DeepMount00

Language model for Italian and English.

Language support for English and Italian.

Library name: transformers. License: apache-2.0. Language:

Language support for Italian and English.

Library for natural language processing using transformers with datasets including DeepMount00/Sonnet-3.5-ITA-INSTRUCTION.

Sibilia-TTS Sibilia-TTS è il miglior modello di sintesi vocale testuale (Text-to-Speech, TTS) per la lingua italiana attualmente disponibile su Hugging Face. Il modello è stato progettato per generare una voce naturale, chiara e scorrevole, mantenendo intonazione, prosodia e pronuncia autentiche dell’italiano standard. Caratteristiche principali - Supporto completo alla lingua italiana con attenzione a fonetica e prosodia. - Voce naturale e gradevole, adatta a diversi contesti (assistenti vocali, audiolibri, accessibility, applicazioni educative). - Prestazioni ottimizzate per qualità vocale, stabilità e fluidità. - Addestrato con tecniche avanzate per garantire una resa espressiva vicina al parlato umano. Benchmark Sibilia-TTS ha dimostrato di superare in modo consistente altri modelli italiani di Text-to-Speech: - `fishaudio/openaudio-s1-mini` - `sesame/csm-1b` - `cartesia/azzurra-voice` - `hexgrad/Kokoro-82M` Rispetto a questi modelli, Sibilia-TTS offre: - Maggiore naturalezza e fluidità della voce. - Migliore gestione della prosodia. - Riduzione significativa di artefatti audio e distorsioni. Utilizzi consigliati - Assistenti virtuali e chatbot vocali. - Audiolibri e narrazione automatica. - Applicazioni educative per l’apprendimento della lingua. - Accessibilità per utenti ipovedenti. - Generazione vocale per contenuti multimediali.

Language model for Italian and English.

ModernBERT is a modernized bidirectional encoder-only Transformer model (BERT-style) pre-trained on 2 trillion tokens of English and code data with a native context length of up to 8,192 tokens. ModernBERT leverages recent architectural improvements such as: - Rotary Positional Embeddings (RoPE) for long-context support. - Local-Global Alternating Attention for efficiency on long inputs. - Unpadding and Flash Attention for efficient inference. ModernBERT’s native long context length makes it ideal for tasks that require processing long documents, such as retrieval, classification, and semantic search within large corpora. The model was trained on a large corpus of text and code, making it suitable for a wide range of downstream tasks, including code retrieval and hybrid (text + code) semantic search. - ModernBERT-base - 22 layers, 149 million parameters - ModernBERT-large - 28 layers, 395 million parameters For more information about ModernBERT, we recommend our release blog post for a high-level overview, and our arXiv pre-print for in-depth information. ModernBERT is a collaboration between Answer.AI, LightOn, and friends. You can use these models directly with the `transformers` library. Until the next `transformers` release, doing so requires installing transformers from main: Since ModernBERT is a Masked Language Model (MLM), you can use the `fill-mask` pipeline or load it via `AutoModelForMaskedLM`. To use ModernBERT for downstream tasks like classification, retrieval, or QA, fine-tune it following standard BERT fine-tuning recipes. ⚠️ If your GPU supports it, we recommend using ModernBERT with Flash Attention 2 to reach the highest efficiency. To do so, install Flash Attention as follows, then use the model as normal: Note: ModernBERT does not use token type IDs, unlike some earlier BERT models. Most downstream usage is identical to standard BERT models on the Hugging Face Hub, except you can omit the `tokentypeids` parameter. We evaluate ModernBERT across a range of tasks, including natural language understanding (GLUE), general retrieval (BEIR), long-context retrieval (MLDR), and code retrieval (CodeSearchNet and StackQA). Key highlights: - On GLUE, ModernBERT-base surpasses other similarly-sized encoder models, and ModernBERT-large is second only to Deberta-v3-large. - For general retrieval tasks, ModernBERT performs well on BEIR in both single-vector (DPR-style) and multi-vector (ColBERT-style) settings. - Thanks to the inclusion of code data in its training mixture, ModernBERT as a backbone also achieves new state-of-the-art code retrieval results on CodeSearchNet and StackQA. ModernBERT’s training data is primarily English and code, so performance may be lower for other languages. While it can handle long sequences efficiently, using the full 8,192 tokens window may be slower than short-context inference. Like any large language model, ModernBERT may produce representations that reflect biases present in its training data. Verify critical or sensitive outputs before relying on them. We release the ModernBERT model architectures, model weights, training codebase under the Apache 2.0 license.