Modelpublic

SimoneAstarita/it-no-bio-20251014-t13

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes8downloads

Model Card

it-no-bio-20251014-t13

Slur reclamation binary classifier Task: LGBTQ+ reclamation vs non-reclamation use of harmful words on social media text.

Trial timestamp (UTC): 2025-10-14 10:40:27 Data case: it

Configuration (trial hyperparameters)

Model: Alibaba-NLP/gte-multilingual-base

Hyperparameter	Value
LANGUAGES	it
LR	1e-05
EPOCHS	5
MAX_LENGTH	256
USE_BIO	False
USELANGTOKEN	False
GATED_BIO	False
FOCAL_LOSS	True
FOCAL_GAMMA	1.5
USE_SAMPLER	True
R_DROP	True
RKLALPHA	1.0
TEXT_NORMALIZE	True

Dev set results (summary)

Metric	Value
f1macrodev_0.5	0.8966749265256728
f1weighteddev_0.5	0.9343820254908982
accuracydev0.5	0.9325153374233128
f1macrodevbestglobal	0.9043240070436314
f1weighteddevbestglobal	0.9428767256476767
accuracydevbest_global	0.9447852760736196
f1macrodevbestby_lang	0.9043240070436314
f1weighteddevbestby_lang	0.9428767256476767
accuracydevbestbylang	0.9447852760736196
default_threshold	0.5
bestthresholdglobal	0.8
thresholdsbylang	{"it": 0.8}

Thresholds

—Default: 0.5
—Best global: 0.8
—Best by language: { "it": 0.8 }

Detailed evaluation

Classification report @ 0.5

text

              precision    recall  f1-score   support

 no-recl (0)     0.9764    0.9394    0.9575       132
    recl (1)     0.7778    0.9032    0.8358        31

    accuracy                         0.9325       163
   macro avg     0.8771    0.9213    0.8967       163
weighted avg     0.9386    0.9325    0.9344       163

Classification report @ best global threshold (t=0.80)

text

              precision    recall  f1-score   support

 no-recl (0)     0.9489    0.9848    0.9665       132
    recl (1)     0.9231    0.7742    0.8421        31

    accuracy                         0.9448       163
   macro avg     0.9360    0.8795    0.9043       163
weighted avg     0.9440    0.9448    0.9429       163

Classification report @ best per-language thresholds

text

              precision    recall  f1-score   support

 no-recl (0)     0.9489    0.9848    0.9665       132
    recl (1)     0.9231    0.7742    0.8421        31

    accuracy                         0.9448       163
   macro avg     0.9360    0.8795    0.9043       163
weighted avg     0.9440    0.9448    0.9429       163

Per-language metrics (at best-by-lang)

lang	n	acc	f1_macro	f1_weighted	prec_macro	rec_macro	prec_weighted	rec_weighted
it	163	0.9448	0.9043	0.9429	0.9360	0.8795	0.9440	0.9448

Data

—Train/Dev: private multilingual splits with ~15% stratified Dev (by (lang,label)).
—Source: merged EN/IT/ES data with bios retained (ignored if unused by model).

Usage

python

from transformers import AutoTokenizer, AutoModelForSequenceClassification, AutoConfig
import torch, numpy as np

repo = "SimoneAstarita/it-no-bio-20251014-t13"
tok = AutoTokenizer.from_pretrained(repo)
cfg = AutoConfig.from_pretrained(repo)
model = AutoModelForSequenceClassification.from_pretrained(repo)

texts = ["example text ..."]
langs = ["en"]

mode = "best_global"  # or "0.5", "by_lang"

enc = tok(texts, truncation=True, padding=True, max_length=256, return_tensors="pt")
with torch.no_grad():
    logits = model(**enc).logits
probs = torch.softmax(logits, dim=-1)[:, 1].cpu().numpy()

if mode == "0.5":
    th = 0.5
    preds = (probs >= th).astype(int)
elif mode == "best_global":
    th = getattr(cfg, "best_threshold_global", 0.5)
    preds = (probs >= th).astype(int)
elif mode == "by_lang":
    th_by_lang = getattr(cfg, "thresholds_by_lang", {})
    preds = np.zeros_like(probs, dtype=int)
    for lg in np.unique(langs):
        t = th_by_lang.get(lg, getattr(cfg, "best_threshold_global", 0.5))
        preds[np.array(langs) == lg] = (probs[np.array(langs) == lg] >= t).astype(int)
print(list(zip(texts, preds, probs)))

Additional files

reports.json: all metrics (macro/weighted/accuracy) for @0.5, @bestglobal, and @bestbylang. config.json: stores thresholds: defaultthreshold, bestthresholdglobal, thresholdsbylang. postprocessing.json: duplicate threshold info for external tools.