GGUF conversions of **AutoArk-AI/ARK-ASR-3B**, a 19-language automatic speech recognition model, for use with **CrispASR** (a ggml/llama.cpp-style C++ engine). > ⚠️ **Experimental / work-in-progress.** Core ASR is validated and accurate (see > below), but the backend has known rough edges — read **Limitations** before relying on it. ARK-ASR-3B = **Whisper-large-v3 encoder with partial rotary position embeddings** → **MLP adapter** → **Qwen2.5-3B decoder** with audio-token injection. A single self-contained GGUF holds the encoder, adapter, and language model.
译文排队中 · 当前显示来源原文
输入 音频→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
分类待补充
输入 待核验→ 输出 待核验
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 音频
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 音频
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 音频→ 输出 待核验
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 向量
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 向量
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 向量
> **Derived from Breeze TTS 2 by BreezeBlue and licensed for research and non-commercial use only.** Architecture: a CSM fork — T5Gemma2 text encoder (26L, bidirectional, symmetric sliding window) → Qwen3 backbone (28L) → depth decoder (12L) over 16 codebooks at 12.5 Hz. Languages: English and Chinese. **The codec is not in these files.** Breeze's bundled audio tokenizer is bit-identical to `Qwen/Qwen3-TTS-Tokenizer-12Hz`, which CrispASR already ships as `cstr/qwen3-tts-tokenizer-12hz-GGUF`. It is wired in as a registry companion and downloaded alongside the model. Roughly 1.16 GB of the original checkpoint (`embed_text_tokens` and a leftover Mimi `codec_model.*`) is unreachable at inference and is dropped by the converter.
译文排队中 · 当前显示来源原文
输入 文本→ 输出 音频
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 音频→ 输出 待核验
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 音频→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
输入 音频→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
分类待补充
输入 待核验→ 输出 待核验
Chatterbox is a full TTS pipeline: character tokenizer → T3 (30-layer Llama AR, 520M) → speech tokens → S3Gen (Conformer encoder + UNet1D CFM denoiser, 10 Euler steps) → HiFTGenerator vocoder (conv chains + Snake activations + iSTFT) → 24 kHz WAV. Distributed under **MIT license**. This is the **multilingual** Chatterbox (23 languages: Arabic, Danish, German, Greek, English, Spanish, Finnish, French, Hebrew, Hindi, Italian, Japanese, Korean, Malay, Dutch, Norwegian, Polish, Portuguese, Russian, Swedish, Swahili, Turkish, Chinese) — the T3 GGUF carries the 2454-token multilingual text tokenizer. Select the language with `-l ` (e.g. `-l de`, `-l ja`); English is the default. Two GGUF files are needed: the **T3 model** (text → speech tokens) and the **S3Gen model** (speech tokens → audio).
译文排队中 · 当前显示来源原文
输入 文本→ 输出 音频
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 音频
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 音频
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 文本→ 输出 文本
GGUF weights for **CohereLabs/cohere-transcribe-03-2026** — Cohere's open-source 2B-parameter ASR model, #1 on the Open ASR Leaderboard (avg WER 5.42, as of March 2026). This conversion enables high-performance CPU inference via **CrispASR** — a crispasr-style C++ runtime for the Cohere Conformer-encoder / Transformer-decoder architecture. `config.json` lists 14 supported languages (`en fr de es it pt nl pl el ar ja zh vi ko`), but the tokenizer carries **all 183 ISO-639-1 ` ` tokens**. So ` ` decodes without error, and the model answers a wrong language *fluently* rather than failing — there is no runtime signal at all. The whitelist therefore has to travel with the weights: these GGUFs carry it as `cohere_transcribe.supported_languages`, and CrispASR substitutes loudly when `-l` names something outside it. For a GGUF converted before that key existed, declare it with `CRISPASR_COHERE_LANGS=en,fr,de,…`.
译文排队中 · 当前显示来源原文
输入 音频→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 音频→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
输入 音频→ 输出 文本
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
输入 文本→ 输出 音频
GGUF conversion of FunAudioLLM/Fun-CosyVoice3-0.5B-2512 for use with CrispASR (`--backend cosyvoice3-tts`). CosyVoice3 is a streaming, multilingual, zero-shot voice-cloning TTS system from Alibaba's FunAudioLLM team. The 0.5B-2512 release is Apache-2.0 licensed and supports **9 languages plus 18 Chinese dialects**. Output is 24 kHz mono. Upstream ships two talker checkpoints in one repo: `llm.pt` (pre-trained) and `llm.rl.pt` (the same architecture after reinforcement learning, tuned by the authors for speech quality, pronunciation accuracy and generation stability). Only the talker differs — **flow, HiFT, CAMPPlus, the speech tokenizer and the voice bank are shared**, so switching is just pointing `-m` at the other LLM GGUF and leaving every companion file in place:
译文排队中 · 当前显示来源原文
输入 文本→ 输出 音频
基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。
基础资料已收录 · 缺失资料后续补充
分类待补充
输入 待核验→ 输出 待核验
Reference-dump archives for the per-backend regression suite in CrispStrobe/CrispASR. See `tests/regression/README.md` in that repo for the full design. Each ` / /ref.gguf` is the output of `tools/dump_reference.py --backend ` against the corresponding source model, capturing the encoder output and a few intermediate stages (per-layer activations for some backends). The C++ runtime's `crispasr-diff ` reads exactly this format and reports per-stage cosine similarity vs the captured tensors. Pinning to revision SHAs from this repo (in `tests/regression/manifest.json`'s `fixtures.revision` field) prevents drift between commits — re-dumping with a different torch / NeMo / transformers version will not silently shift what CI tests against.
译文排队中 · 当前显示来源原文
分类待补充
输入 待核验→ 输出 待核验
