessor, MusicgenMelodyForConditionalGeneration
>>> import torch
>>> processor = AutoProcessor.from_pretrained("facebook/musicgen-melody")
>>> model = MusicgenMelodyForConditionalGeneration.from_pretrained("facebook/musicgen-melody")
>>> inputs = processor(
... text=["80s pop track with bassy drums and synth", "90s rock song with loud guitars and heavy drums"],
... padding=True,
... return_tensors="pt",
... )
>>> pad_token_id = model.generation_config.pad_token_id
>>> decoder_input_ids = (
... torch.ones((inputs.input_ids.shape[0] * model.decoder.num_codebooks, 1), dtype=torch.long)
... * pad_token_id
... )
>>> logits = model(**inputs, decoder_input_ids=decoder_input_ids).logits
>>> logits.shape # (bsz * num_codebooks, encoder_len + tgt_len, vocab_size)
torch.Size([8, 249, 2048])