YuE2-3B needs about 4.0 GB of memory at Q4_K_M (a 2.5 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 8 quants, from 0.1 GB (F16) to 7.3 GB (BF16).
YuE2-3B: 0.677B parameter music model in GGUF format
YuE2-3B is a 0.677 billion parameter model for music generation and SheetSage2 transcription. It supports generation, VAE, transcription, and MIDI export. The model is available in multiple quantizations for different VRAM budgets.
Good at
- Music generation
- SheetSage2 transcription
- MIDI export
Limits
- No vision support
- License is cc-by-nc-4.0
- Requires at least 8k context
Best for
- Transcribing songs to lead sheets
- Generating music with score conditioning
- Exporting MIDI parts
Who it’s for. Developers and researchers working on music AI who need a compact model for generation and transcription. The model is suitable for non‑commercial projects under the cc‑by‑nc‑4.0 license.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| F16 | 0.1 GB | 1.5 GB | Fits | Get file → | |
| F16 | 0.1 GB | 1.6 GB | Fits | Get file → | |
| F16 | 0.3 GB | 1.7 GB | Fits | Get file → | |
| F16 | 1.4 GB | 2.8 GB | Fits | Get file → | |
| F16 | 1.4 GB | 2.8 GB | Fits | Get file → | |
| Q4_K_M | 2.5 GB | 4.0 GB | Fits | Get file → | |
| Q8_0 | 3.9 GB | 5.3 GB | Fits | Get file → | |
| BF16 | 7.3 GB | 8.7 GB | Too big | Get file → |
For 8 GB GPUs, use Q4_K_M (file size 2.53 GB, memory 4 GB at 8k context). For 12 GB GPUs, Q8_0 (3.91 GB, 5.3 GB). For 16 GB or higher, BF16 (7.26 GB, 8.7 GB). Smaller VRAM can use F16 variants, but memory rises to 1.5–2.8 GB.
What GPU do you need for YuE2-3B?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) 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 Q4_K_M (2.5 GB file, about 4.0 GB in use).
- 6 GB VRAM: the best quant that fits is Q8_0 (3.9 GB file, about 5.3 GB in use).
- 8 GB VRAM: the best quant that fits is Q8_0 (3.9 GB file, about 5.3 GB in use).
- 12 GB VRAM: the best quant that fits is BF16 (7.3 GB file, about 8.7 GB in use).
Which YuE2-3B quant should you choose?
Q4_K_M is the usual starting point for YuE2-3B: it balances file size and quality well. Here it is a 2.5 GB file.
The Q8_0 file is 3.9 GB, about 1.5 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.
If memory is tight, the smallest file is F16 at 0.1 GB, with the quality trade-offs that come with fewer bits. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about YuE2-3B
How much VRAM does YuE2-3B need?
About 4.0 GB at Q4_K_M with an 8K context window. That is the 2.5 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 YuE2-3B on an 8 GB GPU?
Yes, with Q8_0: the 3.9 GB file needs about 5.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of YuE2-3B GGUF?
It depends on the quant: F16 is 0.1 GB, F16 is 0.1 GB, F16 is 0.3 GB, F16 is 1.4 GB, F16 is 1.4 GB, Q4_K_M is 2.5 GB and 2 more.
What context length does YuE2-3B support?
The GGUF header lists a trained context length of 24,576 tokens (24K). The KV cache grows with the context you use: about 0.9 GB at 8K and 2.8 GB at 24K.
Where can I download YuE2-3B GGUF?
The files are on Hugging Face in the thepatch/YuE2-3B-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 YuE2-3B use?
Hugging Face lists the license as cc-by-nc-4.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Does the model support vision tasks?
No, the model has no vision capabilities.
What license applies to the model?
The model is released under Creative Commons Attribution‑NonCommercial 4.0.