MA

Marco o1

MarcoGGUF by ZeroWwUpdated 2024-11-27
Download on Hugging Face
Parameters7.62B
Context32K
Layers28
Architectureqwen2
Smallest quant5.1 GB
Quick answer

Marco o1 needs about 8.5 GB of memory at Q6_K (a 7.5 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 6 quants, from 5.1 GB to 15.2 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 1.9 GB.

Marco o1 is a 7.62B-parameter Marco model in GGUF format, quantized by ZeroWw. Its trained context length is 32K tokens. The repository has 6 quant files, from 5.1 GB to 15.2 GB (F16).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—5.1 GB 6.1 GB
Fits Get file →
Q5_K6.8 GB 7.8 GB
Tight Get file →
Q6_K7.5 GB 8.5 GB
Too big Get file →
—8.1 GB 9.1 GB
Too big Get file →
Q8_09.1 GB 10.1 GB
Too big Get file →
F1615.2 GB 16.2 GB
Too big Get file →

What GPU do you need for Marco o1?

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 Q5_K (6.8 GB file, about 7.8 GB in use).
  • 12 GB VRAM: the best quant that fits is Q8_0 (9.1 GB file, about 10.1 GB in use).
  • 16 GB VRAM: the best quant that fits is Q8_0 (9.1 GB file, about 10.1 GB in use).
  • 24 GB VRAM: the best quant that fits is F16 (15.2 GB file, about 16.2 GB in use).

Which Marco o1 quant should you choose?

Q6_K (7.5 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 the first one in the table at 5.1 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

Architectureqwen2
Parameters7.62B
Layers28
Trained context32,768 tokens
KV cache (fp16)56 KB per token

Source

Quantized byZeroWw
FamilyMarco
Licensemit
Downloads (HF)29
Last updated2024-11-27

Frequently asked questions about Marco o1

How much VRAM does Marco o1 need?

About 8.5 GB at Q6_K with an 8K context window. That is the 7.5 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 Marco o1 on an 8 GB GPU?

Yes, with Q5_K: the 6.8 GB file needs about 7.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Marco o1 GGUF?

It depends on the quant: file is 5.1 GB, Q5_K is 6.8 GB, Q6_K is 7.5 GB, file is 8.1 GB, Q8_0 is 9.1 GB, F16 is 15.2 GB.

What context length does Marco o1 support?

The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 0.5 GB at 8K and 1.9 GB at 32K.

Where can I download Marco o1 GGUF?

The files are on Hugging Face in the ZeroWw/Marco-o1-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 Marco o1 use?

Hugging Face lists the license as mit. 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 ZeroWw

All by ZeroWw →

More Marco models

All Marco →

Other small models

All small models →