KO

koala-7B-slerp

KoalaGGUF by s3nhUpdated 2024-01-15
Download on Hugging Face
Parameters8.99B
Context32K
Layers40
Architecturellama
Smallest quant3.9 GB
Quick answer

koala-7B-slerp needs about 7.3 GB of memory at Q4_K_M (a 5.4 GB file) with an 8K context window, so it fits on a 8 GB GPU. This repository has 5 quants, from 3.9 GB (Q3_K_S) to 9.6 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 5.4 GB.

koala-7B-slerp is a 8.99B-parameter Koala model in GGUF format, quantized by s3nh. Its trained context length is 32K tokens. The repository has 5 quant files, from 3.9 GB (Q3_K_S) to 9.6 GB (Q8_0).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q3_K_S3.9 GB 5.8 GB
Fits Get file →
Q4_K_M5.4 GB 7.3 GB
Tight Get file →
Q5_K_M6.4 GB 8.2 GB
Too big Get file →
Q6_K7.4 GB 9.2 GB
Too big Get file →
Q8_09.6 GB 11.4 GB
Too big Get file →

What GPU do you need for koala-7B-slerp?

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: the best quant that fits is Q3_K_S (3.9 GB file, about 5.8 GB in use).
  • 8 GB VRAM: the best quant that fits is Q4_K_M (5.4 GB file, about 7.3 GB in use).
  • 12 GB VRAM: the best quant that fits is Q8_0 (9.6 GB file, about 11.4 GB in use).

Which koala-7B-slerp quant should you choose?

Q4_K_M is the usual starting point for koala-7B-slerp: it balances file size and quality well. Here it is a 5.4 GB file.

The Q8_0 file is 9.6 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 Q3_K_S at 3.9 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters8.99B
Layers40
Trained context32,768 tokens
KV cache (fp16)160 KB per token

Source

Quantized bys3nh
FamilyKoala
Licenseopenrail
Downloads (HF)26
Last updated2024-01-15

Frequently asked questions about koala-7B-slerp

How much VRAM does koala-7B-slerp need?

About 7.3 GB at Q4_K_M with an 8K context window. That is the 5.4 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 koala-7B-slerp on an 8 GB GPU?

Yes, with Q4_K_M: the 5.4 GB file needs about 7.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of koala-7B-slerp GGUF?

It depends on the quant: Q3_K_S is 3.9 GB, Q4_K_M is 5.4 GB, Q5_K_M is 6.4 GB, Q6_K is 7.4 GB, Q8_0 is 9.6 GB.

What context length does koala-7B-slerp support?

The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 1.3 GB at 8K and 5.4 GB at 32K.

Where can I download koala-7B-slerp GGUF?

The files are on Hugging Face in the s3nh/koala-7B-slerp-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 koala-7B-slerp use?

Hugging Face lists the license as openrail. 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 s3nh

All by s3nh →

More Koala models

All Koala →

Other small models

All small models →