0;jiddlZ dejdfdedededejdejf d Z dd ejd ejd edededededefdZddejdedejfdZdS)Ncpusize chunk_sizenum_left_chunksdevicereturnctj||}tj||ddz|z}|d|dk}|S)a7Create mask for subsequent steps (size, size) with chunk size, this is for streaming encoder Args: size (int): size of mask chunk_size (int): size of chunk num_left_chunks (int): number of left chunks <0: use full chunk >=0: use num_left_chunks device (torch.device): "cpu" or "cuda" or torch.Tensor.device Returns: torch.Tensor: mask Examples: >>> subsequent_chunk_mask(4, 2) [[1, 1, 0, 0], [1, 1, 0, 0], [1, 1, 1, 1], [1, 1, 1, 1]] )rtrunc) rounding_moder)torcharangediv unsqueeze)rrrrpos_idx block_valuerets \/root/voice-cloning/.venv/lib/python3.11/site-packages/chatterbox/models/s3gen/utils/mask.pysubsequent_chunk_maskr6sd:l4///G9WjHHH1LPZZK   A  !6!6q!9!9 9C JTxsmasksuse_dynamic_chunkuse_dynamic_left_chunkdecoding_chunk_sizestatic_chunk_sizenum_decoding_left_chunksenable_full_contextc|r|d}|dkr|} d} n}|dkr|} |} nrtjd|d} d} | |dzkr|r|} n:| dzdz} |r0|dz | z} tjd| d} t |d| | |j} | d} || z} nO|dkrG|} t |d|| |j} | d} || z} n|} | jtjksJ| ddk dkr7t dd | | ddk<| S) a Apply optional mask for encoder. Args: xs (torch.Tensor): padded input, (B, L, D), L for max length mask (torch.Tensor): mask for xs, (B, 1, L) use_dynamic_chunk (bool): whether to use dynamic chunk or not use_dynamic_left_chunk (bool): whether to use dynamic left chunk for training. decoding_chunk_size (int): decoding chunk size for dynamic chunk, it's 0: default for training, use random dynamic chunk. <0: for decoding, use full chunk. >0: for decoding, use fixed chunk size as set. static_chunk_size (int): chunk size for static chunk training/decoding if it's greater than 0, if use_dynamic_chunk is true, this parameter will be ignored num_decoding_left_chunks: number of left chunks, this is for decoding, the chunk size is decoding_chunk_size. >=0: use num_decoding_left_chunks <0: use all left chunks enable_full_context (bool): True: chunk size is either [1, 25] or full context(max_len) False: chunk size ~ U[1, 25] Returns: torch.Tensor: chunk mask of the input xs. r rr)r )dimzoget chunk_masks all false at some timestep, force set to true, make sure they are masked in futuer computation!T) rrrandintitemrrrdtypeboolsumloggingwarning) rrrrrrrrmax_lenrrmax_left_chunks chunk_maskss radd_optional_chunk_maskr.YsF#''!**  " " J OO 1 $ $,J6OO q'599>>@@J OGqL((-@($ '"_q0 )B'.{z&AO&+mA49';';;?466$+BGGAJJ ,;,.I77 "++A.. k) Q  2+BGGAJJ8I,;,.I77 "++A.. k)    * * * *B1$))++0022a77J K K K26 KOOO++Q./ rlengthsr+c|}|d}|dkr|n%|}t jd|tj|j}|d ||}|d}||k}|S)aMake mask tensor containing indices of padded part. See description of make_non_pad_mask. Args: lengths (torch.Tensor): Batch of lengths (B,). Returns: torch.Tensor: Mask tensor containing indices of padded part. Examples: >>> lengths = [5, 3, 2] >>> make_pad_mask(lengths) masks = [[0, 0, 0, 0 ,0], [0, 0, 0, 1, 1], [0, 0, 1, 1, 1]] r)r&rr) longrmaxr%rrint64rrexpand)r/r+ batch_size seq_rangeseq_range_expandseq_length_expandmasks r make_pad_maskr:s"llnnGaJ 1gg'++--*<*<*>*>G Q$#(;$+N444I!**1--44ZII))"-- 0 0D Kr)T)r)rrintTensorrr'r.r:rrr>s+"  L "+u|E22        \    T9=KK K#(<K/3K59K25 K 03 K 7: K26KKKK\5<#elr