only used for inference, on autoregressive mode :param B: batch size :param label_B: imagenet label; if None, randomly sample

Compute pairwise distances between two batches of feature vectors.

Compute image features for downstream evals. :param batches: a iterator over NHWC numpy arrays in [0, 255]. :return: a tuple

Evaluate precision and recall efficiently. :param features_1: [N1 x D] feature vectors for reference batch. :param radii_1:

Here, f_BCHW is the feature from inp_img_no_grad, H, W is the size of processing by encoder

Compute the Frechet distance between two sets of statistics.

Evaluate pairwise distances between two batches of feature vectors.

Copied from: https://github.com/numpy/numpy/blob/fb215c76967739268de71aa4bda55dd1b062bc2e/numpy/lib/format.py#L788-L886 Read from file-like

Center cropping implementation from ADM. https://github.com/openai/guided-diffusion/blob/8fb3ad9197f16bbc40620447b2742e13458d2831/guided_diff

Builds a single .npz file from a folder of .png samples. Refer to DiT.

:param label_B: label_B :param x_BlCv_wo_first_l: teacher forcing input (B, self.l, self.Cvae) :return: logits BLV, V is voca

Decay the learning rate with half-cycle cosine after warmup