SK

SkunkApe 16b iMatrix

SkunkapeGGUF by MarsupialAIUpdated 2024-04-18
Download on Hugging Face
Parameters16B
Context4K
Layers72
Architecturellama
Smallest quant4.3 GB
Quick answer

SkunkApe 16b iMatrix needs about 12.5 GB of memory at its main file (a 9.6 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 8 quants, from 4.3 GB to 31.9 GB.

SkunkApe 16b iMatrix is a 16B-parameter Skunkape model in GGUF format, quantized by MarsupialAI. Its trained context length is 4K tokens. The repository has 8 quant files, from 4.3 GB to 31.9 GB.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—4.3 GB 6.0 GB
Fits Get file →
—6.2 GB 7.9 GB
Tight Get file →
—8.6 GB 10.3 GB
Too big Get file →
—9.6 GB 11.3 GB
Too big Get file →
—11.3 GB 13.0 GB
Too big Get file →
—13.1 GB 14.8 GB
Too big Get file →
—17.0 GB 18.7 GB
Too big Get file →
—31.9 GB 33.6 GB
Too big Get file →

What GPU do you need for SkunkApe 16b iMatrix?

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: the best quant that fits is the main file (4.3 GB file, about 7.2 GB in use).
  • 12 GB VRAM: the best quant that fits is the main file (8.6 GB file, about 11.5 GB in use).
  • 16 GB VRAM: the best quant that fits is the main file (13.1 GB file, about 16.0 GB in use).
  • 24 GB VRAM: the best quant that fits is the main file (17.0 GB file, about 19.9 GB in use).
  • 32 GB VRAM: the best quant that fits is the main file (17.0 GB file, about 19.9 GB in use).
  • 48 GB VRAM: the best quant that fits is the main file (31.9 GB file, about 34.9 GB in use).

Which SkunkApe 16b iMatrix quant should you choose?

If memory is tight, the smallest file is the first one in the table at 4.3 GB, with the quality trade-offs that come with fewer bits.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters16B
Layers72
Trained context4,096 tokens
KV cache (fp16)288 KB per token

Source

Quantized byMarsupialAI
FamilySkunkape
Downloads (HF)44
Last updated2024-04-18

Frequently asked questions about SkunkApe 16b iMatrix

How much VRAM does SkunkApe 16b iMatrix need?

About 12.5 GB at its main file with an 8K context window. That is the 9.6 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 SkunkApe 16b iMatrix on an 8 GB GPU?

Yes, with the smaller files: the 4.3 GB file needs about 7.2 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of SkunkApe 16b iMatrix GGUF?

It depends on the quant: file is 4.3 GB, file is 6.2 GB, file is 8.6 GB, file is 9.6 GB, file is 11.3 GB, file is 13.1 GB and 2 more.

What context length does SkunkApe 16b iMatrix support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 2.4 GB at 8K and 1.2 GB at 4K.

Where can I download SkunkApe 16b iMatrix GGUF?

The files are on Hugging Face in the MarsupialAI/SkunkApe-16b_iMatrix_GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

Similar models

Browse all →

More from MarsupialAI

All by MarsupialAI →

More Skunkape models

All Skunkape →

Other medium models

All medium models →