NE

Nethena MLewd Xwin 23B

NethenaGGUF by TheBlokeUpdated 2023-10-31
Download on Hugging Face
Parameters22.8B
Context4K
Layers71
Architecturellama
Smallest quant9.5 GB
Quick answer

Nethena MLewd Xwin 23B needs about 26.2 GB of memory at Q4_K_M (a 13.8 GB file) with an 8K context window, so it fits on a 32 GB GPU. This repository has 12 quants, from 9.5 GB (Q2_K) to 24.3 GB (Q8_0).

Nethena MLewd Xwin 23B – 22.8B param LLaMA model in GGUF format

Nethena MLewd Xwin 23B is a 22.8B parameter LLaMA model quantised for GGUF. It supports a 4096 token context window and is released under cc‑by‑nc‑4.0. The model is available in a range of quantisations from Q2_K to Q8_0.

Good at

  • Large context reasoning
  • Fine‑grained text generation
  • Low‑latency inference on GPUs

Limits

  • No vision support
  • Requires GPU for best performance
  • cc‑by‑nc‑4.0 license limits commercial use

Best for

  • Chatbots needing long context
  • Story generation with 4096 tokens
  • Research prototyping on local hardware

Who it’s for. Developers and researchers who need a large LLaMA model with flexible quantisation options for local inference. The model is suitable for those who can run it on GPU‑enabled systems and comply with the cc‑by‑nc‑4.0 license.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K9.5 GB 15.9 GB
Too big Get file →
Q3_K_S9.9 GB 16.3 GB
Too big Get file →
Q3_K_M11.1 GB 17.5 GB
Too big Get file →
Q3_K_L12.1 GB 18.6 GB
Too big Get file →
Q4_012.9 GB 19.4 GB
Too big Get file →
Q4_K_S12.9 GB 19.4 GB
Too big Get file →
Q4_K_M13.8 GB 20.2 GB
Too big Get file →
Q5_015.7 GB 22.2 GB
Too big Get file →
Q5_K_S15.7 GB 22.2 GB
Too big Get file →
Q5_K_M16.2 GB 22.6 GB
Too big Get file →
Q6_K18.7 GB 25.2 GB
Too big Get file →
Q8_024.3 GB 30.7 GB
Too big Get file →

For 8 GB VRAM, use Q2_K (9.49 GB file, ~21.9 GB memory at 8 k context). For 16 GB VRAM, Q4_K_M (13.77 GB file, ~26.2 GB memory) is a good balance. For 24 GB VRAM, Q6_K (18.75 GB file, ~31.2 GB memory) offers higher quality.

What GPU do you need for Nethena MLewd Xwin 23B?

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 (11.1 GB file, about 23.5 GB in use).
  • 32 GB VRAM: the best quant that fits is Q6_K (18.7 GB file, about 31.2 GB in use).
  • 48 GB VRAM: the best quant that fits is Q8_0 (24.3 GB file, about 36.7 GB in use).

Which Nethena MLewd Xwin 23B quant should you choose?

Q4_K_M is the usual starting point for Nethena MLewd Xwin 23B: it balances file size and quality well. Here it is a 13.8 GB file.

The Q8_0 file is 24.3 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 9.5 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

Architecturellama
Parameters22.8B
Layers71
Trained context4,096 tokens
KV cache (fp16)1420 KB per token

Source

Quantized byTheBloke
FamilyNethena
Downloads (HF)948
Last updated2023-10-31

Frequently asked questions about Nethena MLewd Xwin 23B

How much VRAM does Nethena MLewd Xwin 23B need?

About 26.2 GB at Q4_K_M with an 8K context window. That is the 13.8 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 Nethena MLewd Xwin 23B on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 21.9 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Nethena MLewd Xwin 23B GGUF?

It depends on the quant: Q2_K is 9.5 GB, Q3_K_S is 9.9 GB, Q3_K_M is 11.1 GB, Q3_K_L is 12.1 GB, Q4_0 is 12.9 GB, Q4_K_S is 12.9 GB and 6 more.

What context length does Nethena MLewd Xwin 23B support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 11.9 GB at 8K and 6.0 GB at 4K.

Where can I download Nethena MLewd Xwin 23B GGUF?

The files are on Hugging Face in the TheBloke/Nethena-MLewd-Xwin-23B-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 Nethena MLewd Xwin 23B 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.

Can I run this model on 8 GB VRAM?

Only the Q2_K quantisation will fit, using about 21.9 GB of memory at 8 k context, which exceeds 8 GB VRAM. It is not feasible on 8 GB VRAM.

Does this model support vision tasks?

No, the model has no vision capabilities.

What license applies to this model?

The model is released under cc‑by‑nc‑4.0, which allows non‑commercial use with attribution.

Other GGUFs of Nethena-MLewd-Xwin-23B

Similar models

Browse all →

More from TheBloke

All by TheBloke →

More Nethena models

All Nethena →

Other medium models

All medium models →