|
import torch |
|
import numpy as np |
|
from typing import Union |
|
|
|
|
|
def _to_tuple(x): |
|
if isinstance(x, int): |
|
return x, x |
|
else: |
|
return x |
|
|
|
|
|
def get_fill_resize_and_crop(src, tgt): |
|
th, tw = _to_tuple(tgt) |
|
h, w = _to_tuple(src) |
|
|
|
tr = th / tw |
|
r = h / w |
|
|
|
|
|
if r > tr: |
|
resize_height = th |
|
resize_width = int(round(th / h * w)) |
|
else: |
|
resize_width = tw |
|
resize_height = int(round(tw / w * h)) |
|
|
|
crop_top = int(round((th - resize_height) / 2.0)) |
|
crop_left = int(round((tw - resize_width) / 2.0)) |
|
|
|
return (crop_top, crop_left), (crop_top + resize_height, crop_left + resize_width) |
|
|
|
|
|
def get_meshgrid(start, *args): |
|
if len(args) == 0: |
|
|
|
num = _to_tuple(start) |
|
start = (0, 0) |
|
stop = num |
|
elif len(args) == 1: |
|
|
|
start = _to_tuple(start) |
|
stop = _to_tuple(args[0]) |
|
num = (stop[0] - start[0], stop[1] - start[1]) |
|
elif len(args) == 2: |
|
|
|
start = _to_tuple(start) |
|
stop = _to_tuple(args[0]) |
|
num = _to_tuple(args[1]) |
|
else: |
|
raise ValueError(f"len(args) should be 0, 1 or 2, but got {len(args)}") |
|
|
|
grid_h = np.linspace(start[0], stop[0], num[0], endpoint=False, dtype=np.float32) |
|
grid_w = np.linspace(start[1], stop[1], num[1], endpoint=False, dtype=np.float32) |
|
grid = np.meshgrid(grid_w, grid_h) |
|
grid = np.stack(grid, axis=0) |
|
return grid |
|
|
|
|
|
|
|
|
|
|
|
|
|
def get_2d_sincos_pos_embed(embed_dim, start, *args, cls_token=False, extra_tokens=0): |
|
""" |
|
grid_size: int of the grid height and width |
|
return: |
|
pos_embed: [grid_size*grid_size, embed_dim] or [1+grid_size*grid_size, embed_dim] (w/ or w/o cls_token) |
|
""" |
|
grid = get_meshgrid(start, *args) |
|
|
|
|
|
|
|
|
|
|
|
grid = grid.reshape([2, 1, *grid.shape[1:]]) |
|
pos_embed = get_2d_sincos_pos_embed_from_grid(embed_dim, grid) |
|
if cls_token and extra_tokens > 0: |
|
pos_embed = np.concatenate([np.zeros([extra_tokens, embed_dim]), pos_embed], axis=0) |
|
return pos_embed |
|
|
|
|
|
def get_2d_sincos_pos_embed_from_grid(embed_dim, grid): |
|
assert embed_dim % 2 == 0 |
|
|
|
|
|
emb_h = get_1d_sincos_pos_embed_from_grid(embed_dim // 2, grid[0]) |
|
emb_w = get_1d_sincos_pos_embed_from_grid(embed_dim // 2, grid[1]) |
|
|
|
emb = np.concatenate([emb_h, emb_w], axis=1) |
|
return emb |
|
|
|
|
|
def get_1d_sincos_pos_embed_from_grid(embed_dim, pos): |
|
""" |
|
embed_dim: output dimension for each position |
|
pos: a list of positions to be encoded: size (W,H) |
|
out: (M, D) |
|
""" |
|
assert embed_dim % 2 == 0 |
|
omega = np.arange(embed_dim // 2, dtype=np.float64) |
|
omega /= embed_dim / 2. |
|
omega = 1. / 10000**omega |
|
|
|
pos = pos.reshape(-1) |
|
out = np.einsum('m,d->md', pos, omega) |
|
|
|
emb_sin = np.sin(out) |
|
emb_cos = np.cos(out) |
|
|
|
emb = np.concatenate([emb_sin, emb_cos], axis=1) |
|
return emb |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def get_2d_rotary_pos_embed(embed_dim, start, *args, use_real=True): |
|
""" |
|
This is a 2d version of precompute_freqs_cis, which is a RoPE for image tokens with 2d structure. |
|
|
|
Parameters |
|
---------- |
|
embed_dim: int |
|
embedding dimension size |
|
start: int or tuple of int |
|
If len(args) == 0, start is num; If len(args) == 1, start is start, args[0] is stop, step is 1; |
|
If len(args) == 2, start is start, args[0] is stop, args[1] is num. |
|
use_real: bool |
|
If True, return real part and imaginary part separately. Otherwise, return complex numbers. |
|
|
|
Returns |
|
------- |
|
pos_embed: torch.Tensor |
|
[HW, D/2] |
|
""" |
|
grid = get_meshgrid(start, *args) |
|
grid = grid.reshape([2, 1, *grid.shape[1:]]) |
|
pos_embed = get_2d_rotary_pos_embed_from_grid(embed_dim, grid, use_real=use_real) |
|
return pos_embed |
|
|
|
|
|
def get_2d_rotary_pos_embed_from_grid(embed_dim, grid, use_real=False): |
|
assert embed_dim % 4 == 0 |
|
|
|
|
|
emb_h = get_1d_rotary_pos_embed(embed_dim // 2, grid[0].reshape(-1), use_real=use_real) |
|
emb_w = get_1d_rotary_pos_embed(embed_dim // 2, grid[1].reshape(-1), use_real=use_real) |
|
|
|
if use_real: |
|
cos = torch.cat([emb_h[0], emb_w[0]], dim=1) |
|
sin = torch.cat([emb_h[1], emb_w[1]], dim=1) |
|
return cos, sin |
|
else: |
|
emb = torch.cat([emb_h, emb_w], dim=1) |
|
return emb |
|
|
|
|
|
def get_1d_rotary_pos_embed(dim: int, pos: Union[np.ndarray, int], theta: float = 10000.0, use_real=False): |
|
""" |
|
Precompute the frequency tensor for complex exponentials (cis) with given dimensions. |
|
|
|
This function calculates a frequency tensor with complex exponentials using the given dimension 'dim' |
|
and the end index 'end'. The 'theta' parameter scales the frequencies. |
|
The returned tensor contains complex values in complex64 data type. |
|
|
|
Args: |
|
dim (int): Dimension of the frequency tensor. |
|
pos (np.ndarray, int): Position indices for the frequency tensor. [S] or scalar |
|
theta (float, optional): Scaling factor for frequency computation. Defaults to 10000.0. |
|
use_real (bool, optional): If True, return real part and imaginary part separately. |
|
Otherwise, return complex numbers. |
|
|
|
Returns: |
|
torch.Tensor: Precomputed frequency tensor with complex exponentials. [S, D/2] |
|
|
|
""" |
|
if isinstance(pos, int): |
|
pos = np.arange(pos) |
|
freqs = 1.0 / (theta ** (torch.arange(0, dim, 2)[: (dim // 2)].float() / dim)) |
|
t = torch.from_numpy(pos).to(freqs.device) |
|
freqs = torch.outer(t, freqs).float() |
|
if use_real: |
|
freqs_cos = freqs.cos().repeat_interleave(2, dim=1) |
|
freqs_sin = freqs.sin().repeat_interleave(2, dim=1) |
|
return freqs_cos, freqs_sin |
|
else: |
|
freqs_cis = torch.polar(torch.ones_like(freqs), freqs) |
|
return freqs_cis |
|
|
|
|
|
|
|
def calc_sizes(rope_img, patch_size, th, tw): |
|
if rope_img == 'extend': |
|
|
|
sub_args = [(th, tw)] |
|
elif rope_img.startswith('base'): |
|
|
|
base_size = int(rope_img[4:]) // 8 // patch_size |
|
start, stop = get_fill_resize_and_crop((th, tw), base_size) |
|
sub_args = [start, stop, (th, tw)] |
|
else: |
|
raise ValueError(f"Unknown rope_img: {rope_img}") |
|
return sub_args |
|
|
|
|
|
def init_image_posemb(rope_img, |
|
resolutions, |
|
patch_size, |
|
hidden_size, |
|
num_heads, |
|
log_fn, |
|
rope_real=True, |
|
): |
|
freqs_cis_img = {} |
|
for reso in resolutions: |
|
th, tw = reso.height // 8 // patch_size, reso.width // 8 // patch_size |
|
sub_args = calc_sizes(rope_img, patch_size, th, tw) |
|
freqs_cis_img[str(reso)] = get_2d_rotary_pos_embed(hidden_size // num_heads, *sub_args, use_real=rope_real) |
|
log_fn(f" Using image RoPE ({rope_img}) ({'real' if rope_real else 'complex'}): {sub_args} | ({reso}) " |
|
f"{freqs_cis_img[str(reso)][0].shape if rope_real else freqs_cis_img[str(reso)].shape}") |
|
return freqs_cis_img |
|
|