es-no-bio-sweep-20251013-173901-t18
Slur reclamation binary classifier
Task: LGBTQ+ reclamation vs non-reclamation use of harmful words on social media text.
Trial timestamp (UTC): 2025-10-13 17:39:01
Data case:
es
Configuration (trial hyperparameters)
Model: Alibaba-NLP/gte-multilingual-base
| Hyperparameter | Value |
|---|---|
| LANGUAGES | es |
| LR | 1e-05 |
| EPOCHS | 5 |
| MAX_LENGTH | 256 |
| USE_BIO | False |
| USE_LANG_TOKEN | False |
| GATED_BIO | False |
| FOCAL_LOSS | True |
| FOCAL_GAMMA | 2.5 |
| USE_SAMPLER | False |
| R_DROP | True |
| R_KL_ALPHA | 0.5 |
| TEXT_NORMALIZE | True |
Dev set results (summary)
| Metric | Value |
|---|---|
| f1_macro_dev_0.5 | 0.6507936507936507 |
| f1_weighted_dev_0.5 | 0.7946127946127945 |
| accuracy_dev_0.5 | 0.7727272727272727 |
| f1_macro_dev_best_global | 0.6885456885456885 |
| f1_weighted_dev_best_global | 0.8365508365508365 |
| accuracy_dev_best_global | 0.8333333333333334 |
| f1_macro_dev_best_by_lang | 0.6885456885456885 |
| f1_weighted_dev_best_by_lang | 0.8365508365508365 |
| accuracy_dev_best_by_lang | 0.8333333333333334 |
| default_threshold | 0.5 |
| best_threshold_global | 0.55 |
| thresholds_by_lang | {"es": 0.55} |
Thresholds
- Default:
0.5 - Best global:
0.55 - Best by language:
{ "es": 0.55 }
Detailed evaluation
Classification report @ 0.5
precision recall f1-score support
no-recl (0) 0.9184 0.8036 0.8571 112
recl (1) 0.3529 0.6000 0.4444 20
accuracy 0.7727 132
macro avg 0.6357 0.7018 0.6508 132
weighted avg 0.8327 0.7727 0.7946 132
Classification report @ best global threshold (t=0.55)
precision recall f1-score support
no-recl (0) 0.9091 0.8929 0.9009 112
recl (1) 0.4545 0.5000 0.4762 20
accuracy 0.8333 132
macro avg 0.6818 0.6964 0.6885 132
weighted avg 0.8402 0.8333 0.8366 132
Classification report @ best per-language thresholds
precision recall f1-score support
no-recl (0) 0.9091 0.8929 0.9009 112
recl (1) 0.4545 0.5000 0.4762 20
accuracy 0.8333 132
macro avg 0.6818 0.6964 0.6885 132
weighted avg 0.8402 0.8333 0.8366 132
Per-language metrics (at best-by-lang)
| lang | n | acc | f1_macro | f1_weighted | prec_macro | rec_macro | prec_weighted | rec_weighted |
|---|---|---|---|---|---|---|---|---|
| es | 132 | 0.8333 | 0.6885 | 0.8366 | 0.6818 | 0.6964 | 0.8402 | 0.8333 |
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
from transformers import AutoTokenizer, AutoModelForSequenceClassification, AutoConfig
import torch, numpy as np
repo = "SimoneAstarita/es-no-bio-sweep-20251013-173901-t18"
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, @best_global, and @best_by_lang. config.json: stores thresholds: default_threshold, best_threshold_global, thresholds_by_lang. postprocessing.json: duplicate threshold info for external tools.