LL

Llama 3.2 3B COTv3

LlamaGGUF by ericfloUpdated 2024-12-28
Download on Hugging Face
Parameters3.21B
Context128K
Layers28
Architecturellama
Smallest quant2.0 GB
Quick answer

Llama 3.2 3B COTv3 needs about 3.5 GB of memory at Q4_K_M (a 2.0 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 5 quants, from 2.0 GB (Q4_K_M) to 6.4 GB (F32). Longer context needs more: at 32K tokens the KV cache alone adds about 3.8 GB.

Llama 3.2 3B COTv3 is a 3.21B-parameter Llama model in GGUF format, quantized by ericflo. Its trained context length is 128K tokens. The repository has 5 quant files, from 2.0 GB (Q4_K_M) to 6.4 GB (F32).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K_M2.0 GB 3.5 GB
Fits Get file →
Q6_K2.6 GB 4.1 GB
Fits Get file →
Q8_03.4 GB 4.9 GB
Fits Get file →
BF166.4 GB 7.9 GB
Tight Get file →
F326.4 GB 7.9 GB
Tight Get file →

What GPU do you need for Llama 3.2 3B COTv3?

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: the best quant that fits is Q4_K_M (2.0 GB file, about 3.5 GB in use).
  • 6 GB VRAM: the best quant that fits is Q8_0 (3.4 GB file, about 4.9 GB in use).
  • 8 GB VRAM: the best quant that fits is BF16 (6.4 GB file, about 7.9 GB in use).

Which Llama 3.2 3B COTv3 quant should you choose?

Q4_K_M is the usual starting point for Llama 3.2 3B COTv3: it balances file size and quality well. Here it is a 2.0 GB file.

The Q8_0 file is 3.4 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.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters3.21B
Layers28
Trained context131,072 tokens
KV cache (fp16)112 KB per token

Source

Quantized byericflo
FamilyLlama
Licenseapache-2.0
Downloads (HF)84
Last updated2024-12-28

Frequently asked questions about Llama 3.2 3B COTv3

How much VRAM does Llama 3.2 3B COTv3 need?

About 3.5 GB at Q4_K_M with an 8K context window. That is the 2.0 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 Llama 3.2 3B COTv3 on an 8 GB GPU?

Yes, with BF16: the 6.4 GB file needs about 7.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Llama 3.2 3B COTv3 GGUF?

It depends on the quant: Q4_K_M is 2.0 GB, Q6_K is 2.6 GB, Q8_0 is 3.4 GB, BF16 is 6.4 GB, F32 is 6.4 GB.

What context length does Llama 3.2 3B COTv3 support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 0.9 GB at 8K and 3.8 GB at 32K.

Where can I download Llama 3.2 3B COTv3 GGUF?

The files are on Hugging Face in the ericflo/Llama-3.2-3B-COTv3 repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Llama 3.2 3B COTv3 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.

Other GGUFs of Llama-3.2-3B-Instruct

Similar models

Browse all →

More from ericflo

All by ericflo →

More Llama models

All Llama →

Other tiny models

All tiny models →