MA

mamba-2.8b-hf

MambaGGUF by dranger003Updated 2024-03-09
Download on Hugging Face
Parameters2.77B
Context1024K
Layers64
Architecturemamba
Smallest quant2.0 GB
Quick answer

mamba-2.8b-hf needs about 3.5 GB of memory at Q6_K (a 2.7 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 6 quants, from 2.0 GB (Q4_K) to 11.1 GB (F32).

mamba-2.8b-hf is a 2.77B-parameter Mamba model in GGUF format, quantized by dranger003. Its trained context length is 1024K tokens. The repository has 6 quant files, from 2.0 GB (Q4_K) to 11.1 GB (F32).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K2.0 GB 2.8 GB
Fits Get file →
Q5_K2.3 GB 3.1 GB
Fits Get file →
Q6_K2.7 GB 3.5 GB
Fits Get file →
Q8_03.3 GB 4.2 GB
Fits Get file →
F165.8 GB 7.0 GB
Fits Get file →
F3211.1 GB 12.9 GB
Too big Get file →

What GPU do you need for mamba-2.8b-hf?

These figures add the file size, an fp16 KV cache at an 8K context window 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: the best quant that fits is Q6_K (2.7 GB file, about 3.5 GB in use).
  • 6 GB VRAM: the best quant that fits is Q8_0 (3.3 GB file, about 4.2 GB in use).
  • 8 GB VRAM: the best quant that fits is F16 (5.8 GB file, about 7.0 GB in use).
  • 12 GB VRAM: the best quant that fits is F16 (5.8 GB file, about 7.0 GB in use).
  • 16 GB VRAM: the best quant that fits is F32 (11.1 GB file, about 12.9 GB in use).

Which mamba-2.8b-hf quant should you choose?

Q6_K (2.7 GB) is the middle option in this repository. Very close to the original quality for most uses, at a clearly larger size than 4-bit.

If memory is tight, the smallest file is Q4_K at 2.0 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturemamba
Parameters2.77B
Layers64
Trained context1,048,576 tokens

Source

Quantized bydranger003
FamilyMamba
Downloads (HF)154
Last updated2024-03-09

Frequently asked questions about mamba-2.8b-hf

How much VRAM does mamba-2.8b-hf need?

About 3.5 GB at Q6_K with an 8K context window. That is the 2.7 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 mamba-2.8b-hf on an 8 GB GPU?

Yes, with F16: the 5.8 GB file needs about 7.0 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of mamba-2.8b-hf GGUF?

It depends on the quant: Q4_K is 2.0 GB, Q5_K is 2.3 GB, Q6_K is 2.7 GB, Q8_0 is 3.3 GB, F16 is 5.8 GB, F32 is 11.1 GB.

What context length does mamba-2.8b-hf support?

The GGUF header lists a trained context length of 1,048,576 tokens (1024K).

Where can I download mamba-2.8b-hf GGUF?

The files are on Hugging Face in the dranger003/mamba-2.8b-hf-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

Other GGUFs of mamba-2.8b-hf

Similar models

Browse all →

More from dranger003

All by dranger003 →

More Mamba models

All Mamba →

Other tiny models

All tiny models →