Z j1 S SK rS SKrSSKJr SSKJr SSKJr SSK J
r
J r SSK J
r
\ R " \5 rS\S \S
\S \S \S
\R& 4 S jr\
" SS9 " S S\5 5 rS/rg) N ) hertz_to_mel)SequenceFeatureExtractor) BatchFeature)
TensorTypelogging)requires num_mel_binsnum_spectrogram_bins sample_ratelower_edge_hertzupper_edge_hertzreturnc b [ R nSnUS- n[ R " SXUS9US n [ U SS9SS2[ R 4 n
[ R " [ USS9[ USS9U S- US9n U SS [ R SS24 U SS
[ R SS24 U SS [ R SS24 pn X-
X-
- nX-
X-
- n[ R
" S[ R " UU5 5 n[ R " UUS /S S //5 R U5 $ ) z.NumPy-port of the JAX mel weight matrix logic. g @ )dtypeNkaldi) mel_scale r ) npfloat64linspacer newaxismaximumminimumpadastype)r
r r r
r r internal_dtype
bands_to_zero
nyquist_hertzlinear_frequenciesspectrogram_bins_meledgeslower_edge_mel
center_melupper_edge_mel lower_slopes upper_slopesmel_weights_matrixs ځ/root/GenerationalWealth/GenerationalWealth/venv/lib/python3.13/site-packages/transformers/models/lasr/feature_extraction_lasr.pylinear_to_mel_weight_matrixr. s< ZZN M #%MS-Ucderest'(:gNqRTR\R\}] KK%9%9q
E cr
2::q=!
a BJJM"
ab "**a- !/N )9j>YZL"9n>YZLCL,)OP
66$
q'9Aq6&B C J J5 QQ )torch)backendsc . ^ \ rS rSrSrSS/r SU 4S jjrSS jr SS\R \
\ - \
\R - \
\
\ - S \ S
\
S- S \\- S- S \ S- S
\S- S\
S- S\
S- S\ S- S\S- S\ S- S\4S jjrSrU =r$ )LasrFeatureExtractorB a
Constructs a LASR feature extractor.
This feature extractor inherits from [`~feature_extraction_sequence_utils.SequenceFeatureExtractor`] which contains
most of the main methods. Users should refer to this superclass for more information regarding those methods.
This class extracts mel-filter bank features from raw speech using a custom numpy implementation of the `Short Time
Fourier Transform` which should match pytorch's `torch.stft` equivalent.
Args:
feature_size (`int`, *optional*, defaults to 128):
The feature dimension of the extracted features.
sampling_rate (`int`, *optional*, defaults to 16000):
The sampling rate at which the audio files should be digitalized expressed in hertz (Hz).
hop_length (`int`, *optional*, defaults to 160):
Length of the overlapping windows for the STFT used to obtain the Mel Frequency coefficients.
n_fft (`int`, *optional*, defaults to 512):
Size of the Fourier transform.
win_length (`int`, *optional*, defaults to 400):
The window length for the STFT computation.
padding_value (`float`, *optional*, defaults to 0.0):
Padding value used to pad the audio. Should correspond to silences.
input_featuresattention_maskc > [ TU ] " SXUS.UD6 X0l X@l XPl [
R " [ UUS- S- USS[ R S95 U l
g )N) feature_size
sampling_rate
padding_valuer r g @_@g L@)r
r r r
r r ) super__init__
hop_lengthn_fft
win_lengthr0
from_numpyr. r r mel_filters) selfr8 r9 r> r? r@ r: kwargs __class__s r- r= LasrFeatureExtractor.__init__^ sf wl_lwpvw$
$ ++ ' )%*aZ!^ )!&!'jj
r/ c [ R " U R SU[ R S9nUR [ R 5 nUR SU R U R 5 n[ R R X4- U R S9n[ R " U5 S- nU R R U5 n[ R " Xg- SS9n[ R " U5 nU$ )NF)periodicdevicer r )nr gh㈵>)min)r0 hann_windowr@ r tounfoldr> fftrfftr? absrB clamplog) rC waveformrI windowframesstft
power_specrB mel_specs r- _torch_extract_fbank_features2LasrFeatureExtractor._torch_extract_fbank_featuresx s ""4??U6Y^YfYfg;;u}}-
T__dooFyy~~fo~<YYt_)
&&))&1 ;;z 7TB99X&r/ N
raw_speech
truncationpad_to_multiple_ofreturn_tensorsreturn_attention_maskpadding
max_lengthr9 do_normalizerI return_token_timestampsr c $ UbP XR :w a@ [ SU R R SU R SU R SU S3 5 eO-[ R SU R R S35 [
U[ R 5 ( a [ R " U5 nO[
U[ [ 45 ( ae [
US [ [ R 45 ( a' U V
s/ s H n
[ R " U
5 PM nn
O[ R " U5 n[
U[ R 5 =( a [ UR 5 S
: nU( aW [ UR 5 S : a> [ R S U R R S
35 UR! S5 n[
U[ [ 45 nU( ab U H\ n
[ U
R 5 S
: d M [ R S U R R S
35 U
R! S5 n
M^ U( d U( a6 U V
s/ s H( oSS2S4 R# [ R$ 5 PM* nn
O'USS2S4 R# [ R$ 5 /n[' SU05 nU R) UUUUUUSS9nUR* R- S5 nU R/ UU
5 nSUR# [ R$ 5 0nU( aO UR0 SS2U R2 S
-
SU R4 24 nUR# [ R6 5 US' [' UUS9$ s sn
f s sn
f )a
Main method to featurize and prepare for the model one or several sequence(s). Implementation uses PyTorch for
the STFT computation if available, otherwise a slower NumPy based one.
Args:
raw_speech (`np.ndarray`, `list[float]`, `list[np.ndarray]`, `list[list[float]]`):
The sequence or batch of sequences to be padded. Each sequence can be a numpy array, a list of float
values, a list of numpy arrays or a list of list of float values. Must be mono channel audio, not
stereo, i.e. single float per timestep.
truncation (`bool`, *optional*, default to `True`):
Activates truncation to cut input sequences longer than *max_length* to *max_length*.
pad_to_multiple_of (`int`, *optional*, defaults to None):
If set will pad the sequence to a multiple of the provided value.
This is especially useful to enable the use of Tensor Cores on NVIDIA hardware with compute capability
`>= 7.5` (Volta), or on TPUs which benefit from having sequence lengths be a multiple of 128.
return_attention_mask (`bool`, *optional*):
Whether to return the attention mask. If left to the default, will return the attention mask according
to the specific feature_extractor's default.
[What are attention masks?](../glossary#attention-mask)
For Parakeet models, `attention_mask` should always be passed for batched inference, to avoid subtle
bugs.
return_tensors (`str` or [`~utils.TensorType`], *optional*):
If set, will return tensors instead of list of python integers. Acceptable values are:
- `'tf'`: Return TensorFlow `tf.constant` objects.
- `'pt'`: Return PyTorch `torch.Tensor` objects.
- `'np'`: Return Numpy `np.ndarray` objects.
sampling_rate (`int`, *optional*):
The sampling rate at which the `raw_speech` input was sampled. It is strongly recommended to pass
`sampling_rate` at the forward call to prevent silent errors and allow automatic speech recognition
pipeline.
padding_value (`float`, *optional*, defaults to 0.0):
The value that is used to fill the padding values / vectors.
do_normalize (`bool`, *optional*, defaults to `False`):
Whether or not to zero-mean unit-variance normalize the input. Normalizing can help to significantly
improve the performance of the model.
device (`str`, *optional*, defaults to `'cpu'`):
Specifies the device for computation of the log-mel spectrogram of audio signals in the
`_torch_extract_fbank_features` method. (e.g., "cpu", "cuda")
return_token_timestamps (`bool`, *optional*, defaults to `None`):
Deprecated. Use `return_attention_mask` instead from which the number of frames can be inferred.
Whether or not to return the number of frames of the input raw_speech.
These num_frames can be used by the model to compute word level timestamps.
Nz3The model corresponding to this feature extractor: z& was trained using a sampling rate of zI. Please make sure that the provided `raw_speech` input was sampled with z and not .zDIt is strongly recommended to pass the `sampling_rate` argument to `zN()`. Failing to do so can result in silent errors that might be hard to debug.r r r z2Only mono-channel audio is supported for input to z;. We will take the mean of the channels to convert to mono.r r5 pt)ra rb r] r^ r` r_ r6 )data tensor_type) r9
ValueErrorrE __name__loggerwarning
isinstancer ndarrayr0 tensorlisttupleTensorlenshapemeanrM float32r r r5 squeezerZ r6 r@ r> bool)rC r\ r] r^ r_ r` ra rb r9 rc rI rd rD speechis_batched_torchis_batched_sequencebatched_speech
padded_inputsr5 rh r6 s r- __call__ LasrFeatureExtractor.__call__ s, H $ 2 22 I$..JaJaIb c))-););(< =))-););(
! * 9 9!T__q=P=cTXTcTc=c:c dN%3%6%6uzz%BD! ">BB_ M. Vs > N /N
)r> rB r? r@ ) i> i i r )cpu)
FNNNlongestNNNr N)rk
__module__ __qualname____firstlineno____doc__model_input_namesr= rZ r ro rq floatry intstrr r r __static_attributes__
__classcell__)rE s @r- r3 r3 B sC 0 *+;<
4, !)-26-1 '!%$($( "/3ECJJe,tBJJ/??$tE{BSSEC EC $J EC
j(4/ EC $d{
EC tEC $JEC TzEC TkEC d
EC "&EC
EC ECr/ r3 )numpyr r0 audio_utilsr !feature_extraction_sequence_utilsr feature_extraction_utilsr utilsr r utils.import_utilsr
get_loggerrk rl r r ro r. r3 __all__r; r/ r- r s ' I 4 ( *
H %"R"R "R "R "R
"R ZZ"RJ
: MC3 MC MC` "
"r/