AY

Ayla Light Extra v2

AylaGGUF by mradermacherUpdated 2025-02-19
Download on Hugging Face
Parameters18.8B
Context1000K
Layers64
Architecturellama
Smallest quant7.2 GB
Quick answer

Ayla Light Extra v2 needs about 14.0 GB of memory at Q4_K_M (a 11.4 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 11 quants, from 7.2 GB (Q2_K) to 20.0 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 8.6 GB.

Ayla Light Extra v2 is a 18.8B-parameter Ayla model in GGUF format, quantized by mradermacher. Its trained context length is 1000K tokens. The repository has 11 quant files, from 7.2 GB (Q2_K) to 20.0 GB (Q8_0).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K7.2 GB 9.8 GB
Too big Get file →
Q3_K_S8.3 GB 11.0 GB
Too big Get file →
Q3_K_M9.2 GB 11.9 GB
Too big Get file →
Q3_K_L10.0 GB 12.6 GB
Too big Get file →
IQ4_XS10.3 GB 13.0 GB
Too big Get file →
Q4_K_S10.8 GB 13.5 GB
Too big Get file →
Q4_K_M11.4 GB 14.0 GB
Too big Get file →
Q5_K_S13.0 GB 15.7 GB
Too big Get file →
Q5_K_M13.4 GB 16.0 GB
Too big Get file →
Q6_K15.4 GB 18.1 GB
Too big Get file →
Q8_020.0 GB 22.6 GB
Too big Get file →

What GPU do you need for Ayla Light Extra v2?

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 Q3_K_M (9.2 GB file, about 11.9 GB in use).
  • 16 GB VRAM: the best quant that fits is Q5_K_M (13.4 GB file, about 16.0 GB in use).
  • 24 GB VRAM: the best quant that fits is Q8_0 (20.0 GB file, about 22.6 GB in use).

Which Ayla Light Extra v2 quant should you choose?

Q4_K_M is the usual starting point for Ayla Light Extra v2: it balances file size and quality well. Here it is a 11.4 GB file.

The Q8_0 file is 20.0 GB, about 1.8 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 7.2 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
Parameters18.8B
Layers64
Trained context1,024,000 tokens
KV cache (fp16)256 KB per token

Source

Quantized bymradermacher
FamilyAyla
Downloads (HF)48
Last updated2025-02-19

Frequently asked questions about Ayla Light Extra v2

How much VRAM does Ayla Light Extra v2 need?

About 14.0 GB at Q4_K_M with an 8K context window. That is the 11.4 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 Ayla Light Extra v2 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 9.8 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Ayla Light Extra v2 GGUF?

It depends on the quant: Q2_K is 7.2 GB, Q3_K_S is 8.3 GB, Q3_K_M is 9.2 GB, Q3_K_L is 10.0 GB, IQ4_XS is 10.3 GB, Q4_K_S is 10.8 GB and 5 more.

What context length does Ayla Light Extra v2 support?

The GGUF header lists a trained context length of 1,024,000 tokens (1000K). The KV cache grows with the context you use: about 2.1 GB at 8K and 8.6 GB at 32K.

Where can I download Ayla Light Extra v2 GGUF?

The files are on Hugging Face in the mradermacher/Ayla-Light-Extra-v2-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 Ayla-Light-Extra-v2

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Ayla models

All Ayla →

Other medium models

All medium models →