gongchenting
commented
1 year ago I have the same problem. The output of the logsumexp function may be less than 0, but the input of the Weibull distribution needs to be greater than 0. Have you solved it？
ValueError: Expected value argument (Tensor of shape (100,)) to be within the support (GreaterThan(lower_bound=0.0)) of the distribution Weibull(scale: 3.6472277641296387, concentration: 2.517430067062378), but found invalid values: tensor([-5.6482, -5.6381, -5.6280, -5.6179, -5.6078, -5.5977, -5.5876, -5.5775, -5.5674, -5.5573, -5.5472, -5.5371, -5.5270, -5.5169, -5.5068, -5.4967, -5.4866, -5.4765, -5.4664, -5.4563, -5.4462, -5.4361, -5.4260, -5.4159, -5.4058, -5.3957, -5.3856, -5.3755, -5.3654, -5.3553, -5.3452, -5.3351, -5.3250, -5.3149, -5.3048, -5.2947, -5.2846, -5.2745, -5.2644, -5.2543, -5.2442, -5.2341, -5.2240, -5.2139, -5.2038, -5.1937, -5.1836, -5.1735, -5.1634, -5.1533, -5.1432, -5.1331, -5.1230, -5.1129, -5.1028, -5.0927, -5.0826, -5.0725, -5.0624, -5.0523, -5.0422, -5.0321, -5.0220, -5.0118, -5.0017, -4.9916, -4.9815, -4.9714, -4.9613, -4.9512, -4.9411, -4.9310, -4.9209, -4.9108, -4.9007, -4.8906, -4.8805, -4.8704, -4.8603, -4.8502, -4.8401, -4.8300, -4.8199, -4.8098, -4.7997, -4.7896, -4.7795, -4.7694, -4.7593, -4.7492, -4.7391, -4.7290, -4.7189, -4.7088, -4.6987, -4.6886,