保存済みツールを推薦
FOXAI SEARCH

サイト内検索

AI モデル 149

  • tara↗

    Trelis

    **Tara is a frontier automatic speech-recognition model for Hindi and mixed-code (Hinglish) transcription.** On the AI4Bharat Vistaar Hindi benchmark suite it achieves state-of-the-art aggregate accuracy, outperforming leading commercial Hindi ASR systems on the 7-benchmark Vistaar mean, while natively handling Hindi–English code-switched speech through a dedicated **mixed-code mode** that renders English words in Latin script and Hindi in Devanagari, the way real Hinglish is written. The mixed-code mode (the ` ` prefix above) conditions generation only: on pure-Hindi audio it neither degrades accuracy nor forces transliteration; on mixed-code audio it renders English words in Latin script. Evaluation code, the exact text normalizer, and Tara's per-utterance predictions for every benchmark below are published at TrelisResearch/tara, so all numbers can be reproduced or re-scored under alternative normalizers.

  • typhoon-whisper-turbo↗

    typhoon-ai

    **Typhoon Whisper Turbo** is a state-of-the-art Thai Automatic Speech Recognition (ASR) model fine-tuned on the OpenAI Whisper Large v3 Turbo architecture. It is engineered to deliver a balance of high accuracy and low latency for offline transcription, significantly outperforming standard Whisper models in throughput while maintaining robust performance on Thai speech. The model was presented in the paper Typhoon ASR Real-time: FastConformer-Transducer for Thai Automatic Speech Recognition. By using this model, you agree to the OpenTyphoon Terms and Conditions and acknowledge the Privacy Notice: https://opentyphoon.ai/tac · https://opentyphoon.ai/privacy

  • Parable-Granite-4.1-3B-Claude-Fable-5-GGUF↗

    AnkitAI

    > **~3 GB of RAM is all you need.** Laptop, old GPU, Raspberry-Pi-class boxes with swap — the Q4 build runs > anywhere. One command and you have a private, offline reasoning model on your machine: > > v2 is a full retrain: **13× more genuine Fable 5 trace data** (11,574 sessions, 16.8M tokens — corpus published) and a rebuilt recipe (completion-masked loss, replay mixing, benchmark-gated checkpoints, seed-averaged weights). Clean answers, structured reasoning, agent instincts — and the transcript artifacts that leaked into v1's replies are gone. *One trade, made on purpose: raw HumanEval-style function synthesis stays the base model's turf (81.7 vs 70.1) — v2 spends that capacity on agent behavior instead, and spends half as much as v1 did.* Measurement notes below. 👇

  • silero-vad-v6.2-gguf↗

    handy-computer

    Ported from upstream commit 5cd7945, pinned 2026-10-09. Validated against the silero-vad 6.2.3 (TorchScript, CPU, F32) reference at transcribe.cpp commit 2674b95d on 2026-10-10. Voice activity detection by the Silero Team. Not a transcription model: the VAD role produces per-frame speech probabilities (offline or streaming) and offline speech segments. CPU is the default backend for reference fidelity. transcribe.cpp also loads whisper.cpp's `ggml-silero-v5.1.2.bin` / `v6.2.0.bin`. From the C API, use the VAD role (`include/transcribe/vad.h`): offline segments with `transcribe_vad_run`, live per-frame probabilities with `transcribe_vad_stream_feed`.

  • Voxtral-Mini-4B-Realtime-2602↗

    Mistral AI

    Voxtral Mini 4B Realtime 2602 is a **multilingual, realtime speech-transcription model** and among the first open-source solutions to achieve accuracy comparable to offline systems with a delay of ** = 3600 / 0.8 = 45000`. In theory, you should be able to record with no limit; in practice, pre-allocations of RoPE parameters among other things limits `--max-model-len`. For the best user experience, we recommend to simply instantiate vLLM with the default parameters which will automatically set a maximum model length of 131072 (~ca. 3h). - We strongly recommend using websockets to set up audio streaming sessions. For more info on how to do so, check Usage. - We recommend using a delay of 480ms as we found it to be the sweet spot of performance and low latency. If, however, you want to adapt the delay, you can change the `"transcription_delay_ms": 480` parameter in the tekken.json file to any multiple of 80ms between 80 and 1200, as well as 2400 as a standalone value. We compare Voxtral Mini 4B Realtime to similar models - both offline models and realtime. Voxtral Mini 4B Realtime is competitive to leading offline models and shows significant gains over existing open-source realtime solutions. > [!Warning] > Due to its novel architecture, Voxtral Realtime is currently only support in vLLM. We very much welcome community contributions > to add the architecture to Transformers and Llama.cpp.

  • parakeet-primeline-gguf↗

    handy-computer

    GGUF conversions of primeline/parakeet-primeline for use with transcribe.cpp. Ported from upstream commit 3f1a9bc, pinned 2026-08-16. Validated against the NeMo reference at transcribe.cpp commit 856d7c1 on 2026-08-16. primeLine's German fine-tune of NVIDIA's parakeet-tdt-0.6b-v3. A FastConformer encoder with a TDT/RNNT transducer decoder, taking 16 kHz mono WAV and producing a punctuated, cased transcript with optional token-level timestamps. Tuned for German, but the fine-tune did not collapse the base model's multilingual ability: it still transcribes the other 24 v3 languages with correct per-language casing and punctuation. Not a streaming model and does not translate.

  • whisperkit-coreml↗

    argmaxinc

    WhisperKit is part of Argmax OSS, an On-device Speech AI SDK for Apple Silicon: https://github.com/argmaxinc/argmax-oss-swift For real-time transcription with speakers and custom vocabulary, check out Argmax Pro SDK: https://www.argmaxinc.com/blog/argmax-sdk-2

  • llmlingua-2-xlm-roberta-large-meetingbank↗

    Microsoft

    This model was introduced in the paper **LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression** (Pan et al, 2024). It is a XLM-RoBERTa (large-sized model) finetuned to perform token classification for task agnostic prompt compression. The probability $p_{preserve}$ of each token $x_i$ is used as the metric for compression. This model is trained on the extractive text compression dataset constructed with the methodology proposed in the **LLMLingua-2**, using training examples from MeetingBank (Hu et al, 2023) as the seed data. You can evaluate the model on downstream tasks such as question answering (QA) and summarization over compressed meeting transcripts using this dataset. For more details, please check the home page of LLMLingua-2 and LLMLingua Series.

  • parakeet-tdt-1.1b-gguf↗

    handy-computer

    GGUF conversions of nvidia/parakeet-tdt-1.1b for use with transcribe.cpp. Ported from upstream commit 53276c6, pinned 2026-05-10. Validated against the NeMo reference at transcribe.cpp commit 42528dd on 2026-05-10. Offline English speech-to-text. A FastConformer-XL encoder with a TDT/RNNT transducer decoder. Takes a 16 kHz mono WAV and produces a transcript with optional token-level timestamps. Not a streaming model and does not translate.

  • gemma-4-E2B-it-RKLLM↗

    HanzoHuang

    Text generation, image questions, and short audio transcription on Rockchip RK3576 and RK3588, converted from google/gemma-4-E2B-it. **RK3576 preview:** image and audio requests work in the functional tests, but the vision encoder misses the numerical accuracy target. RK3576 W8A8 also fails one exact-recall check. See the measured results below before choosing this variant. RK3588 W8A8 passes the functional and encoder checks performed here. The new matched multimodal set is under **`v1.3.1/`**. Use the exact SoC for your board. Each variant needs its language file and vision encoder; add the audio encoder to enable speech input. The two RK3576 variants share the same encoders.