hayeong0 / DDDM-VC

Official Pytorch Implementation for "DDDM-VC: Decoupled Denoising Diffusion Models with Disentangled Representation and Prior Mixup for Verified Robust Voice Conversion" (AAAI 2024)
https://hayeong0.github.io/DDDM-VC-demo/
179 stars 19 forks source link

Some debug information appears during training #14

Closed isMoJo closed 1 month ago

isMoJo commented 3 months ago

Hello, thank you very much for sharing. I encountered these information outputs during the training process. Is this normal? Is there something wrong with my configuration?

DEBUG:numba.interpreter:label 0: x = arg(0, name=x) ['x'] y = arg(1, name=y) ['y'] sample_ratio = arg(2, name=sample_ratio) ['sample_ratio'] interp_win = arg(3, name=interp_win) ['interp_win'] interp_delta = arg(4, name=interp_delta) ['interp_delta'] num_table = arg(5, name=num_table) ['num_table'] $2load_global.0 = global(min: ) ['$2load_global.0'] $const4.1 = const(float, 1.0) ['$const4.1'] $8call_function.3 = call $2load_global.0($const4.1, sample_ratio, func=$2load_global.0, args=[Var($const4.1, interpn.py:10), Var(sample_ratio, interpn.py:10)], kws=(), vararg=None) ['$2load_global.0', '$8call_function.3', '$const4.1', 'sample_ratio'] scale = $8call_function.3 ['$8call_function.3', 'scale'] $const12.4 = const(float, 1.0) ['$const12.4'] $16binary_true_divide.6 = $const12.4 / sample_ratio ['$16binary_true_divide.6', '$const12.4', 'sample_ratio'] time_increment = $16binary_true_divide.6 ['$16binary_true_divide.6', 'time_increment'] $20load_global.7 = global(int: <class 'int'>) ['$20load_global.7'] $26binary_multiply.10 = scale * num_table ['$26binary_multiply.10', 'num_table', 'scale'] $28call_function.11 = call $20load_global.7($26binary_multiply.10, func=$20load_global.7, args=[Var($26binary_multiply.10, interpn.py:12)], kws=(), vararg=None) ['$20load_global.7', '$26binary_multiply.10', '$28call_function.11'] index_step = $28call_function.11 ['$28call_function.11', 'index_step'] $const32.12 = const(float, 0.0) ['$const32.12'] time_register = $const32.12 ['$const32.12', 'time_register'] $const36.13 = const(int, 0) ['$const36.13'] n = $const36.13 ['$const36.13', 'n'] $const40.14 = const(float, 0.0) ['$const40.14'] frac = $const40.14 ['$const40.14', 'frac'] $const44.15 = const(float, 0.0) ['$const44.15'] index_frac = $const44.15 ['$const44.15', 'index_frac'] $const48.16 = const(int, 0) ['$const48.16'] offset = $const48.16 ['$const48.16', 'offset'] $const52.17 = const(float, 0.0) ['$const52.17'] eta = $const52.17 ['$const52.17', 'eta'] $const56.18 = const(float, 0.0) ['$const56.18'] weight = $const56.18 ['$const56.18', 'weight'] $62load_attr.20 = getattr(value=interp_win, attr=shape) ['$62load_attr.20', 'interp_win'] $const64.21 = const(int, 0) ['$const64.21'] $66binary_subscr.22 = getitem(value=$62load_attr.20, index=$const64.21) ['$62load_attr.20', '$66binary_subscr.22', '$const64.21'] nwin = $66binary_subscr.22 ['$66binary_subscr.22', 'nwin'] $72load_attr.24 = getattr(value=x, attr=shape) ['$72load_attr.24', 'x'] $const74.25 = const(int, 0) ['$const74.25'] $76binary_subscr.26 = getitem(value=$72load_attr.24, index=$const74.25) ['$72load_attr.24', '$76binary_subscr.26', '$const74.25'] n_orig = $76binary_subscr.26 ['$76binary_subscr.26', 'n_orig'] $82load_attr.28 = getattr(value=y, attr=shape) ['$82load_attr.28', 'y'] $const84.29 = const(int, 0) ['$const84.29'] $86binary_subscr.30 = getitem(value=$82load_attr.28, index=$const84.29) ['$82load_attr.28', '$86binary_subscr.30', '$const84.29'] n_out = $86binary_subscr.30 ['$86binary_subscr.30', 'n_out'] $92load_attr.32 = getattr(value=y, attr=shape) ['$92load_attr.32', 'y'] $const94.33 = const(int, 1) ['$const94.33'] $96binary_subscr.34 = getitem(value=$92load_attr.32, index=$const94.33) ['$92load_attr.32', '$96binary_subscr.34', '$const94.33'] n_channels = $96binary_subscr.34 ['$96binary_subscr.34', 'n_channels'] $100load_global.35 = global(range: <class 'range'>) ['$100load_global.35'] $104call_function.37 = call $100load_global.35(n_out, func=$100load_global.35, args=[Var(n_out, interpn.py:24)], kws=(), vararg=None) ['$100load_global.35', '$104call_function.37', 'n_out'] $106get_iter.38 = getiter(value=$104call_function.37) ['$104call_function.37', '$106get_iter.38'] $phi108.0 = $106get_iter.38 ['$106get_iter.38', '$phi108.0']

hayeong0 commented 3 months ago

Hello, sorry for the late reply. It seems that the debug logs are appearing due to the numba version update. Although I haven't used the updated version before, it doesn't seem to be a major issue! Thanks.