# --------------------------------------------------------
# BEIT: BERT Pre-Training of Image Transformers (https://arxiv.org/abs/2106.08254)
# Github source: https://github.com/microsoft/unilm/tree/master/beit
# Copyright (c) 2021 Microsoft
# Licensed under The MIT License [see LICENSE for details]
# By Hangbo Bao
# Based on timm and DeiT code bases
# https://github.com/rwightman/pytorch-image-models/tree/master/timm
# https://github.com/facebookresearch/deit/
# https://github.com/facebookresearch/dino
# --------------------------------------------------------'
import math
from functools import partial
from scipy import interpolate

import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
from timm.models.layers import drop_path, to_2tuple, trunc_normal_
#from timm.models.registry import register_model


def _cfg(url='', **kwargs):
    return {
        'url': url,
        'num_classes': 1000, 'input_size': (3, 224, 224), 'pool_size': None,
        'crop_pct': .9, 'interpolation': 'bicubic',
        'mean': (0.5, 0.5, 0.5), 'std': (0.5, 0.5, 0.5),
        **kwargs
    }


class DropPath(nn.Module):
    """Drop paths (Stochastic Depth) per sample  (when applied in main path of residual blocks).
    """
    def __init__(self, drop_prob=None):
        super(DropPath, self).__init__()
        self.drop_prob = drop_prob

    def forward(self, x):
        return drop_path(x, self.drop_prob, self.training)

    def extra_repr(self) -> str:
        return 'p={}'.format(self.drop_prob)


class Mlp(nn.Module):
    def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
        super().__init__()
        out_features = out_features or in_features
        hidden_features = hidden_features or in_features
        self.fc1 = nn.Linear(in_features, hidden_features)
        self.act = act_layer()
        self.fc2 = nn.Linear(hidden_features, out_features)
        self.drop = nn.Dropout(drop)

    def forward(self, x):
        x = self.fc1(x)
        x = self.act(x)
        # x = self.drop(x)
        # commit this for the orignal BERT implement
        x = self.fc2(x)
        x = self.drop(x)
        return x


class Attention(nn.Module):
    def __init__(
            self, dim, num_heads=8, qkv_bias=False, qk_scale=None, attn_drop=0.,
            proj_drop=0., window_size=None, attn_head_dim=None):
        super().__init__()
        self.num_heads = num_heads
        head_dim = dim // num_heads
        if attn_head_dim is not None:
            head_dim = attn_head_dim
        all_head_dim = head_dim * self.num_heads
        self.scale = qk_scale or head_dim ** -0.5

        self.qkv = nn.Linear(dim, all_head_dim * 3, bias=False)
        if qkv_bias:
            self.q_bias = nn.Parameter(torch.zeros(all_head_dim))
            self.v_bias = nn.Parameter(torch.zeros(all_head_dim))
        else:
            self.q_bias = None
            self.v_bias = None

        if window_size:
            self.window_size = window_size
            self.num_relative_distance = (2 * window_size[0] - 1) * (2 * window_size[1] - 1) + 3
            self.relative_position_bias_table = nn.Parameter(
                torch.zeros(self.num_relative_distance, num_heads))  # 2*Wh-1 * 2*Ww-1, nH
            # cls to token & token 2 cls & cls to cls

            # get pair-wise relative position index for each token inside the window
            coords_h = torch.arange(window_size[0])
            coords_w = torch.arange(window_size[1])
            coords = torch.stack(torch.meshgrid([coords_h, coords_w]))  # 2, Wh, Ww
            coords_flatten = torch.flatten(coords, 1)  # 2, Wh*Ww
            relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]  # 2, Wh*Ww, Wh*Ww
            relative_coords = relative_coords.permute(1, 2, 0).contiguous()  # Wh*Ww, Wh*Ww, 2
            relative_coords[:, :, 0] += window_size[0] - 1  # shift to start from 0
            relative_coords[:, :, 1] += window_size[1] - 1
            relative_coords[:, :, 0] *= 2 * window_size[1] - 1
            relative_position_index = \
                torch.zeros(size=(window_size[0] * window_size[1] + 1, ) * 2, dtype=relative_coords.dtype)
            relative_position_index[1:, 1:] = relative_coords.sum(-1)  # Wh*Ww, Wh*Ww
            relative_position_index[0, 0:] = self.num_relative_distance - 3
            relative_position_index[0:, 0] = self.num_relative_distance - 2
            relative_position_index[0, 0] = self.num_relative_distance - 1

            self.register_buffer("relative_position_index", relative_position_index)
        else:
            self.window_size = None
            self.relative_position_bias_table = None
            self.relative_position_index = None

        self.attn_drop = nn.Dropout(attn_drop)
        self.proj = nn.Linear(all_head_dim, dim)
        self.proj_drop = nn.Dropout(proj_drop)

    def forward(self, x, rel_pos_bias=None):
        B, N, C = x.shape
        qkv_bias = None
        if self.q_bias is not None:
            qkv_bias = torch.cat((self.q_bias, torch.zeros_like(self.v_bias, requires_grad=False), self.v_bias))
        # qkv = self.qkv(x).reshape(B, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
        qkv = F.linear(input=x, weight=self.qkv.weight, bias=qkv_bias)
        qkv = qkv.reshape(B, N, 3, self.num_heads, -1).permute(2, 0, 3, 1, 4)
        q, k, v = qkv[0], qkv[1], qkv[2]   # make torchscript happy (cannot use tensor as tuple)

        q = q * self.scale
        attn = (q @ k.transpose(-2, -1))

        if self.relative_position_bias_table is not None:
            relative_position_bias = \
                self.relative_position_bias_table[self.relative_position_index.view(-1)].view(
                    self.window_size[0] * self.window_size[1] + 1,
                    self.window_size[0] * self.window_size[1] + 1, -1)  # Wh*Ww,Wh*Ww,nH
            relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()  # nH, Wh*Ww, Wh*Ww
            attn = attn + relative_position_bias.unsqueeze(0)

        if rel_pos_bias is not None:
            attn = attn + rel_pos_bias

        attn = attn.softmax(dim=-1)
        attn = self.attn_drop(attn)

        x = (attn @ v).transpose(1, 2).reshape(B, N, -1)
        x = self.proj(x)
        x = self.proj_drop(x)
        return x


class Block(nn.Module):

    def __init__(self, dim, num_heads, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop=0., attn_drop=0.,
                 drop_path=0., init_values=None, act_layer=nn.GELU, norm_layer=nn.LayerNorm,
                 window_size=None, attn_head_dim=None):
        super().__init__()
        self.norm1 = norm_layer(dim)
        self.attn = Attention(
            dim, num_heads=num_heads, qkv_bias=qkv_bias, qk_scale=qk_scale,
            attn_drop=attn_drop, proj_drop=drop, window_size=window_size, attn_head_dim=attn_head_dim)
        # NOTE: drop path for stochastic depth, we shall see if this is better than dropout here
        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
        self.norm2 = norm_layer(dim)
        mlp_hidden_dim = int(dim * mlp_ratio)
        self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)

        if init_values > 0:
            self.gamma_1 = nn.Parameter(init_values * torch.ones((dim)),requires_grad=True)
            self.gamma_2 = nn.Parameter(init_values * torch.ones((dim)),requires_grad=True)
        else:
            self.gamma_1, self.gamma_2 = None, None

    def forward(self, x, rel_pos_bias=None):
        if self.gamma_1 is None:
            x = x + self.drop_path(self.attn(self.norm1(x), rel_pos_bias=rel_pos_bias))
            x = x + self.drop_path(self.mlp(self.norm2(x)))
        else:
            x = x + self.drop_path(self.gamma_1 * self.attn(self.norm1(x), rel_pos_bias=rel_pos_bias))
            x = x + self.drop_path(self.gamma_2 * self.mlp(self.norm2(x)))
        return x


class PatchEmbed(nn.Module):
    """ Image to Patch Embedding
    """
    def __init__(self, img_size=224, patch_size=16, in_chans=3, embed_dim=768):
        super().__init__()
        img_size = to_2tuple(img_size)
        patch_size = to_2tuple(patch_size)
        num_patches = (img_size[1] // patch_size[1]) * (img_size[0] // patch_size[0])
        self.patch_shape = (img_size[0] // patch_size[0], img_size[1] // patch_size[1])
        self.img_size = img_size
        self.patch_size = patch_size
        self.num_patches = num_patches

        self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)

    def forward(self, x, **kwargs):
        B, C, H, W = x.shape
        # FIXME look at relaxing size constraints
        assert H == self.img_size[0] and W == self.img_size[1], \
            f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."
        x = self.proj(x).flatten(2).transpose(1, 2)
        return x


class RelativePositionBias(nn.Module):

    def __init__(self, window_size, num_heads):
        super().__init__()
        self.window_size = window_size
        self.num_relative_distance = (2 * window_size[0] - 1) * (2 * window_size[1] - 1) + 3
        self.relative_position_bias_table = nn.Parameter(
            torch.zeros(self.num_relative_distance, num_heads))  # 2*Wh-1 * 2*Ww-1, nH
        # cls to token & token 2 cls & cls to cls

        # get pair-wise relative position index for each token inside the window
        coords_h = torch.arange(window_size[0])
        coords_w = torch.arange(window_size[1])
        coords = torch.stack(torch.meshgrid([coords_h, coords_w]))  # 2, Wh, Ww
        coords_flatten = torch.flatten(coords, 1)  # 2, Wh*Ww
        relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]  # 2, Wh*Ww, Wh*Ww
        relative_coords = relative_coords.permute(1, 2, 0).contiguous()  # Wh*Ww, Wh*Ww, 2
        relative_coords[:, :, 0] += window_size[0] - 1  # shift to start from 0
        relative_coords[:, :, 1] += window_size[1] - 1
        relative_coords[:, :, 0] *= 2 * window_size[1] - 1
        relative_position_index = \
            torch.zeros(size=(window_size[0] * window_size[1] + 1,) * 2, dtype=relative_coords.dtype)
        relative_position_index[1:, 1:] = relative_coords.sum(-1)  # Wh*Ww, Wh*Ww
        relative_position_index[0, 0:] = self.num_relative_distance - 3
        relative_position_index[0:, 0] = self.num_relative_distance - 2
        relative_position_index[0, 0] = self.num_relative_distance - 1

        self.register_buffer("relative_position_index", relative_position_index)

        # trunc_normal_(self.relative_position_bias_table, std=.02)

    def forward(self):
        relative_position_bias = \
            self.relative_position_bias_table[self.relative_position_index.view(-1)].view(
                self.window_size[0] * self.window_size[1] + 1,
                self.window_size[0] * self.window_size[1] + 1, -1)  # Wh*Ww,Wh*Ww,nH
        return relative_position_bias.permute(2, 0, 1).contiguous()  # nH, Wh*Ww, Wh*Ww


class VisionTransformer(nn.Module):
    """ Vision Transformer with support for patch or hybrid CNN input stage
    """
    def __init__(self, img_size=224, patch_size=16, in_chans=3, num_classes=1000, embed_dim=768, depth=12,
                 num_heads=12, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop_rate=0., attn_drop_rate=0.,
                 drop_path_rate=0., norm_layer=nn.LayerNorm, init_values=None,
                 use_abs_pos_emb=True, use_rel_pos_bias=False, use_shared_rel_pos_bias=False,
                 use_mean_pooling=True, init_scale=0.001):
        super().__init__()
        self.num_classes = num_classes
        self.num_features = self.embed_dim = embed_dim  # num_features for consistency with other models

        self.patch_embed = PatchEmbed(
            img_size=img_size, patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim)
        num_patches = self.patch_embed.num_patches

        self.cls_token = nn.Parameter(torch.zeros(1, 1, embed_dim))
        # self.mask_token = nn.Parameter(torch.zeros(1, 1, embed_dim))
        if use_abs_pos_emb:
            self.pos_embed = nn.Parameter(torch.zeros(1, num_patches + 1, embed_dim))
        else:
            self.pos_embed = None
        self.pos_drop = nn.Dropout(p=drop_rate)

        if use_shared_rel_pos_bias:
            self.rel_pos_bias = RelativePositionBias(window_size=self.patch_embed.patch_shape, num_heads=num_heads)
        else:
            self.rel_pos_bias = None

        dpr = [x.item() for x in torch.linspace(0, drop_path_rate, depth)]  # stochastic depth decay rule
        self.use_rel_pos_bias = use_rel_pos_bias
        self.blocks = nn.ModuleList([
            Block(
                dim=embed_dim, num_heads=num_heads, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
                drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[i], norm_layer=norm_layer,
                init_values=init_values, window_size=self.patch_embed.patch_shape if use_rel_pos_bias else None)
            for i in range(depth)])
        self.norm = nn.Identity() if use_mean_pooling else norm_layer(embed_dim)
        self.fc_norm = norm_layer(embed_dim) if use_mean_pooling else None
        self.head = nn.Linear(embed_dim, num_classes) if num_classes > 0 else nn.Identity()

        if self.pos_embed is not None:
            trunc_normal_(self.pos_embed, std=.02)
        trunc_normal_(self.cls_token, std=.02)
        # trunc_normal_(self.mask_token, std=.02)
        self.apply(self._init_weights)
        self.fix_init_weight()

        if num_classes > 0:
            trunc_normal_(self.head.weight, std=.02)
            self.head.weight.data.mul_(init_scale)
            self.head.bias.data.mul_(init_scale)

    def fix_init_weight(self):
        def rescale(param, layer_id):
            param.div_(math.sqrt(2.0 * layer_id))

        for layer_id, layer in enumerate(self.blocks):
            rescale(layer.attn.proj.weight.data, layer_id + 1)
            rescale(layer.mlp.fc2.weight.data, layer_id + 1)

    def _init_weights(self, m):
        if isinstance(m, nn.Linear):
            trunc_normal_(m.weight, std=.02)
            if isinstance(m, nn.Linear) and m.bias is not None:
                nn.init.constant_(m.bias, 0)
        elif isinstance(m, nn.LayerNorm):
            nn.init.constant_(m.bias, 0)
            nn.init.constant_(m.weight, 1.0)

    def get_num_layers(self):
        return len(self.blocks)

    @torch.jit.ignore
    def no_weight_decay(self):
        return {'pos_embed', 'cls_token'}

    def get_classifier(self):
        return self.head

    def reset_classifier(self, num_classes, global_pool=''):
        self.num_classes = num_classes
        self.head = nn.Linear(self.embed_dim, num_classes) if num_classes > 0 else nn.Identity()

    def forward_features(self, x):
        x = self.patch_embed(x)
        batch_size, seq_len, _ = x.size()

        cls_tokens = self.cls_token.expand(batch_size, -1, -1)  # stole cls_tokens impl from Phil Wang, thanks
        x = torch.cat((cls_tokens, x), dim=1)
        if self.pos_embed is not None:
            x = x + self.pos_embed
        x = self.pos_drop(x)

        rel_pos_bias = self.rel_pos_bias() if self.rel_pos_bias is not None else None
        for blk in self.blocks:
            x = blk(x, rel_pos_bias=rel_pos_bias)

        x = self.norm(x)
        if self.fc_norm is not None:
            t = x[:, 1:, :]
            return self.fc_norm(t.mean(1))
        else:
            return x[:, 0]

    def forward(self, x):
        x = self.forward_features(x)
        x = self.head(x)
        return x


#@register_model
def beit_base_patch16_224(pretrained=False, **kwargs):
    model = VisionTransformer(
        patch_size=16, embed_dim=768, depth=12, num_heads=12, mlp_ratio=4, qkv_bias=True,
        norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    model.default_cfg = _cfg()
    return model


#@register_model
def beit_base_patch16_384(pretrained=False, **kwargs):
    model = VisionTransformer(
        img_size=384, patch_size=16, embed_dim=768, depth=12, num_heads=12, mlp_ratio=4, qkv_bias=True,
        norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    model.default_cfg = _cfg()
    return model


#@register_model
def beit_large_patch16_224(pretrained=False, **kwargs):
    model = VisionTransformer(
        patch_size=16, embed_dim=1024, depth=24, num_heads=16, mlp_ratio=4, qkv_bias=True,
        norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    model.default_cfg = _cfg()
    return model


#@register_model
def beit_large_patch16_384(pretrained=False, **kwargs):
    model = VisionTransformer(
        img_size=384, patch_size=16, embed_dim=1024, depth=24, num_heads=16, mlp_ratio=4, qkv_bias=True,
        norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    model.default_cfg = _cfg()
    return model


#@register_model
def beit_large_patch16_512(pretrained=False, **kwargs):
    model = VisionTransformer(
        img_size=512, patch_size=16, embed_dim=1024, depth=24, num_heads=16, mlp_ratio=4, qkv_bias=True,
        norm_layer=partial(nn.LayerNorm, eps=1e-6), **kwargs)
    model.default_cfg = _cfg()
    return model


def load_state_dict(model, state_dict, prefix='', ignore_missing="relative_position_index"):
    missing_keys = []
    unexpected_keys = []
    error_msgs = []
    # copy state_dict so _load_from_state_dict can modify it
    metadata = getattr(state_dict, '_metadata', None)
    state_dict = state_dict.copy()
    if metadata is not None:
        state_dict._metadata = metadata

    def _load(module, prefix=''):
        local_metadata = {} if metadata is None else metadata.get(
            prefix[:-1], {})
        module._load_from_state_dict(
            state_dict, prefix, local_metadata, True, missing_keys, unexpected_keys, error_msgs)
        for name, child in module._modules.items():
            if child is not None:
                _load(child, prefix + name + '.')

    _load(model, prefix=prefix)

    warn_missing_keys = []
    ignore_missing_keys = []
    for key in missing_keys:
        keep_flag = True
        for ignore_key in ignore_missing.split('|'):
            if ignore_key in key:
                keep_flag = False
                break
        if keep_flag:
            warn_missing_keys.append(key)
        else:
            ignore_missing_keys.append(key)

    missing_keys = warn_missing_keys

    if len(missing_keys) > 0:
        print("Weights of {} not initialized from pretrained model: {}".format(
            model.__class__.__name__, missing_keys))
    if len(unexpected_keys) > 0:
        print("Weights from pretrained model not used in {}: {}".format(
            model.__class__.__name__, unexpected_keys))
    if len(ignore_missing_keys) > 0:
        print("Ignored weights of {} not initialized from pretrained model: {}".format(
            model.__class__.__name__, ignore_missing_keys))
    if len(error_msgs) > 0:
        print('\n'.join(error_msgs))


def default_pretrained_model(args):
    model = beit_base_patch16_224(
        pretrained=False,
        img_size=args.image_size,
        num_classes=0,
        drop_rate=0.,
        drop_path_rate=0.1,
        attn_drop_rate=0.,
        #drop_block_rate=None,
        use_mean_pooling=True,
        init_scale=0.001,
        use_rel_pos_bias=True,
        use_abs_pos_emb=False,
        init_values=0.1,
    )

    #url = 'https://unilm.blob.core.windows.net/beit/beit_base_patch16_224_pt22k.pth'
    url = 'https://unilm.blob.core.windows.net/beit/beit_base_patch16_224_pt22k_ft22k.pth'

    checkpoint = torch.hub.load_state_dict_from_url(
        url, map_location='cpu', check_hash=True)
    print('Pretrained weights found at {}'.format(url))

    # select key
    checkpoint_model = None
    for model_key in ['model', 'module']:
        if model_key in checkpoint:
            checkpoint_model = checkpoint[model_key]
            print("Load state_dict by model_key = %s" % model_key)
            break
    if checkpoint_model is None:
        checkpoint_model = checkpoint

    # remove head
    state_dict = model.state_dict()
    for k in ['head.weight', 'head.bias']:
        #if k in checkpoint_model and checkpoint_model[k].shape != state_dict[k].shape:
        if k in checkpoint_model:
            print(f"Removing key {k} from pretrained checkpoint")
            del checkpoint_model[k]

    # resize rel_pos_bias
    if model.use_rel_pos_bias and "rel_pos_bias.relative_position_bias_table" in checkpoint_model:
        print("Expand the shared relative position embedding to each transformer block. ")
        num_layers = model.get_num_layers()
        rel_pos_bias = checkpoint_model["rel_pos_bias.relative_position_bias_table"]
        for i in range(num_layers):
            checkpoint_model["blocks.%d.attn.relative_position_bias_table" % i] = rel_pos_bias.clone()

        checkpoint_model.pop("rel_pos_bias.relative_position_bias_table")

    all_keys = list(checkpoint_model.keys())
    for key in all_keys:
        if "relative_position_index" in key:
            checkpoint_model.pop(key)

        if "relative_position_bias_table" in key:
            rel_pos_bias = checkpoint_model[key]
            src_num_pos, num_attn_heads = rel_pos_bias.size()
            dst_num_pos, _ = model.state_dict()[key].size()
            dst_patch_shape = model.patch_embed.patch_shape
            if dst_patch_shape[0] != dst_patch_shape[1]:
                raise NotImplementedError()
            num_extra_tokens = dst_num_pos - (dst_patch_shape[0] * 2 - 1) * (dst_patch_shape[1] * 2 - 1)
            src_size = int((src_num_pos - num_extra_tokens) ** 0.5)
            dst_size = int((dst_num_pos - num_extra_tokens) ** 0.5)
            if src_size != dst_size:
                print("Position interpolate for %s from %dx%d to %dx%d" % (
                    key, src_size, src_size, dst_size, dst_size))
                extra_tokens = rel_pos_bias[-num_extra_tokens:, :]
                rel_pos_bias = rel_pos_bias[:-num_extra_tokens, :]

                def geometric_progression(a, r, n):
                    return a * (1.0 - r ** n) / (1.0 - r)

                left, right = 1.01, 1.5
                while right - left > 1e-6:
                    q = (left + right) / 2.0
                    gp = geometric_progression(1, q, src_size // 2)
                    if gp > dst_size // 2:
                        right = q
                    else:
                        left = q

                # if q > 1.090307:
                #     q = 1.090307

                dis = []
                cur = 1
                for i in range(src_size // 2):
                    dis.append(cur)
                    cur += q ** (i + 1)

                r_ids = [-_ for _ in reversed(dis)]

                x = r_ids + [0] + dis
                y = r_ids + [0] + dis

                t = dst_size // 2.0
                dx = np.arange(-t, t + 0.1, 1.0)
                dy = np.arange(-t, t + 0.1, 1.0)

                print("Original positions = %s" % str(x))
                print("Target positions = %s" % str(dx))

                all_rel_pos_bias = []

                for i in range(num_attn_heads):
                    z = rel_pos_bias[:, i].view(src_size, src_size).float().numpy()
                    f = interpolate.interp2d(x, y, z, kind='cubic')
                    all_rel_pos_bias.append(
                        torch.Tensor(f(dx, dy)).contiguous().view(-1, 1).to(rel_pos_bias.device))

                rel_pos_bias = torch.cat(all_rel_pos_bias, dim=-1)

                new_rel_pos_bias = torch.cat((rel_pos_bias, extra_tokens), dim=0)
                checkpoint_model[key] = new_rel_pos_bias

    # interpolate position embedding
    if 'pos_embed' in checkpoint_model:
        pos_embed_checkpoint = checkpoint_model['pos_embed']
        embedding_size = pos_embed_checkpoint.shape[-1]
        num_patches = model.patch_embed.num_patches
        num_extra_tokens = model.pos_embed.shape[-2] - num_patches
        # height (== width) for the checkpoint position embedding
        orig_size = int((pos_embed_checkpoint.shape[-2] - num_extra_tokens) ** 0.5)
        # height (== width) for the new position embedding
        new_size = int(num_patches ** 0.5)
        # class_token and dist_token are kept unchanged
        if orig_size != new_size:
            print("Position interpolate from %dx%d to %dx%d" % (orig_size, orig_size, new_size, new_size))
            extra_tokens = pos_embed_checkpoint[:, :num_extra_tokens]
            # only the position tokens are interpolated
            pos_tokens = pos_embed_checkpoint[:, num_extra_tokens:]
            pos_tokens = pos_tokens.reshape(-1, orig_size, orig_size, embedding_size).permute(0, 3, 1, 2)
            pos_tokens = torch.nn.functional.interpolate(
                pos_tokens, size=(new_size, new_size), mode='bicubic', align_corners=False)
            pos_tokens = pos_tokens.permute(0, 2, 3, 1).flatten(1, 2)
            new_pos_embed = torch.cat((extra_tokens, pos_tokens), dim=1)
            checkpoint_model['pos_embed'] = new_pos_embed

    load_state_dict(model, checkpoint_model)
    return model