Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

dev(narugo): norm + mean #1

Merged
merged 3 commits into from
Apr 28, 2024
Merged

dev(narugo): norm + mean #1

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
1a02c85
dev(narugo): calculate mean
narugo1992 Apr 28, 2024
328c506
dev(narugo): align the embs norm
narugo1992 Apr 28, 2024
fd7f4cb
dev(narugo): mean better, ci skip
narugo1992 Apr 28, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 4 additions & 19 deletions ccip_merge/measure.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,6 @@
import pandas as pd
from ditk import logging
from imgutils.metrics import ccip_batch_differences, ccip_default_threshold
from scipy.optimize import minimize
from tqdm import tqdm

from .index import get_np_feats
Expand Down Expand Up @@ -35,24 +34,10 @@ def measure_tag_via_func(tag, func):


def ccip_merge_func(embs):
# TODO: just replace your function here!!!
def objective_function(x):
total_similarity = 0
for vector in embs:
total_similarity += np.dot(x, vector) / (np.linalg.norm(x) * np.linalg.norm(vector))
return -total_similarity / embs.shape[0]

initial_guess = np.random.rand(768)
constraints = ({'type': 'eq', 'fun': lambda x: np.linalg.norm(x) - 1.0})
result = minimize(
objective_function, initial_guess,
constraints=constraints,
)

mean_length = np.linalg.norm(embs, axis=1).mean()
best_vec = result.x
best_vec = best_vec * mean_length
return best_vec
lengths = np.linalg.norm(embs, axis=-1)
embs = embs / lengths.reshape(-1, 1)
ret_embedding = embs.mean(axis=0)
return ret_embedding / np.linalg.norm(ret_embedding) * lengths.mean()


def get_metrics_of_tags(n: int = 100) -> pd.DataFrame:
Expand Down
Loading