alt.freeweights.vision / distilbert-multilingual-toxicity-classifier
From: gravitee-io <huggingface.co/gravitee-io>Newsgroups: alt.freeweights.visionSubject:

distilbert-multilingual-toxicity-classifier

Date: 2026-11-15Commit: 865ad1bSigned: 2DC28CC7B26A283B ✓

Text classification model with 135M parameters. Onnx and safetensors weights, 1.22 GB in all. Published by gravitee-io under OpenRAIL++-M (custom).

135M params1.22GBdistilbert-multilingual-toxicity-classifieronnx safetensorscustom terms
FactsParams135MSize1.22 GBFormatsonnx, safetensorsTasktext classificationLicenseOpenRAIL++-M (custom)From sourcecustom termsread terms
# custom terms: read the license before you download
$ fw get gravitee-io/distilbert-multilingual-toxicity-classifier
# manifest sha256 5e59bcd51f59… · 1.22 GB · 11 files
$ fw verify ./distilbert-multilingual-toxicity-classifier gravitee-io/distilbert-multilingual-toxicity-classifier
From sourceHosted (signed)gravitee-io/distilbert-multilingual-toxicity-classifierLast check: 2026-11-15 04:56 UTC

Get help: custom fine-tune or private deploy. We review your setup before we quote.

Quantized of distilbert-base-multilingual-cased.

Overview

Size
1.22 GB
Weights
541 MB (safetensors)
Parameters
135M
Formats
onnx, safetensors
Task
text classification
License
OpenRAIL++-M (custom) (custom)
Files
11, each with a sha256
Publisher
gravitee-io
Added
2026-11-15
Section
alt.freeweights.vision
Lineage

Original release from the publisher.

Base model: distilbert/distilbert-base-multilingual-cased

Where these bytes come from
Publisher at the pinned commit→You

Sources are tried in order. A file counts only when its size and sha256 match the manifest. Sources are checked when you ask for them, not on a timer.

Sources
BadgeHostLast checkResult
From sourcehuggingface.co2026-11-15 04:56 UTCpassed

Model card

From the publisher at commit 865ad1b. Shown as plain text with links; images open at the publisher.

distillbert-multilingual-toxicity-classifier#

This is a DistilBERT-based multilingual toxicity classifier fine-tuned using the gravitee-io/textdetox-multilingual-toxicity-dataset. The model supports a wide range of languages and is trained for toxicity classification ("not-toxic", "toxic").

We perform an 85/15 train-test split per language based on the textdetox dataset. All credits go to the authors of the original corpora.

Performance Overview#

Original model#
LanguageValidation F1Training F1ΔF1
Russian0.95720.9897-0.0324
English0.95280.9853-0.0325
Hindi0.92480.9599-0.0351
Armenian0.65130.6915-0.0402
French0.94460.9874-0.0428
Tatar0.92000.9682-0.0482
Ukrainian0.89970.9511-0.0514
Japanese0.86580.9253-0.0595
German0.89040.9547-0.0643
Spanish0.85640.9399-0.0835
Chinese0.68650.7807-0.0942
Arabic0.75630.8550-0.0987
Italian0.82230.9271-0.1048
Hinglish0.72340.8533-0.1299
Hebrew0.64550.8441-0.1987
Quantized model (ONNX)#
LanguageVal F1Quantized Val F1Δ Val F1Train F1Quantized Train F1Δ Train F1
Russian0.95720.9609+0.00370.98970.9875−0.0022
English0.95280.9495−0.00330.98530.9857+0.0004
German0.89040.8842−0.00620.95470.9369−0.0178
Hindi0.92480.8940−0.03000.95990.9335−0.0264
French0.94460.9351−0.00950.98740.9814−0.0060
Japanese0.86580.8584−0.00740.92530.9081−0.0172
Ukrainian0.89970.8988−0.00090.95110.9476−0.0035
Tatar0.92000.9148−0.00520.96820.9631−0.0051
Amharic0.65130.6377−0.01360.69150.6863−0.0052
Spanish0.85640.8439−0.01250.93990.9273−0.0126
Chinese0.68650.6697−0.01680.78070.7596−0.0211
Arabic0.75630.7535−0.00280.85500.8481−0.0069
Italian0.82230.8033−0.01900.92710.9193−0.0078
Hinglish0.72340.7260+0.00260.85330.8436−0.0097
Hebrew0.64550.6190−0.02650.84410.8204−0.0237

The quantized model maintains high accuracy and generalization on the evaluation set, with negligible performance loss for most languages.

🤗 Usage#

from transformers import AutoTokenizer
from optimum.onnxruntime import ORTModelForSequenceClassification
import numpy as np

# Load model and tokenizer using optimum
model = ORTModelForSequenceClassification.from_pretrained(
 "gravitee-io/distilbert-multilingual-toxicity-classifier",
 file_name="model.quant.onnx"
)
tokenizer = AutoTokenizer.from_pretrained("gravitee-io/distilbert-multilingual-toxicity-classifier")

# Tokenize input
text = "Your text here"
inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True)

# Run inference
outputs = model(**inputs)
logits = outputs.logits

# Optional: convert to probabilities
probs = 1 / (1 + np.exp(-logits))
print(probs)

Github Repository#

You can check details on how the model was fine-tuned and evaluated on the Github Repository

License#

This model is licensed under OpenRAIL++

Citation#

@inproceedings{dementieva2024overview,
  title={Overview of the Multilingual Text Detoxification Task at PAN 2024},
  author={Dementieva, Daryna and Moskovskiy, Daniil and Babakov, Nikolay and Ayele, Abinew Ali and Rizwan, Naquee and Schneider, Frolian and Wang, Xintog and Yimam, Seid Muhie and Ustalov, Dmitry and Stakovskii, Elisei and Smirnova, Alisa and Elnagar, Ashraf and Mukherjee, Animesh and Panchenko, Alexander},
  booktitle={Working Notes of CLEF 2024 - Conference and Labs of the Evaluation Forum},
  editor={Guglielmo Faggioli and Nicola Ferro and Petra Galu{{s}}{{c}}{'a}kov{'a} and Alba Garc{'i}a Seco de Herrera},
  year={2024},
  organization={CEUR-WS.org}
}

@inproceedings{dementieva-etal-2024-toxicity,
  title = "Toxicity Classification in {U}krainian",
  author = "Dementieva, Daryna and Khylenko, Valeriia and Babakov, Nikolay and Groh, Georg",
  booktitle = "Proceedings of the 8th Workshop on Online Abuse and Harms (WOAH 2024)",
  month = jun,
  year = "2024",
  address = "Mexico City, Mexico",
  publisher = "Association for Computational Linguistics",
  url = "https://aclanthology.org/2024.woah-1.19/",
  doi = "10.18653/v1/2024.woah-1.19",
  pages = "244--255"
}

@inproceedings{DBLP:conf/ecir/BevendorffCCDEFFKMMPPRRSSSTUWZ24,
  author = {Janek Bevendorff and et al.},
  title = {Overview of {PAN} 2024: Multi-author Writing Style Analysis, Multilingual Text Detoxification, Oppositional Thinking Analysis, and Generative {AI} Authorship Verification - Extended Abstract},
  booktitle = {ECIR 2024, Glasgow, UK, March 24-28, 2024, Proceedings, Part {VI}},
  series = {Lecture Notes in Computer Science},
  volume = {14613},
  pages = {3--10},
  publisher = {Springer},
  year = {2024},
  doi = {10.1007/978-3-031-56072-9_1}
}

Files

11 files · 1.22 GBWeights 541 MB (safetensors). Sizes and sha256 from the signed manifest.
Hide all 11 files with their sha256
FileSizesha256
.gitattributes1.52 KB11ad7efa24975ee4b0c3c3a38ed18737f0658a5f75a0a96787b576a78a023361
LICENSE14.1 KB5f44a64473fdb1019fc616d999d351f60bf061a4d77272b4b7309f799dba3676
README.md6.63 KB3673b2900c4af7d11c2b084209b14e4dfc746809e72e9adc0036174ac6be60d4
config.json711 B7e96ac0dba63fb0bfe1f65fa533e7e0720b609fe88ed658050f01a9a13e4bccf
model.onnx541 MBbee01a9d2d474a589c5db631d99c382d5502bb81a0fe318d247f46c16b25f060
model.quant.onnx136 MBc8169782758fe097c6cf036ddc48f6ff7b3c861764d6124c20d8b31bc0ba93d7
model.safetensors541 MBb78c8a01991037caecadaa5055092e64923b52286612e87ba32e9d36f8209630
special_tokens_map.json125 Bb6d346be366a7d1d48332dbc9fdf3bf8960b5d879522b7799ddba59e76237ee3
tokenizer.json2.92 MBc2c0e4d392f08c2c2eb950156eb26e4e5ddd15b557faa45ed6a0decaf45b5f98
tokenizer_config.json1.23 KB91955163d30880d1ef0928cfdd729baaf413401dc4456f3ab1a48b3eec2ec624
vocab.txt996 KBfe0fda7c425b48c516fc8f160d594c8022a0808447475c1a7c6d6479763f310c
VersionsEach time the pinned upstream commit or file list changed in this repo. Newest first.
Upstream commitDateContents
865ad1b453 pinned now2026-11-15 04:56 UTC11 files · 1.22 GB

Use it

# custom terms: read the license before you download
$ fw get gravitee-io/distilbert-multilingual-toxicity-classifier
# manifest sha256 5e59bcd51f59… · 1.22 GB · 11 files
$ fw verify ./distilbert-multilingual-toxicity-classifier gravitee-io/distilbert-multilingual-toxicity-classifier
Terms that come with these files. OpenRAIL++-M: the use restrictions in Attachment A pass on to you and to anyone you share the model with. Keep a copy of the license with the files.
Custom license (custom). Read the license text first. Open this panel to show the commands once you accept its terms.
fw.pyChecks the signature and every sha256, tries each source in order, resumes.
$ python3 fw.py get https://freeweights.allternit.com/manifests/gravitee-io/distilbert-multilingual-toxicity-classifier.json --dir ./distilbert-multilingual-toxicity-classifier $ python3 fw.py verify ./distilbert-multilingual-toxicity-classifier ./gravitee-io__distilbert-multilingual-toxicity-classifier.json
hfThe Hugging Face CLI with one setting. Same repo id.
$ HF_ENDPOINT=https://freeweights.allternit.com hf download gravitee-io/distilbert-multilingual-toxicity-classifier --revision 865ad1b453a49e4a7f2471fc40b727ececd28314 --local-dir ./distilbert-multilingual-toxicity-classifier

License

Name
CreativeML Open RAIL++-M
Id
LicenseRef-OpenRAIL++-M
Class
Custom terms
Redistribution
click-through
Listing
Hosted (signed)
Terms
OpenRAIL++-M: the use restrictions in Attachment A pass on to you and to anyone you share the model with. Keep a copy of the license with the files.
Decided
License as declared by the publisher. No person read it. 2026-11-15
Text in
LICENSE
Text sha256
5f44a64473fdb1019fc616d999d351f60bf061a4d77272b4b7309f799dba3676
Terms gate
Yes. The commands stay folded until you open them after reading the license.

Read the license text

Licenses are checked against the license table. A custom license passes only after a person reads it.

Provenance

From: gravitee-io <huggingface.co/gravitee-io>Newsgroups: alt.freeweights.visionSubject: distilbert-multilingual-toxicity-classifierDate: 2026-11-15Commit: 865ad1bSigned: 2DC28CC7B26A283B ✓
Upstream
gravitee-io/distilbert-multilingual-toxicity-classifier
Pinned commit
865ad1b453a49e4a7f2471fc40b727ececd28314
Manifest
distilbert-multilingual-toxicity-classifier.json · .minisig
Manifest sha256
5e59bcd51f5960a307ee176b6e3b8e83038d45af569c404fca8302970206f2ed
Signature
minisign, Ed25519, key 2DC28CC7B26A283B
Created
2026-11-15 04:56 UTC
Generator
freeweights-manifest/0.1.0
Takedowns
None.

How to verify

$ curl -O https://freeweights.allternit.com/manifests/gravitee-io/distilbert-multilingual-toxicity-classifier.json $ curl -O https://freeweights.allternit.com/manifests/gravitee-io/distilbert-multilingual-toxicity-classifier.json.minisig $ minisign -Vm distilbert-multilingual-toxicity-classifier.json -P RWQ7KGqyx4zCLXtUFOqvfv8L3g+8fpKJ1FsuxPyOYMF+4KufbUxKq1Oq $ shasum -a 256 <file> # compare with the Files tab

A manifest whose signature fails must be rejected. A file whose size or sha256 does not match must be discarded. Verify a download

Model tree

Parent

distilbert/distilbert-base-multilingual-cased · this model is a quantization of it

1 model in the catalog name this one as their parent.

Quantizations
1
Get it
Read license, then download
$ HF_ENDPOINT=https://freeweights.allternit.com hf download gravitee-io/distilbert-multilingual-toxicity-classifier
Open in Freeweights Studio

Pull, verify and run it in the Allternit app on Desktop, web or phone.

Signed manifest
minisign, key 2DC28CC7B26A283B. distilbert-multilingual-toxicity-classifier.json. Provenance. Takedown history: none.

Variants

1 quantized or adapter build of distilbert-multilingual-toxicity-classifier. Each has its own license check and page.

From: dima806 <huggingface.co/dima806>Subject: ai_vs_real_image_detection
From: legekka <huggingface.co/legekka>Subject: AI-Anime-Image-Detector-ViT
From: onnx-community <huggingface.co/onnx-community>Subject: AI-Anime-Image-Detector-ViT-ONNX
From: AICodexLab <huggingface.co/AICodexLab>Subject: answerdotai-ModernBERT-base-ai-detector
From: onnx-community <huggingface.co/onnx-community>Subject: answerdotai-ModernBERT-base-ai-detector-ONNX
From: onnx-community <huggingface.co/onnx-community>Subject: bart-large-cnn-ONNX