GP

GPT-NeoX-20B-Erebus

GptGGUF by YorkieOH10Updated 2024-05-07
Download on Hugging Face
Parameters20.6B
Context2K
Layers44
Architecturegptneox
Smallest quant21.9 GB
Quick answer

GPT-NeoX-20B-Erebus needs about 31.2 GB of memory at Q8_0 (a 21.9 GB file) with an 8K context window, so it fits on a 32 GB GPU.

GPT-NeoX-20B-Erebus is a 20.6B-parameter Gpt model in GGUF format, quantized by YorkieOH10. Its trained context length is 2K tokens. The repository has 1 quant file, from 21.9 GB (Q8_0) to 21.9 GB (Q8_0).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q8_021.9 GB 24.6 GB
Too big Get file →

What GPU do you need for GPT-NeoX-20B-Erebus?

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: the best quant that fits is Q8_0 (21.9 GB file, about 31.2 GB in use).

Which GPT-NeoX-20B-Erebus quant should you choose?

Q8_0 (21.9 GB) is the middle option in this repository. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturegptneox
Parameters20.6B
Layers44
Trained context2,048 tokens
KV cache (fp16)1056 KB per token

Source

Quantized byYorkieOH10
FamilyGpt
Licenseapache-2.0
Downloads (HF)41
Last updated2024-05-07

Frequently asked questions about GPT-NeoX-20B-Erebus

How much VRAM does GPT-NeoX-20B-Erebus need?

About 31.2 GB at Q8_0 with an 8K context window. That is the 21.9 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 GPT-NeoX-20B-Erebus on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 31.2 GB, so you would need a larger GPU or CPU offloading.

What is the file size of GPT-NeoX-20B-Erebus GGUF?

It depends on the quant: Q8_0 is 21.9 GB.

What context length does GPT-NeoX-20B-Erebus support?

The GGUF header lists a trained context length of 2,048 tokens (2K). The KV cache grows with the context you use: about 8.9 GB at 8K and 2.2 GB at 2K.

Where can I download GPT-NeoX-20B-Erebus GGUF?

The files are on Hugging Face in the YorkieOH10/GPT-NeoX-20B-Erebus-Q8_0-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 GPT-NeoX-20B-Erebus 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.

Similar models

Browse all →

More from YorkieOH10

All by YorkieOH10 →

More Gpt models

All Gpt →

Other medium models

All medium models →