AB

Abacus

AbacusGGUF by HIT-SCIRUpdated 2024-09-26
Download on Hugging Face
Parameters2.72B
Context4K
Layers40
Architectureminicpm
Smallest quant1.8 GB
Quick answer

Abacus needs about 5.3 GB of memory at Q4_K_M (a 1.8 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 2 quants, from 1.8 GB (Q4_K_M) to 2.9 GB (Q8_0).

Abacus is a 2.72B-parameter Abacus model in GGUF format, quantized by HIT-SCIR. Its trained context length is 4K tokens. The repository has 2 quant files, from 1.8 GB (Q4_K_M) to 2.9 GB (Q8_0).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K_M1.8 GB 3.8 GB
Fits Get file →
Q8_02.9 GB 4.9 GB
Fits Get file →

What GPU do you need for Abacus?

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 Q4_K_M (1.8 GB file, about 5.3 GB in use).
  • 8 GB VRAM: the best quant that fits is Q8_0 (2.9 GB file, about 6.4 GB in use).

Which Abacus quant should you choose?

Q4_K_M is the usual starting point for Abacus: it balances file size and quality well. Here it is a 1.8 GB file.

The Q8_0 file is 2.9 GB, about 1.6 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.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureminicpm
Parameters2.72B
Layers40
Trained context4,096 tokens
KV cache (fp16)360 KB per token

Source

Quantized byHIT-SCIR
FamilyAbacus
Downloads (HF)27
Last updated2024-09-26

Frequently asked questions about Abacus

How much VRAM does Abacus need?

About 5.3 GB at Q4_K_M with an 8K context window. That is the 1.8 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 Abacus on an 8 GB GPU?

Yes, with Q8_0: the 2.9 GB file needs about 6.4 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Abacus GGUF?

It depends on the quant: Q4_K_M is 1.8 GB, Q8_0 is 2.9 GB.

What context length does Abacus support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 3.0 GB at 8K and 1.5 GB at 4K.

Where can I download Abacus GGUF?

The files are on Hugging Face in the HIT-SCIR/Abacus-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 HIT-SCIR

All by HIT-SCIR →

More Abacus models

All Abacus →

Other tiny models

All tiny models →