CE

Mt5 Xxl

CendolGGUF by mradermacherUpdated 2024-12-04
Download on Hugging Face
Parameters12.9B
Context512
Layers24
Architecturet5
Smallest quant5.1 GB
Quick answer

Mt5 Xxl needs about 11.8 GB of memory at Q4_K_M (a 8.1 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 12 quants, from 5.1 GB (Q2_K) to 13.7 GB (Q8_0).

Mt5 Xxl is a 12.9B-parameter Cendol model in GGUF format, quantized by mradermacher. Its trained context length is 512 tokens. The repository has 12 quant files, from 5.1 GB (Q2_K) to 13.7 GB (Q8_0).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K5.1 GB 5.8 GB
Fits Get file →
Q3_K_S6.0 GB 6.7 GB
Fits Get file →
Q3_K_M6.4 GB 7.1 GB
Fits Get file →
Q3_K_L6.8 GB 7.5 GB
Tight Get file →
IQ4_XS7.2 GB 7.9 GB
Tight Get file →
Q4_07.5 GB 8.2 GB
Too big Get file →
Q4_K_S7.6 GB 8.3 GB
Too big Get file →
Q4_K_M8.1 GB 8.8 GB
Too big Get file →
Q5_K_S9.0 GB 9.7 GB
Too big Get file →
Q5_K_M9.3 GB 10.0 GB
Too big Get file →
Q6_K10.6 GB 11.3 GB
Too big Get file →
Q8_013.7 GB 14.4 GB
Too big Get file →

What GPU do you need for Mt5 Xxl?

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: the best quant that fits is Q4_K_M (8.1 GB file, about 11.8 GB in use).
  • 16 GB VRAM: the best quant that fits is Q6_K (10.6 GB file, about 14.3 GB in use).
  • 24 GB VRAM: the best quant that fits is Q8_0 (13.7 GB file, about 17.5 GB in use).

Which Mt5 Xxl quant should you choose?

Q4_K_M is the usual starting point for Mt5 Xxl: it balances file size and quality well. Here it is a 8.1 GB file.

The Q8_0 file is 13.7 GB, about 1.7 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 Q2_K at 5.1 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

Architecturet5
Parameters12.9B
Layers24
Trained context512 tokens
KV cache (fp16)384 KB per token

Source

Quantized bymradermacher
FamilyCendol
Licenseapache-2.0
Downloads (HF)151
Last updated2024-12-04

Frequently asked questions about Mt5 Xxl

How much VRAM does Mt5 Xxl need?

About 11.8 GB at Q4_K_M with an 8K context window. That is the 8.1 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 Mt5 Xxl on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 8.8 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Mt5 Xxl GGUF?

It depends on the quant: Q2_K is 5.1 GB, Q3_K_S is 6.0 GB, Q3_K_M is 6.4 GB, Q3_K_L is 6.8 GB, IQ4_XS is 7.2 GB, Q4_0 is 7.5 GB and 6 more.

What context length does Mt5 Xxl support?

The GGUF header lists a trained context length of 512 tokens (512). The KV cache grows with the context you use: about 3.2 GB at 8K and 0.2 GB at 512.

Where can I download Mt5 Xxl GGUF?

The files are on Hugging Face in the mradermacher/cendol-mt5-xxl-merged-inst-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 Mt5 Xxl use?

Hugging Face lists the license as apache-2.0. 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 mradermacher

All by mradermacher →

More Cendol models

All Cendol →

Other medium models

All medium models →