Jevhakim - a TigreGotico Collection

TigreGotico 's Collections

Onnx feature extractors

Arabic Diacritizers (tashkeel)

Question Classifiers

Grapheme2Phoneme datasets

speechonnxmetrics models

Phoonnx TTS models

Portuguese Heterophone Homographs

Synthetic TTS datasets

TTS Voice Converted Datasets

Synthetic Wake Word Datasets — English

Synthetic "Wake Up" Datasets

Synthetic Wake Word Datasets — Portuguese

Synthetic Wake Word Datasets — Spanish

Portuguese NLP Datasets

Mirandese Datasets

NotWakeWord datasets

voiceclonnx — pure-ONNX voice conversion

ATC-model2vec-role-classification

Jevhakim

updated 2 days ago

Decision models for Arabic, MSA and dialects: Kev fine-tuned on typed decisions from 80 public datasets.

Running on Zero

Agents

Jevhakim

⚖

Typed decisions for Arabic text, MSA and dialects

Note Try the models in the browser: an Arabic state, typed questions, and a probability per option.
TigreGotico/jevhakim-4b

Text Classification • Updated 2 days ago • 19 • 1

Note The larger model: 0.765 accuracy on the Arabic test split, and 0.727 on the 22 evaluation-only sources.
TigreGotico/jevhakim-0.8b

Text Classification • Updated 2 days ago • 14

Note The small model: 0.661 accuracy on the Arabic test split, against 0.503 for stock Kev-0.8B.
TigreGotico/jevhakim-decisions

Updated 2 days ago • 34

Note The training and evaluation data: 442k typed decisions in Arabic from 58 training and 22 evaluation-only public sources.