{"v":1,"id":"model:hf:FunAudioLLM/Fun-ASR-Nano-2512-GGUF","slug":"model-funaudiollm-fun-asr-nano-2512-gguf","kind":"model","category":"speech","title":"Fun-ASR-Nano-2512-GGUF","summary":"Standalone audio.cpp GGUF builds of FunAudioLLM/Fun-ASR-Nano-2512-hf. Each file embeds the model configuration, processor configuration, tokenizer, chat template, and the audio.cpp model package spec…","source":{"provider":"hf","ref":"FunAudioLLM/Fun-ASR-Nano-2512-GGUF","url":"https://huggingface.co/FunAudioLLM/Fun-ASR-Nano-2512-GGUF","rev":"ce72677f84900f0dc57f498ace253bfb3c9155b6","fetchedAt":"2026-10-02T20:59:57.536Z","etag":"W/\"734-6dA8Dydhklul67vZuk5B0/KXuVs\""},"author":{"name":"FunAudioLLM","url":"https://huggingface.co/FunAudioLLM"},"license":{"spdx":null,"raw":"other","url":"https://huggingface.co/FunAudioLLM/Fun-ASR-Nano-2512-hf/blob/main/README.md","open":null,"note":"custom license: read it at the source before installing"},"metrics":{"downloads":63470,"downloadsWeek":15195,"likes":8,"takenAt":"2026-10-02T20:59:57.536Z"},"tags":["audio.cpp","gguf","speech-recognition","multilingual","funasr","automatic-speech-recognition"],"pipeline":"automatic-speech-recognition","links":{"github":"ggml-org/whisper.cpp","npm":"nodejs-whisper"},"updatedAt":"2026-07-29T17:35:39.000Z","collectedAt":"2026-10-02T20:59:57.536Z","review":{"numbers":["63,470 downloads on Hugging Face","8 likes","license other","1.0 GB for fun-asr-nano-2512-q8_0.gguf","15,195 npm downloads a week for nodejs-whisper","latest nodejs-whisper@0.3.1"],"log":null},"trust":"unlabeled","health":{"status":"alive","checkedAt":"2026-10-02T20:59:57.536Z","http":200},"description":"# Fun-ASR-Nano-2512 GGUF\n\nStandalone audio.cpp GGUF builds of\n[FunAudioLLM/Fun-ASR-Nano-2512-hf](https://huggingface.co/FunAudioLLM/Fun-ASR-Nano-2512-hf).\nEach file embeds the model configuration, processor configuration, tokenizer,\nchat template, and the audio.cpp model package specification.\n\n## Files\n\n| File | Size | SHA256 |\n| --- | ---: | --- |\n| `fun-asr-nano-2512-q8_0.gguf` | 1,045,334,432 bytes | `4d727357574b079b7f43336b2930f39da086ca02f5d8d50872090b4c1c3d5e0a` |\n| `fun-asr-nano-2512-f16.gguf` | 1,675,708,832 bytes | `3d906c3ccfed07efef88ff53d6cc94b788b9d2edf1492a5679d041b43e98c5be` |\n\nThe source checkpoint is pinned to revision\n`854d88f94205cd17d2afdb24332130d86fbe654a`. The source\n`model.safetensors` SHA256 is\n`335ca3e74917f1156690400e2c344350112950165789cf78ce3d0a367affd821`.\n\n## audio.cpp\n\n```bash\naudiocpp_cli \\\n  --task asr \\\n  --family fun_asr_nano \\\n  --model fun-asr-nano-2512-q8_0.gguf \\\n  --backend cuda \\\n  --audio speech.wav\n```\n\nFun-ASR-Nano currently provides offline multilingual ASR. It does not expose\nstreaming or timestamp output. On CUDA, audio.cpp keeps the Q8_0 encoder and\nadaptor weights native and loads decoder weights as BF16 by default for stable\nlogits. An explicit `fun_asr_nano.decoder_weight_type` session option overrides\nthat default.\n\n## Reproducibility\n\nThe files were generated with audio.cpp's `audiocpp_gguf` converter:\n\n```bash\naudiocpp_gguf \\\n  --input model.safetensors \\\n  --root /path/to/Fun-ASR-Nano-2512-hf \\\n  --output fun-asr-nano-2512-q8_0.gguf \\\n  --type q8_0 \\\n  --family fun_asr_nano \\\n  --model-spec model_specs/fun_asr_nano.json\n```\n\nBoth formats were checked with `audiocpp_gguf --inspect` and full reference\naudio transcription on CPU and NVIDIA H100 CUDA.\n\n## License\n\nThe original model and these converted weights are governed by the\nFunASR Model Open Source License Agreement v1.1 distributed with the source\nmodel. Review that agreement before using…\n\nSource: https://huggingface.co/FunAudioLLM/Fun-ASR-Nano-2512-GGUF","install":{"kind":"model","hfId":"FunAudioLLM/Fun-ASR-Nano-2512-GGUF","gated":false,"format":"gguf","files":[{"name":"fun-asr-nano-2512-f16.gguf","size":1675708832,"quant":"F16","sha256":"3d906c3ccfed07efef88ff53d6cc94b788b9d2edf1492a5679d041b43e98c5be"},{"name":"fun-asr-nano-2512-q8_0.gguf","size":1045334432,"quant":"Q8_0","sha256":"4d727357574b079b7f43336b2930f39da086ca02f5d8d50872090b4c1c3d5e0a"}],"totalBytes":2721043264,"suggestedFile":"fun-asr-nano-2512-q8_0.gguf","requirements":{"ramGb":2,"diskBytes":1045334432,"note":"estimate: suggested file size × 1.15 + 0.5 GB; a real measurement comes with lsh models"},"runWith":["whisper.cpp"],"command":"lsh models install hf:FunAudioLLM/Fun-ASR-Nano-2512-GGUF"}}