259 lines
11 KiB
Python
259 lines
11 KiB
Python
|
# coding=utf-8
|
||
|
|
# Copyright 2023 Meta AI and The HuggingFace Inc. team. All rights reserved.
|
||
|
|
#
|
||
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
||
|
|
# you may not use this file except in compliance with the License.
|
||
|
|
# You may obtain a copy of the License at
|
||
|
|
#
|
||
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
||
|
|
#
|
||
|
|
# Unless required by applicable law or agreed to in writing, software
|
||
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
||
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||
|
|
# See the License for the specific language governing permissions and
|
||
|
|
# limitations under the License.
|
||
|
|
""" MusicGen model configuration"""
|
||
|
|
|
||
|
|
from ...configuration_utils import PretrainedConfig
|
||
|
|
from ...utils import logging
|
||
|
|
from ..auto.configuration_auto import AutoConfig
|
||
|
|
|
||
|
|
|
||
|
|
logger = logging.get_logger(__name__)
|
||
|
|
|
||
|
|
|
||
|
|
from ..deprecated._archive_maps import MUSICGEN_PRETRAINED_CONFIG_ARCHIVE_MAP # noqa: F401, E402
|
||
|
|
|
||
|
|
|
||
|
|
class MusicgenDecoderConfig(PretrainedConfig):
|
||
|
|
r"""
|
||
|
|
This is the configuration class to store the configuration of an [`MusicgenDecoder`]. It is used to instantiate a
|
||
|
|
MusicGen 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
|
||
|
|
[facebook/musicgen-small](https://huggingface.co/facebook/musicgen-small) architecture.
|
||
|
|
|
||
|
|
Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
|
||
|
|
documentation from [`PretrainedConfig`] for more information.
|
||
|
|
|
||
|
|
|
||
|
|
Args:
|
||
|
|
vocab_size (`int`, *optional*, defaults to 2048):
|
||
|
|
Vocabulary size of the MusicgenDecoder model. Defines the number of different tokens that can be
|
||
|
|
represented by the `inputs_ids` passed when calling [`MusicgenDecoder`].
|
||
|
|
hidden_size (`int`, *optional*, defaults to 1024):
|
||
|
|
Dimensionality of the layers and the pooler layer.
|
||
|
|
num_hidden_layers (`int`, *optional*, defaults to 24):
|
||
|
|
Number of decoder layers.
|
||
|
|
num_attention_heads (`int`, *optional*, defaults to 16):
|
||
|
|
Number of attention heads for each attention layer in the Transformer block.
|
||
|
|
ffn_dim (`int`, *optional*, defaults to 4096):
|
||
|
|
Dimensionality of the "intermediate" (often named feed-forward) layer in the Transformer block.
|
||
|
|
activation_function (`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.
|
||
|
|
dropout (`float`, *optional*, defaults to 0.1):
|
||
|
|
The dropout probability for all fully connected layers in the embeddings, text_encoder, and pooler.
|
||
|
|
attention_dropout (`float`, *optional*, defaults to 0.0):
|
||
|
|
The dropout ratio for the attention probabilities.
|
||
|
|
activation_dropout (`float`, *optional*, defaults to 0.0):
|
||
|
|
The dropout ratio for activations inside the fully connected layer.
|
||
|
|
max_position_embeddings (`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).
|
||
|
|
initializer_factor (`float`, *optional*, defaults to 0.02):
|
||
|
|
The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
|
||
|
|
layerdrop (`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.
|
||
|
|
scale_embedding (`bool`, *optional*, defaults to `False`):
|
||
|
|
Scale embeddings by diving by sqrt(hidden_size).
|
||
|
|
use_cache (`bool`, *optional*, defaults to `True`):
|
||
|
|
Whether the model should return the last key/values attentions (not used by all models)
|
||
|
|
num_codebooks (`int`, *optional*, defaults to 4):
|
||
|
|
The number of parallel codebooks forwarded to the model.
|
||
|
|
tie_word_embeddings(`bool`, *optional*, defaults to `False`):
|
||
|
|
Whether input and output word embeddings should be tied.
|
||
|
|
audio_channels (`int`, *optional*, defaults to 1
|
||
|
|
Number of channels in the audio data. Either 1 for mono or 2 for stereo. Stereo models generate a separate
|
||
|
|
audio stream for the left/right output channels. Mono models generate a single audio stream output.
|
||
|
|
"""
|
||
|
|
|
||
|
|
model_type = "musicgen_decoder"
|
||
|
|
keys_to_ignore_at_inference = ["past_key_values"]
|
||
|
|
|
||
|
|
def __init__(
|
||
|
|
self,
|
||
|
|
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,
|
||
|
|
):
|
||
|
|
self.vocab_size = vocab_size
|
||
|
|
self.max_position_embeddings = max_position_embeddings
|
||
|
|
self.hidden_size = hidden_size
|
||
|
|
self.ffn_dim = ffn_dim
|
||
|
|
self.num_hidden_layers = num_hidden_layers
|
||
|
|
self.num_attention_heads = num_attention_heads
|
||
|
|
self.dropout = dropout
|
||
|
|
self.attention_dropout = attention_dropout
|
||
|
|
self.activation_dropout = activation_dropout
|
||
|
|
self.activation_function = activation_function
|
||
|
|
self.initializer_factor = initializer_factor
|
||
|
|
self.layerdrop = layerdrop
|
||
|
|
self.use_cache = use_cache
|
||
|
|
self.scale_embedding = scale_embedding # scale factor will be sqrt(d_model) if True
|
||
|
|
self.num_codebooks = num_codebooks
|
||
|
|
|
||
|
|
if audio_channels not in [1, 2]:
|
||
|
|
raise ValueError(f"Expected 1 (mono) or 2 (stereo) audio channels, got {audio_channels} channels.")
|
||
|
|
self.audio_channels = audio_channels
|
||
|
|
|
||
|
|
super().__init__(
|
||
|
|
pad_token_id=pad_token_id,
|
||
|
|
bos_token_id=bos_token_id,
|
||
|
|
eos_token_id=eos_token_id,
|
||
|
|
tie_word_embeddings=tie_word_embeddings,
|
||
|
|
**kwargs,
|
||
|
|
)
|
||
|
|
|
||
|
|
|
||
|
|
class MusicgenConfig(PretrainedConfig):
|
||
|
|
r"""
|
||
|
|
This is the configuration class to store the configuration of a [`MusicgenModel`]. It is used to instantiate a
|
||
|
|
MusicGen model according to the specified arguments, defining the text encoder, audio encoder and MusicGen decoder
|
||
|
|
configs.
|
||
|
|
|
||
|
|
Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
|
||
|
|
documentation from [`PretrainedConfig`] for more information.
|
||
|
|
|
||
|
|
Args:
|
||
|
|
kwargs (*optional*):
|
||
|
|
Dictionary of keyword arguments. Notably:
|
||
|
|
|
||
|
|
- **text_encoder** ([`PretrainedConfig`], *optional*) -- An instance of a configuration object that
|
||
|
|
defines the text encoder config.
|
||
|
|
- **audio_encoder** ([`PretrainedConfig`], *optional*) -- An instance of a configuration object that
|
||
|
|
defines the audio encoder config.
|
||
|
|
- **decoder** ([`PretrainedConfig`], *optional*) -- An instance of a configuration object that defines
|
||
|
|
the decoder config.
|
||
|
|
|
||
|
|
Example:
|
||
|
|
|
||
|
|
```python
|
||
|
|
>>> from transformers import (
|
||
|
|
... MusicgenConfig,
|
||
|
|
... MusicgenDecoderConfig,
|
||
|
|
... T5Config,
|
||
|
|
... EncodecConfig,
|
||
|
|
... MusicgenForConditionalGeneration,
|
||
|
|
... )
|
||
|
|
|
||
|
|
>>> # Initializing text encoder, audio encoder, and decoder model configurations
|
||
|
|
>>> text_encoder_config = T5Config()
|
||
|
|
>>> audio_encoder_config = EncodecConfig()
|
||
|
|
>>> decoder_config = MusicgenDecoderConfig()
|
||
|
|
|
||
|
|
>>> configuration = MusicgenConfig.from_sub_models_config(
|
||
|
|
... text_encoder_config, audio_encoder_config, decoder_config
|
||
|
|
... )
|
||
|
|
|
||
|
|
>>> # Initializing a MusicgenForConditionalGeneration (with random weights) from the facebook/musicgen-small style configuration
|
||
|
|
>>> model = MusicgenForConditionalGeneration(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-model")
|
||
|
|
|
||
|
|
>>> # loading model and config from pretrained folder
|
||
|
|
>>> musicgen_config = MusicgenConfig.from_pretrained("musicgen-model")
|
||
|
|
>>> model = MusicgenForConditionalGeneration.from_pretrained("musicgen-model", config=musicgen_config)
|
||
|
|
```"""
|
||
|
|
|
||
|
|
model_type = "musicgen"
|
||
|
|
is_composition = True
|
||
|
|
|
||
|
|
def __init__(self, **kwargs):
|
||
|
|
super().__init__(**kwargs)
|
||
|
|
if "text_encoder" not in kwargs or "audio_encoder" not in kwargs or "decoder" not in kwargs:
|
||
|
|
raise ValueError("Config has to be initialized with text_encoder, audio_encoder and decoder config")
|
||
|
|
|
||
|
|
text_encoder_config = kwargs.pop("text_encoder")
|
||
|
|
text_encoder_model_type = text_encoder_config.pop("model_type")
|
||
|
|
|
||
|
|
audio_encoder_config = kwargs.pop("audio_encoder")
|
||
|
|
audio_encoder_model_type = audio_encoder_config.pop("model_type")
|
||
|
|
|
||
|
|
decoder_config = kwargs.pop("decoder")
|
||
|
|
|
||
|
|
self.text_encoder = AutoConfig.for_model(text_encoder_model_type, **text_encoder_config)
|
||
|
|
self.audio_encoder = AutoConfig.for_model(audio_encoder_model_type, **audio_encoder_config)
|
||
|
|
self.decoder = MusicgenDecoderConfig(**decoder_config)
|
||
|
|
self.is_encoder_decoder = True
|
||
|
|
|
||
|
|
@classmethod
|
||
|
|
def from_sub_models_config(
|
||
|
|
cls,
|
||
|
|
text_encoder_config: PretrainedConfig,
|
||
|
|
audio_encoder_config: PretrainedConfig,
|
||
|
|
decoder_config: MusicgenDecoderConfig,
|
||
|
|
**kwargs,
|
||
|
|
):
|
||
|
|
r"""
|
||
|
|
Instantiate a [`MusicgenConfig`] (or a derived class) from text encoder, audio encoder and decoder
|
||
|
|
configurations.
|
||
|
|
|
||
|
|
Returns:
|
||
|
|
[`MusicgenConfig`]: An instance of a configuration object
|
||
|
|
"""
|
||
|
|
|
||
|
|
return cls(
|
||
|
|
text_encoder=text_encoder_config.to_dict(),
|
||
|
|
audio_encoder=audio_encoder_config.to_dict(),
|
||
|
|
decoder=decoder_config.to_dict(),
|
||
|
|
**kwargs,
|
||
|
|
)
|
||
|
|
|
||
|
|
@property
|
||
|
|
# This is a property because you might want to change the codec model on the fly
|
||
|
|
def sampling_rate(self):
|
||
|
|
return self.audio_encoder.sampling_rate
|
||
|
|
|
||
|
|
@property
|
||
|
|
def _attn_implementation(self):
|
||
|
|
# This property is made private for now (as it cannot be changed and a PreTrainedModel.use_attn_implementation method needs to be implemented.)
|
||
|
|
if hasattr(self, "_attn_implementation_internal"):
|
||
|
|
if self._attn_implementation_internal is None:
|
||
|
|
# `config.attn_implementation` should never be None, for backward compatibility.
|
||
|
|
return "eager"
|
||
|
|
else:
|
||
|
|
return self._attn_implementation_internal
|
||
|
|
else:
|
||
|
|
return "eager"
|
||
|
|
|
||
|
|
@_attn_implementation.setter
|
||
|
|
def _attn_implementation(self, value):
|
||
|
|
self._attn_implementation_internal = value
|
||
|
|
self.decoder._attn_implementation = value
|