Z j3 S SK JrJrJr SSKJr SSKJr SSKJ r SSK
J r J r SSK
Jr \ " 5 ( a S S KrSS
KJr SS KJr S r " S
S\SS9r " S S\5 rg ) )Any TypedDictoverload )
AudioInput)GenerationConfig)is_torch_available)ChatChatType )PipelineN)%MODEL_FOR_TEXT_TO_SPECTROGRAM_MAPPING)SpeechT5HifiGanzmicrosoft/speecht5_hifiganc . \ rS rSr% Sr\\S' \\S' Srg) AudioOutput! z
audio (`AudioInput`):
The generated audio waveform.
sampling_rate (`int`):
The sampling rate of the generated audio waveform.
audio
sampling_rate N) __name__
__module__ __qualname____firstlineno____doc__r __annotations__int__static_attributes__r u/root/GenerationalWealth/GenerationalWealth/venv/lib/python3.13/site-packages/transformers/pipelines/text_to_audio.pyr r ! s r r F)totalc ^ \ rS rSrSrSrSrSrSrSr \
" SS9r SSS.U 4S jjr S
r
S r\S \S
\S\4S j5 r\S \\ S
\S\\ 4S j5 r\S \S
\S\4S j5 r\S \\ S
\S\\ 4S j5 rU 4S jr SS jrS rSrU =r$ )TextToAudioPipeline- al
Text-to-audio generation pipeline using any `AutoModelForTextToWaveform` or `AutoModelForTextToSpectrogram`. This
pipeline generates an audio file from an input text and optional other conditional inputs.
Unless the model you're using explicitly sets these generation parameters in its configuration files
(`generation_config.json`), the following default values will be used:
- max_new_tokens: 256
Example:
```python
>>> from transformers import pipeline
>>> pipe = pipeline(model="suno/bark-small")
>>> output = pipe("Hey it's HuggingFace on the phone!")
>>> audio = output["audio"]
>>> sampling_rate = output["sampling_rate"]
```
Learn more about the basics of using a pipeline in the [pipeline tutorial](../pipeline_tutorial)
You can specify parameters passed to the model by using [`TextToAudioPipeline.__call__.forward_params`] or
[`TextToAudioPipeline.__call__.generate_kwargs`].
Example:
```python
>>> from transformers import pipeline
>>> music_generator = pipeline(task="text-to-audio", model="facebook/musicgen-small")
>>> # diversify the music generation by adding randomness with a high temperature and set a maximum music length
>>> generate_kwargs = {
... "do_sample": True,
... "temperature": 0.7,
... "max_new_tokens": 35,
... }
>>> outputs = music_generator("Techno music with high melodic riffs", generate_kwargs=generate_kwargs)
```
This pipeline can currently be loaded from [`pipeline`] using the following task identifiers: `"text-to-speech"` or
`"text-to-audio"`.
See the list of available models on [huggingface.co/models](https://huggingface.co/models?filter=text-to-speech).
TNF )max_new_tokens)vocoderr c r > [ T
U ] " U0 UD6 S U l U R R [
R " 5 ; aG Uc= [ R " [ 5 R U R R 5 OUU l U R R R S; a S U l X l U R b% U R R R U l U R c U R R nU R R R# SS 5 nUbG UR% UR' 5 R) 5 VVs0 s H u pxUc M
Xx_M snn5 S HL n [+ XYS 5 nUb X l M [+ USS 5 c M* [+ UR, U S 5 nUc MF X l MN U R cP U R bB [/ U R S5 ( a& U R R0 R U l g g g g s snnf )N)musicgenspeecht5generation_config) sample_rater codec_configfeature_extractor)super__init__r& model __class__r valuesr from_pretrainedDEFAULT_VOCODER_IDtodeviceconfig
model_type processorr __dict__getupdateto_dictitemsgetattrr, hasattrr- ) selfr& r argskwargsr7
gen_configkvsampling_rate_namer1 s r r/ TextToAudioPipeline.__init__m s
$)&) :: #H#O#O#Q Q ? //0BCFFtzzGXGXY
L :: ' '+C C !DN * << #!% !4!4!B!BD % ZZ&&F ,,001DdKJ%
0B0B0D0J0J0L ^0LPQtqt0L ^_&F" 'D I
,)6& V^T:F$+F,?,?ASUY$ZM$0-:* 'G %$..*DQUQ_Q_atIuIu!% !A!A!O!OD Jv*D % _s ( H3
5H3
c x [ U[ 5 ( a U/nU R R R S:X aX Sn[ U R S5 ( a! [ U R R SS5 nUSSSS.nUR U5 UnU R b U R O U R n[ U[ 5 ( a" UR " UR 4SSS.UD6nU$ U R R R S :X a> U Vs/ s H owR S
5 ( d S U 3OUPM! nnUR! S S5 U R R R S
:X a, U Vs/ s H owR S
5 ( d SU 3OUPM! nnU" U40 UDSS0D6nU$ s snf s snf )Nbarkr$ semantic_configmax_input_semantic_lengthFT)
max_lengthadd_special_tokensreturn_attention_maskreturn_token_type_ids)tokenize return_dictcsm[z[0]rN diaz[S1] return_tensorspt)
isinstancestrr0 r7 r8 r@ r* r? rK r< r9 tokenizerr
apply_chat_templatemessages
startswith
setdefault)rA textrC rM
new_kwargs preprocessoroutputts r
preprocess TextToAudioPipeline.preprocess s dC 6D :: ' '6 1 Jt--/@AA $T%;%;%K%KMhjm n
(&+)-). J
f % F)- )Ct~~ dD ! !!55
F
zz ++u4KOP4a c):):#aS A4P!!"6=zz ++u4MQRT<<+<+<%s !CTR!$F&FFF
Q Ss &F2;&F7c U R X R S9nUS nUS nU R R 5 ( a U R X@R S9nSU; a U R US' UR U5 UR SS05 U R R R S; a SU; a SUS' U R R " S 0 UDUD6nOC[ U5 ( a [ S UR 5 35 eU R " S 0 UDUD6S
nU R b U R U5 nU$ ) N)r6 forward_paramsgenerate_kwargsr* return_dict_in_generateT)rS output_audiozYou're using the `TextToAudioPipeline` with a forward-only model, but `generate_kwargs` is non empty. For forward-only TTA models, please use `forward_params` instead of `generate_kwargs`. For reference, the `generate_kwargs` used here are: r r )
_ensure_tensor_on_devicer6 r0 can_generater* r< r7 r8 generatelen
ValueErrorkeysr& )rA model_inputsrC rg rh rb s r _forward TextToAudioPipeline._forward sQ ..vkk.J 01 !23 :: " " $ $ " ; ;OT_T_ ; `O #/97;7M7M 34
! !/ 2
! !#2ZZ((J<J>JF?## KKZK_K_KaJbd
ZZA,A.A!DF << #\\&)F
r text_inputsrg returnc g Nr rA rt rg s r __call__ TextToAudioPipeline.__call__ s PSr c g rw r rx s r ry rz s \_r c g rw r rx s r ry rz s UXr c g rw r rx s r ry rz s adr c & > [ TU ] " U40 UD6$ )a
Generates speech/audio from the inputs. See the [`TextToAudioPipeline`] documentation for more information.
Args:
text_inputs (`str`, `list[str]`, `ChatType`, or `list[ChatType]`):
One or several texts to generate. If strings or a list of string are passed, this pipeline will
generate the corresponding text. Alternatively, a "chat", in the form of a list of dicts with "role"
and "content" keys, can be passed, or a list of such chats. When chats are passed, the model's chat
template will be used to format them before passing them to the model.
forward_params (`dict`, *optional*):
Parameters passed to the model generation/forward method. `forward_params` are always passed to the
underlying model.
generate_kwargs (`dict`, *optional*):
The dictionary of ad-hoc parametrization of `generate_config` to be used for the generation call. For a
complete overview of generate, check the [following
guide](https://huggingface.co/docs/transformers/en/main_classes/text_generation). `generate_kwargs` are
only passed to the underlying model if the latter is a generative model.
Return:
`AudioOutput` or a list of `AudioOutput`, which is a `TypedDict` with two keys:
- **audio** (`np.ndarray` of shape `(nb_channels, audio_length)`) -- The generated audio waveform.
- **sampling_rate** (`int`) -- The sampling rate of the generated audio waveform.
)r. ry )rA rt rg r1 s r ry rz s 2 w >~>>r c [ U SS 5 b U R US' [ U SS 5 b U R US' U R US' U( a UO0 U( a UO0 S.nUc 0 n0 nXU4$ )Nassistant_modelassistant_tokenizerrZ )rg rh )r? r rZ r )rA preprocess_paramsrg rh paramspostprocess_paramss r _sanitize_parameters(TextToAudioPipeline._sanitize_parameters s 4*D 1 =151E1EO - . 4. 5 A+/>>OK (595M5MO 1 2 1?nB2Ar
$ " *<< nn[ U5 S: a UOUS nO:UR
S[ R S9R 5 R 5 n[ UU R S 9$ s snf )
NFr T sequencesr cpu)r6 dtyper )r r )rX dicttupler9 decodelistr5 torchfloatnumpysqueezern r r )rA r needs_decodingels r postprocess TextToAudioPipeline.postprocess% s eT " "% g!% k*
u
%
%!HE dnn 8NN))%0E eT " "X]^X]RTUU%u{{U;AACKKMX]E^ Z!^EqEHHEH=CCEMMOE ,,
_s AD8)r9 r r& )NNN)r r r r r _pipeline_calls_generate_load_processor_load_image_processor_load_feature_extractor_load_tokenizerr _default_generation_configr/ rd rr r rY r r ry r r r r r
__classcell__)r1 s @r r" r" - s 2h $ O ! #O "2" '+$ &P &PP&P(T SCS3S;S S
_DI__kIZ_ _
XHXX X X
dDNdcddS^N_d d?: =.
r r" )typingr r r audio_utilsr
generationr utilsr utils.chat_template_utilsr
r baser
r models.auto.modeling_autor !models.speecht5.modeling_speecht5r r4 r r" r r r r sR , + $ ) & 6 QC1 )5 O
( O
r