KI

KitchenSink 103b

KitchensinkGGUF by bcseUpdated 2024-01-23
Download on Hugging Face
Parameters103B
Context4K
Layers120
Architecturellama
Smallest quant37.9 GB
Quick answer

KitchenSink 103b needs about 46.5 GB of memory at Q3_K_XS (a 42.0 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 4 quants, from 37.9 GB (Q2_K) to 49.6 GB (Q3_K_M).

KitchenSink 103b is a 103B-parameter Kitchensink model in GGUF format, quantized by bcse. Its trained context length is 4K tokens. The repository has 4 quant files, from 37.9 GB (Q2_K) to 49.6 GB (Q3_K_M).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K37.9 GB 40.4 GB
Too big Get file →
Q3_K_XS42.0 GB 44.5 GB
Too big Get file →
Q3_K_S44.5 GB 47.0 GB
Too big Get file →
Q3_K_M49.6 GB 52.1 GB
Too big Get file →

What GPU do you need for KitchenSink 103b?

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 Q3_K_XS (42.0 GB file, about 46.5 GB in use).

Which KitchenSink 103b quant should you choose?

Q3_K_XS (42.0 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.

If memory is tight, the smallest file is Q2_K at 37.9 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
Parameters103B
Layers120
Trained context4,096 tokens
KV cache (fp16)480 KB per token

Source

Quantized bybcse
Licensellama2
Downloads (HF)32
Last updated2024-01-23

Frequently asked questions about KitchenSink 103b

How much VRAM does KitchenSink 103b need?

About 46.5 GB at Q3_K_XS with an 8K context window. That is the 42.0 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 KitchenSink 103b on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 42.4 GB, so you would need a larger GPU or CPU offloading.

What is the file size of KitchenSink 103b GGUF?

It depends on the quant: Q2_K is 37.9 GB, Q3_K_XS is 42.0 GB, Q3_K_S is 44.5 GB, Q3_K_M is 49.6 GB.

What context length does KitchenSink 103b support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 4.0 GB at 8K and 2.0 GB at 4K.

Where can I download KitchenSink 103b GGUF?

The files are on Hugging Face in the bcse/KitchenSink-103b-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 KitchenSink 103b 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.

Similar models

Browse all →

More from bcse

All by bcse →

More Kitchensink models

All Kitchensink →

Other large models

All large models →