Stheno-Mega-False-49B-L2 needs about 56.3 GB of memory at Q4_K_M (a 29.8 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 13 quants, from 18.3 GB (Q2_K) to 40.6 GB (Q6_K).
Stheno-Mega-False-49B-L2 is a 49.5B-parameter Stheno model in GGUF format, quantized by mradermacher. Its trained context length is 4K tokens. The repository has 13 quant files, from 18.3 GB (Q2_K) to 40.6 GB (Q6_K).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 18.3 GB | 31.8 GB | Too big | Get file → | |
| IQ3_XS | 20.2 GB | 33.7 GB | Too big | Get file → | |
| IQ3_S | 21.3 GB | 34.8 GB | Too big | Get file → | |
| Q3_K_S | 21.3 GB | 34.8 GB | Too big | Get file → | |
| IQ3_M | 22.6 GB | 36.1 GB | Too big | Get file → | |
| Q3_K_M | 24.0 GB | 37.5 GB | Too big | Get file → | |
| Q3_K_L | 26.3 GB | 39.8 GB | Too big | Get file → | |
| IQ4_XS | 26.6 GB | 40.1 GB | Too big | Get file → | |
| Q4_K_S | 28.1 GB | 41.6 GB | Too big | Get file → | |
| Q4_K_M | 29.8 GB | 43.3 GB | Too big | Get file → | |
| Q5_K_S | 34.1 GB | 47.6 GB | Too big | Get file → | |
| Q5_K_M | 35.0 GB | 48.5 GB | Too big | Get file → | |
| Q6_K | 40.6 GB | 54.1 GB | Too big | Get file → |
What GPU do you need for Stheno-Mega-False-49B-L2?
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: no quant fits at an 8K context.
- 32 GB VRAM: no quant fits at an 8K context.
- 48 GB VRAM: the best quant that fits is IQ3_S (21.3 GB file, about 47.8 GB in use).
Which Stheno-Mega-False-49B-L2 quant should you choose?
Q4_K_M is the usual starting point for Stheno-Mega-False-49B-L2: it balances file size and quality well. Here it is a 29.8 GB file.
If memory is tight, the smallest file is Q2_K at 18.3 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 Stheno-Mega-False-49B-L2
How much VRAM does Stheno-Mega-False-49B-L2 need?
About 56.3 GB at Q4_K_M with an 8K context window. That is the 29.8 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 Stheno-Mega-False-49B-L2 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 44.8 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Stheno-Mega-False-49B-L2 GGUF?
It depends on the quant: Q2_K is 18.3 GB, IQ3_XS is 20.2 GB, IQ3_S is 21.3 GB, Q3_K_S is 21.3 GB, IQ3_M is 22.6 GB, Q3_K_M is 24.0 GB and 7 more.
What context length does Stheno-Mega-False-49B-L2 support?
The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 26.0 GB at 8K and 13.0 GB at 4K.
Where can I download Stheno-Mega-False-49B-L2 GGUF?
The files are on Hugging Face in the mradermacher/Stheno-Mega-False-49B-L2-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 Stheno-Mega-False-49B-L2 use?
Hugging Face lists the license as llama2. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.