SakaMoe 3x14B Instruct needs about 29.6 GB of memory at Q4_K_M (a 27.5 GB file) with an 8K context window, so it fits on a 32 GB GPU. This repository has 11 quants, from 16.8 GB (Q2_K) to 48.2 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 6.4 GB.
SakaMoe 3x14B Instruct is a 45.3B-parameter Sakamoe model in GGUF format, quantized by mradermacher. Its trained context length is 128K tokens. The repository has 11 quant files, from 16.8 GB (Q2_K) to 48.2 GB (Q8_0).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 16.8 GB | 19.0 GB | Too big | Get file → | |
| Q3_K_S | 19.8 GB | 21.9 GB | Too big | Get file → | |
| Q3_K_M | 21.9 GB | 24.0 GB | Too big | Get file → | |
| Q3_K_L | 23.7 GB | 25.8 GB | Too big | Get file → | |
| IQ4_XS | 24.6 GB | 26.7 GB | Too big | Get file → | |
| Q4_K_S | 25.9 GB | 28.0 GB | Too big | Get file → | |
| Q4_K_M | 27.5 GB | 29.6 GB | Too big | Get file → | |
| Q5_K_S | 31.3 GB | 33.4 GB | Too big | Get file → | |
| Q5_K_M | 32.2 GB | 34.3 GB | Too big | Get file → | |
| Q6_K | 37.2 GB | 39.3 GB | Too big | Get file → | |
| Q8_0 | 48.2 GB | 50.3 GB | Too big | Get file → |
What GPU do you need for SakaMoe 3x14B Instruct?
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: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: the best quant that fits is Q3_K_M (21.9 GB file, about 24.0 GB in use).
- 32 GB VRAM: the best quant that fits is Q4_K_M (27.5 GB file, about 29.6 GB in use).
- 48 GB VRAM: the best quant that fits is Q6_K (37.2 GB file, about 39.3 GB in use).
Which SakaMoe 3x14B Instruct quant should you choose?
Q4_K_M is the usual starting point for SakaMoe 3x14B Instruct: it balances file size and quality well. Here it is a 27.5 GB file.
The Q8_0 file is 48.2 GB, about 1.8 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 Q2_K at 16.8 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about SakaMoe 3x14B Instruct
How much VRAM does SakaMoe 3x14B Instruct need?
About 29.6 GB at Q4_K_M with an 8K context window. That is the 27.5 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 SakaMoe 3x14B Instruct on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 19.0 GB, so you would need a larger GPU or CPU offloading.
What is the file size of SakaMoe 3x14B Instruct GGUF?
It depends on the quant: Q2_K is 16.8 GB, Q3_K_S is 19.8 GB, Q3_K_M is 21.9 GB, Q3_K_L is 23.7 GB, IQ4_XS is 24.6 GB, Q4_K_S is 25.9 GB and 5 more.
What context length does SakaMoe 3x14B Instruct support?
The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 1.6 GB at 8K and 6.4 GB at 32K.
Where can I download SakaMoe 3x14B Instruct GGUF?
The files are on Hugging Face in the mradermacher/SakaMoe-3x14B-Instruct-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 SakaMoe 3x14B Instruct 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.