Vh> NdZddlZddlZd dZejdZ d d ZdS) z Ways to make the model stronger.Nc|dksJ|jd|jdksJ|jd}tj|||j|j}t |D]6}||}|dd}|d|zz }7| S)zXThis is the power method. batch size is used to try multiple starting point in parallel.rr)devicedtypeT)dimkeepdimg|=) rshapetorchrandnrrrangemmnormmean)mnitersbsrb_nrs P/var/www/html/movieo_spanner_bot/venv/lib/python3.11/site-packages/demucs/svd.pypower_iterationr s 5577a<<<< 71: # # # # '!*C CAHAG<<| dvr(| dd } | dkr$| t| d } nZ| d kr$| t| d } n| dkr$|r(| d ksJ| | jf|rBt j| d dd }n|rq| j\}}||kr(| | }n'| | }t+||| }n5t j| ||ddd }| |z } D_| |z S) a Penalty on the largest singular value for a layer. Args: - model: model to penalize - min_size: minimum size in MB of a layer to penalize. - dim: projection dimension for the svd_lowrank. Higher is better but slower. - niters: number of iterations in the algorithm used by svd_lowrank. - powm: use power method instead of lowrank SVD, my own experience is that it is both slower and less stable. - convtr: when True, differentiate between Conv and Transposed Conv. this is kept for compatibility with older experiments. - proba: probability to apply the penalty. - conv_only: only apply to conv and conv transposed, not LSTM (might not be reliable for other models than Demucs). - exact: use exact SVD (slow but useful at validation). - bs: batch_size for power method. rgF)recursei)rrrr) compute_uv) penalty_rngrandommodulesnamed_parametersnumel isinstancer nnConvTranspose1dConvTranspose2dr transpose contiguousviewlenr svdpowmaxrtr svd_lowrank)modelmin_sizerrpowmconvtrproba conv_onlyexactrtotalrnamepestimatearrs r svd_penaltyr?sU& Ee##r ]]__))%)88  GD!wwyy5 8++ ;a%(":EHrCsn '&     "fmD!! JN:;444444r