|
|
|
"""AutoAnchor utils.""" |
|
|
|
import random |
|
|
|
import numpy as np |
|
import torch |
|
import yaml |
|
from tqdm import tqdm |
|
|
|
from utils import TryExcept |
|
from utils.general import LOGGER, TQDM_BAR_FORMAT, colorstr |
|
|
|
PREFIX = colorstr("AutoAnchor: ") |
|
|
|
|
|
def check_anchor_order(m): |
|
|
|
a = m.anchors.prod(-1).mean(-1).view(-1) |
|
da = a[-1] - a[0] |
|
ds = m.stride[-1] - m.stride[0] |
|
if da and (da.sign() != ds.sign()): |
|
LOGGER.info(f"{PREFIX}Reversing anchor order") |
|
m.anchors[:] = m.anchors.flip(0) |
|
|
|
|
|
@TryExcept(f"{PREFIX}ERROR") |
|
def check_anchors(dataset, model, thr=4.0, imgsz=640): |
|
|
|
m = model.module.model[-1] if hasattr(model, "module") else model.model[-1] |
|
shapes = imgsz * dataset.shapes / dataset.shapes.max(1, keepdims=True) |
|
scale = np.random.uniform(0.9, 1.1, size=(shapes.shape[0], 1)) |
|
wh = torch.tensor(np.concatenate([l[:, 3:5] * s for s, l in zip(shapes * scale, dataset.labels)])).float() |
|
|
|
def metric(k): |
|
r = wh[:, None] / k[None] |
|
x = torch.min(r, 1 / r).min(2)[0] |
|
best = x.max(1)[0] |
|
aat = (x > 1 / thr).float().sum(1).mean() |
|
bpr = (best > 1 / thr).float().mean() |
|
return bpr, aat |
|
|
|
stride = m.stride.to(m.anchors.device).view(-1, 1, 1) |
|
anchors = m.anchors.clone() * stride |
|
bpr, aat = metric(anchors.cpu().view(-1, 2)) |
|
s = f"\n{PREFIX}{aat:.2f} anchors/target, {bpr:.3f} Best Possible Recall (BPR). " |
|
if bpr > 0.98: |
|
LOGGER.info(f"{s}Current anchors are a good fit to dataset ✅") |
|
else: |
|
LOGGER.info(f"{s}Anchors are a poor fit to dataset ⚠️, attempting to improve...") |
|
na = m.anchors.numel() // 2 |
|
anchors = kmean_anchors(dataset, n=na, img_size=imgsz, thr=thr, gen=1000, verbose=False) |
|
new_bpr = metric(anchors)[0] |
|
if new_bpr > bpr: |
|
anchors = torch.tensor(anchors, device=m.anchors.device).type_as(m.anchors) |
|
m.anchors[:] = anchors.clone().view_as(m.anchors) |
|
check_anchor_order(m) |
|
m.anchors /= stride |
|
s = f"{PREFIX}Done ✅ (optional: update model *.yaml to use these anchors in the future)" |
|
else: |
|
s = f"{PREFIX}Done ⚠️ (original anchors better than new anchors, proceeding with original anchors)" |
|
LOGGER.info(s) |
|
|
|
|
|
def kmean_anchors(dataset="./data/coco128.yaml", n=9, img_size=640, thr=4.0, gen=1000, verbose=True): |
|
""" |
|
Creates kmeans-evolved anchors from training dataset. |
|
|
|
Arguments: |
|
dataset: path to data.yaml, or a loaded dataset |
|
n: number of anchors |
|
img_size: image size used for training |
|
thr: anchor-label wh ratio threshold hyperparameter hyp['anchor_t'] used for training, default=4.0 |
|
gen: generations to evolve anchors using genetic algorithm |
|
verbose: print all results |
|
|
|
Return: |
|
k: kmeans evolved anchors |
|
|
|
Usage: |
|
from utils.autoanchor import *; _ = kmean_anchors() |
|
""" |
|
from scipy.cluster.vq import kmeans |
|
|
|
npr = np.random |
|
thr = 1 / thr |
|
|
|
def metric(k, wh): |
|
r = wh[:, None] / k[None] |
|
x = torch.min(r, 1 / r).min(2)[0] |
|
|
|
return x, x.max(1)[0] |
|
|
|
def anchor_fitness(k): |
|
_, best = metric(torch.tensor(k, dtype=torch.float32), wh) |
|
return (best * (best > thr).float()).mean() |
|
|
|
def print_results(k, verbose=True): |
|
k = k[np.argsort(k.prod(1))] |
|
x, best = metric(k, wh0) |
|
bpr, aat = (best > thr).float().mean(), (x > thr).float().mean() * n |
|
s = ( |
|
f"{PREFIX}thr={thr:.2f}: {bpr:.4f} best possible recall, {aat:.2f} anchors past thr\n" |
|
f"{PREFIX}n={n}, img_size={img_size}, metric_all={x.mean():.3f}/{best.mean():.3f}-mean/best, " |
|
f"past_thr={x[x > thr].mean():.3f}-mean: " |
|
) |
|
for x in k: |
|
s += "%i,%i, " % (round(x[0]), round(x[1])) |
|
if verbose: |
|
LOGGER.info(s[:-2]) |
|
return k |
|
|
|
if isinstance(dataset, str): |
|
with open(dataset, errors="ignore") as f: |
|
data_dict = yaml.safe_load(f) |
|
from utils.dataloaders import LoadImagesAndLabels |
|
|
|
dataset = LoadImagesAndLabels(data_dict["train"], augment=True, rect=True) |
|
|
|
|
|
shapes = img_size * dataset.shapes / dataset.shapes.max(1, keepdims=True) |
|
wh0 = np.concatenate([l[:, 3:5] * s for s, l in zip(shapes, dataset.labels)]) |
|
|
|
|
|
i = (wh0 < 3.0).any(1).sum() |
|
if i: |
|
LOGGER.info(f"{PREFIX}WARNING ⚠️ Extremely small objects found: {i} of {len(wh0)} labels are <3 pixels in size") |
|
wh = wh0[(wh0 >= 2.0).any(1)].astype(np.float32) |
|
|
|
|
|
|
|
try: |
|
LOGGER.info(f"{PREFIX}Running kmeans for {n} anchors on {len(wh)} points...") |
|
assert n <= len(wh) |
|
s = wh.std(0) |
|
k = kmeans(wh / s, n, iter=30)[0] * s |
|
assert n == len(k) |
|
except Exception: |
|
LOGGER.warning(f"{PREFIX}WARNING ⚠️ switching strategies from kmeans to random init") |
|
k = np.sort(npr.rand(n * 2)).reshape(n, 2) * img_size |
|
wh, wh0 = (torch.tensor(x, dtype=torch.float32) for x in (wh, wh0)) |
|
k = print_results(k, verbose=False) |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
f, sh, mp, s = anchor_fitness(k), k.shape, 0.9, 0.1 |
|
pbar = tqdm(range(gen), bar_format=TQDM_BAR_FORMAT) |
|
for _ in pbar: |
|
v = np.ones(sh) |
|
while (v == 1).all(): |
|
v = ((npr.random(sh) < mp) * random.random() * npr.randn(*sh) * s + 1).clip(0.3, 3.0) |
|
kg = (k.copy() * v).clip(min=2.0) |
|
fg = anchor_fitness(kg) |
|
if fg > f: |
|
f, k = fg, kg.copy() |
|
pbar.desc = f"{PREFIX}Evolving anchors with Genetic Algorithm: fitness = {f:.4f}" |
|
if verbose: |
|
print_results(k, verbose) |
|
|
|
return print_results(k).astype(np.float32) |
|
|