CO

CohereForAI aya 23 35B

CohereforaiGGUF by RichardErkhovUpdated 2024-06-01
Download on Hugging Face
Parameters35B
Context—
Layers—
Architecture—
Smallest quant0.0 GB
Quick answer

CohereForAI aya 23 35B needs about 0.5 GB of memory at Q4_K_M (a 0.0 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 22 quants, from 0.0 GB (IQ4_NL) to 16.7 GB (IQ3_M).

Aya‑23‑35B: a 35‑billion‑parameter multilingual instruction model

Aya‑23‑35B is a 35‑billion‑parameter auto‑regressive language model. It supports 23 languages and is fine‑tuned for instruction following. The model is available in multiple GGUF quantizations for local deployment.

Good at

  • Multilingual instruction following
  • Large‑scale text generation
  • Supports 23 languages

Limits

  • No vision support
  • License CC‑BY‑NC‑4.0
  • Requires 8k context for memory estimates

Best for

  • Developing multilingual chatbots
  • Fine‑tuning on custom instruction data
  • Running local inference on powerful GPUs

Who it’s for. Developers and researchers needing a large multilingual model for local inference. The model is suitable for those who can comply with the CC‑BY‑NC‑4.0 license and have sufficient GPU memory.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ4_NL0.0 GB 0.5 GB
Fits Get file →
Q4_10.0 GB 0.5 GB
Fits Get file →
Q4_K0.0 GB 0.5 GB
Fits Get file →
Q4_K_M0.0 GB 0.5 GB
Fits Get file →
Q4_K_S0.0 GB 0.5 GB
Fits Get file →
Q5_00.0 GB 0.5 GB
Fits Get file →
Q5_10.0 GB 0.5 GB
Fits Get file →
Q5_K0.0 GB 0.5 GB
Fits Get file →
Q5_K_M0.0 GB 0.5 GB
Fits Get file →
Q5_K_S0.0 GB 0.5 GB
Fits Get file →
Q6_K0.0 GB 0.5 GB
Fits Get file →
Q8_00.0 GB 0.5 GB
Fits Get file →
Q4_00.0 GB 0.5 GB
Fits Get file →
IQ4_XS0.7 GB 1.2 GB
Fits Get file →
Q3_K_L1.5 GB 2.2 GB
Fits Get file →
Q3_K_M2.6 GB 3.4 GB
Fits Get file →
Q3_K7.7 GB 9.1 GB
Too big Get file →
Q3_K_S9.9 GB 11.6 GB
Too big Get file →
Q2_K13.8 GB 16.0 GB
Too big Get file →
IQ3_XS15.1 GB 17.4 GB
Too big Get file →
IQ3_S15.9 GB 18.3 GB
Too big Get file →
IQ3_M16.7 GB 19.2 GB
Too big Get file →

For 8 GB VRAM, use Q4_0 (0.01 GB) or IQ4_NL (0.5 GB). For 16 GB VRAM, Q3_K_M (2.59 GB) or Q3_K (7.69 GB) are suitable. For 32 GB VRAM, Q3_K_S (9.91 GB) or Q2_K (13.82 GB) provide higher quality. Memory estimates are approximate.

What GPU do you need for CohereForAI aya 23 35B?

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 Q3_K_M (2.6 GB file, about 3.4 GB in use).
  • 6 GB VRAM: the best quant that fits is Q3_K_M (2.6 GB file, about 3.4 GB in use).
  • 8 GB VRAM: the best quant that fits is Q3_K_M (2.6 GB file, about 3.4 GB in use).
  • 12 GB VRAM: the best quant that fits is Q3_K_S (9.9 GB file, about 11.6 GB in use).
  • 16 GB VRAM: the best quant that fits is Q2_K (13.8 GB file, about 16.0 GB in use).
  • 24 GB VRAM: the best quant that fits is IQ3_M (16.7 GB file, about 19.2 GB in use).

Which CohereForAI aya 23 35B quant should you choose?

Q4_K_M is the usual starting point for CohereForAI aya 23 35B: it balances file size and quality well. Here it is a 0.0 GB file.

The Q8_0 file is 0.0 GB, about NaN 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 IQ4_NL at 0.0 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

Parameters35B

Source

Quantized byRichardErkhov
Downloads (HF)5,829
Last updated2024-06-01

Frequently asked questions about CohereForAI aya 23 35B

How much VRAM does CohereForAI aya 23 35B need?

About 0.5 GB at Q4_K_M with an 8K context window. That is the 0.0 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 CohereForAI aya 23 35B on an 8 GB GPU?

Yes, with Q3_K_M: the 2.6 GB file needs about 3.4 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of CohereForAI aya 23 35B GGUF?

It depends on the quant: IQ4_NL is 0.0 GB, Q4_1 is 0.0 GB, Q4_K is 0.0 GB, Q4_K_M is 0.0 GB, Q4_K_S is 0.0 GB, Q5_0 is 0.0 GB and 16 more.

Where can I download CohereForAI aya 23 35B GGUF?

The files are on Hugging Face in the RichardErkhov/CohereForAI_-_aya-23-35B-gguf repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

Can I run Aya‑23‑35B on 8 GB VRAM?

Use the Q4_0 or IQ4_NL quantization; each requires about 0.5 GB of VRAM at 8k context.

Does Aya‑23‑35B support image input?

No, the model accepts text only and does not provide vision capabilities.

What license applies to Aya‑23‑35B?

The model is released under CC‑BY‑NC‑4.0 and requires adherence to C4AI's Acceptable Use Policy.

Similar models

Browse all →

More from RichardErkhov

All by RichardErkhov →

Other large models

All large models →