MOSS-Transcribe-Diarize 0.9B needs about 1.2 GB of memory at Q4_K_M (a 0.6 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 6 quants, from 0.6 GB (Q4_K_M) to 1.8 GB (F16).
MOSS-Transcribe-Diarize 0.9B – end‑to‑end long‑form audio transcription with speaker diarization
MOSS-Transcribe-Diarize 0.9B is a 0.9B‑parameter audio‑LLM that transcribes and diarizes long audio recordings in a single pass. It supports over 50 languages, 90‑minute recordings, and custom hotword prompting. The model outputs timestamped text with anonymous speaker labels.
Good at
- Long‑form multi‑speaker transcription
- Speaker‑aware diarization in one pass
- Custom hotword prompting
Limits
- No vision support
- Requires 8‑k context for estimates
- License is Apache‑2.0
Best for
- Meeting and call transcription
- Podcast and interview diarization
- Video subtitle generation
Who it’s for. Developers and content creators needing structured, speaker‑aware transcripts from long audio or video files. The model is suitable for applications that require time‑aligned text and speaker attribution without separate diarization pipelines.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_K_M | 0.6 GB | 1.2 GB | Fits | Get file → | |
| Q5_K_M | 0.7 GB | 1.3 GB | Fits | Get file → | |
| Q6_K | 0.8 GB | 1.4 GB | Fits | Get file → | |
| Q8_0 | 1.0 GB | 1.6 GB | Fits | Get file → | |
| BF16 | 1.8 GB | 2.5 GB | Fits | Get file → | |
| F16 | 1.8 GB | 2.6 GB | Fits | Get file → |
For 8 GB VRAM, use Q4_K_M (0.62 GB, ~1.2 GB RAM) or Q5_K_M (0.70 GB, ~1.3 GB RAM). For 12 GB VRAM, Q6_K (0.77 GB, ~1.4 GB RAM) or Q8_0 (0.99 GB, ~1.6 GB RAM) provide higher precision. BF16 and F16 offer the best quality but need ~2.5–2.6 GB RAM.
What GPU do you need for MOSS-Transcribe-Diarize 0.9B?
These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: the best quant that fits is F16 (1.8 GB file, about 2.6 GB in use).
Which MOSS-Transcribe-Diarize 0.9B quant should you choose?
Q4_K_M is the usual starting point for MOSS-Transcribe-Diarize 0.9B: it balances file size and quality well. Here it is a 0.6 GB file.
The Q8_0 file is 1.0 GB, about 1.6 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about MOSS-Transcribe-Diarize 0.9B
How much VRAM does MOSS-Transcribe-Diarize 0.9B need?
About 1.2 GB at Q4_K_M with an 8K context window. That is the 0.6 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run MOSS-Transcribe-Diarize 0.9B on an 8 GB GPU?
Yes, with F16: the 1.8 GB file needs about 2.6 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of MOSS-Transcribe-Diarize 0.9B GGUF?
It depends on the quant: Q4_K_M is 0.6 GB, Q5_K_M is 0.7 GB, Q6_K is 0.8 GB, Q8_0 is 1.0 GB, BF16 is 1.8 GB, F16 is 1.8 GB.
Where can I download MOSS-Transcribe-Diarize 0.9B GGUF?
The files are on Hugging Face in the handy-computer/moss-transcribe-diarize-gguf repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does MOSS-Transcribe-Diarize 0.9B use?
Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run MOSS-Transcribe-Diarize on 8 GB VRAM?
Yes, using Q4_K_M or Q5_K_M quantizations keeps VRAM usage below 8 GB and requires about 1.2–1.3 GB of RAM.
Does the model support video files?
Yes, it accepts audio or video inputs and outputs a speaker‑aware transcript with timestamps.
What license applies to this model?
The model is released under the Apache‑2.0 license.