0;ji,(dZddlZddlmZmZddlZddlmcmZ ddl Z Gddejj Z Gdde ZGdd e ZGd d e ZGd d ejj ZGddejj ZdS)zPositonal Encoding Module.N)TupleUnionc eZdZdZ ddedededeffd Z dd ej d e eej fd e ej ej ffd Z dd e eej fdeded ej fdZ xZS)PositionalEncodingzPositional encoding. :param int d_model: embedding dim :param float dropout_rate: dropout rate :param int max_len: maximum input length PE(pos, 2i) = sin(pos/(10000^(2i/dmodel))) PE(pos, 2i+1) = cos(pos/(10000^(2i/dmodel))) Fd_model dropout_ratemax_lenreversec(t||_tj|j|_t j||_ ||_ t j |j |j|_ t j d|j t jd}t jt j d|jdt jtjd|jz z}t j||z|j dddddf<t j||z|j dddddf<|j d|_ dS)'Construct an PositionalEncoding object.prdtype@N)super__init__rmathsqrtxscaletorchnnDropoutdropoutr zerospearangefloat32 unsqueezeexplogsincos)selfrr r r positiondiv_term __class__s g/root/voice-cloning/.venv/lib/python3.11/site-packages/chatterbox/models/s3gen/transformer/embedding.pyrzPositionalEncoding.__init__%sR  i -- x'','77  +dlDL99<4<&+m5555>Yq\\ 9 LDL!5= A A Ahw$,. / 011!9X%89914a4 9X%89914a4'##A&&rxoffsetreturnc |j|j|_|||dd}||jz|z}||||fS)aNAdd positional encoding. Args: x (torch.Tensor): Input. Its shape is (batch, time, ...) offset (int, torch.tensor): position offset Returns: torch.Tensor: Encoded tensor. Its shape is (batch, time, ...) torch.Tensor: for compatibility to RelPositionalEncoding rF)rtodeviceposition_encodingsizerrr'r-r.pos_embs r+forwardzPositionalEncoding.forward;sm'**QX&&((EBB  Og %||A W 5 555r,Tr4 apply_dropoutc~t|tr'||z|jksJ|jdd|||zf}nt|tjr?|dkr'||z|jksJ|jdd|||zf}nt j||z|jksJ|dt j d| |j z}|dk}||z}tj ||jd}|r||}|S) For getting encoding in a streaming fashion Attention!!!!! we apply dropout only once at the whole utterance level in a none streaming way, but will call this function several times with increasing input size in a streaming scenario, so the dropout will be applied several times. Args: offset (int or torch.tensor): start offset size (int): required size of position encoding Returns: torch.Tensor: Corresponding encoding Nrr) isinstanceintr rrTensordimmaxr"r r1r2F embeddingr)r'r.r4r8r6indexflags r+r3z$PositionalEncoding.position_encodingOsJ* fc " " 5D=DL0000gaaa !556GG  - - 5&**,,!2C2CD=DL0000gaaa !556GG9V$$t+t|;;;;$$Q'' Q%%((778E19DDLEk%44G  ,ll7++Gr,)rFr)T)__name__ __module__ __qualname____doc__r<floatboolrrr=rrr7r3 __classcell__r*s@r+rrs!%!& '''$'' ''''''04566<6c5</06U\5</06666.15&&"'U\(9":& #&*.&:?&&&&&&&&r,rc eZdZdZd dededeffd Z d dejd e eejfd e ejejffd Z xZ S)RelPositionalEncodingzRelative positional encoding module. See : Appendix B in https://arxiv.org/abs/1901.02860 Args: d_model (int): Embedding dimension. dropout_rate (float): Dropout rate. max_len (int): Maximum input length. rrr r cRt|||ddS)zInitialize class.T)r N)rrr'rr r r*s r+rzRelPositionalEncoding.__init__s* ,FFFFFr,rr-r.r/c|j|j|_||jz}|||dd}||||fS)aCompute positional encoding. Args: x (torch.Tensor): Input tensor (batch, time, `*`). Returns: torch.Tensor: Encoded tensor (batch, time, `*`). torch.Tensor: Positional embedding tensor (1, time, `*`). rF)rr1r2rr3r4rr5s r+r7zRelPositionalEncoding.forwardsh'**QX&&  O((EBB||A W 5 555r,rrD) rErFrGrHr<rIrrr=rrr7rKrLs@r+rNrNxsGGG5G3GGGGGG 4566<6c5</06U\5</066666666r,rNc2eZdZdZddededeffd ZxZS)WhisperPositionalEncodingz@ Sinusoids position encoding used in openai-whisper.encoder rr r cXt|||d|_tjd|dzdz z }t j| t j|dzz}t j|ddtjf|tjddfz}t j t j |t j |gd}t|d| d|ddS)N?i'rrr>rr)rrrnpr$rr#r newaxiscatr%r&delattrregister_bufferr") r'rr r log_timescale_incrementinv_timescales scaled_timerr*s r+rz"WhisperPositionalEncoding.__init__s ,888 "$&--7a??l7++AAArzM: 2:qqq= )* Y +.. +0F0FGQ O O Od T2<<??33333r,)rUrErFrGrHr<rIrrKrLs@r+rTrTsa 4 4 45 43 4 4 4 4 4 4 4 4 4 4r,rTc2eZdZdZddededeffd ZxZS)LearnablePositionalEncodingz@ Learnable position encoding used in openai-whisper.decoder rr r ct|||tjtjd|||_d|_dS)NrrW)rrrr ParameteremptyrrrPs r+rz$LearnablePositionalEncoding.__init__sO ,888($$U[GW%E%EFF r,)rdrarLs@r+rcrcsa53r,rcc eZdZdZdedeffd Z d dejde eejfde ejejffd Z de eejfd edejfd Z xZ S) NoPositionalEncodingz No position encoding rr ct||_tj||_dS)Nr)rrrrrrr)r'rr r*s r+rzNoPositionalEncoding.__init__s>  x'','77 r,rr-r.r/ctjd|d|j|j}|||fS)z= Just return zero vector for interface compatibility r)rrr4rr1r2rr5s r+r7zNoPositionalEncoding.forwardsH +aDL99<rs !  [[[[[[[[|66666.666<44444 2444""42222258?222.]]]]]%(/]]]]]r,