LL

Llama-3-3x8B-multilingual

LlamaGGUF by mradermacherUpdated 2024-08-02
Download on Hugging Face
Parameters19.3B
Context80K
Layers32
Architecturellama
Smallest quant4.3 GB
Quick answer

Llama-3-3x8B-multilingual needs about 13.3 GB of memory at Q4_K_M (a 11.7 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 21 quants, from 4.3 GB (IQ1_S) to 15.8 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 4.3 GB.

Llama-3-3x8B-multilingual is a 19.3B-parameter Llama model in GGUF format, quantized by mradermacher. Its trained context length is 80K tokens. The repository has 21 quant files, from 4.3 GB (IQ1_S) to 15.8 GB (Q6_K).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S4.3 GB 5.9 GB
Fits Get file →
IQ1_M4.7 GB 6.3 GB
Fits Get file →
IQ2_XXS5.3 GB 6.9 GB
Fits Get file →
IQ2_XS5.9 GB 7.5 GB
Tight Get file →
IQ2_S6.1 GB 7.7 GB
Tight Get file →
IQ2_M6.6 GB 8.2 GB
Too big Get file →
Q2_K7.3 GB 8.8 GB
Too big Get file →
IQ3_XXS7.6 GB 9.2 GB
Too big Get file →
IQ3_XS8.1 GB 9.7 GB
Too big Get file →
Q3_K_S8.5 GB 10.1 GB
Too big Get file →
IQ3_S8.5 GB 10.1 GB
Too big Get file →
IQ3_M8.7 GB 10.3 GB
Too big Get file →
Q3_K_M9.4 GB 11.0 GB
Too big Get file →
Q3_K_L10.1 GB 11.7 GB
Too big Get file →
IQ4_XS10.4 GB 12.0 GB
Too big Get file →
Q4_011.0 GB 12.6 GB
Too big Get file →
Q4_K_S11.1 GB 12.7 GB
Too big Get file →
Q4_K_M11.7 GB 13.3 GB
Too big Get file →
Q5_K_S13.4 GB 14.9 GB
Too big Get file →
Q5_K_M13.7 GB 15.3 GB
Too big Get file →
Q6_K15.8 GB 17.4 GB
Too big Get file →

What GPU do you need for Llama-3-3x8B-multilingual?

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: the best quant that fits is IQ1_S (4.3 GB file, about 5.9 GB in use).
  • 8 GB VRAM: the best quant that fits is IQ2_S (6.1 GB file, about 7.7 GB in use).
  • 12 GB VRAM: the best quant that fits is IQ4_XS (10.4 GB file, about 12.0 GB in use).
  • 16 GB VRAM: the best quant that fits is Q5_K_M (13.7 GB file, about 15.3 GB in use).
  • 24 GB VRAM: the best quant that fits is Q6_K (15.8 GB file, about 17.4 GB in use).

Which Llama-3-3x8B-multilingual quant should you choose?

Q4_K_M is the usual starting point for Llama-3-3x8B-multilingual: it balances file size and quality well. Here it is a 11.7 GB file.

If memory is tight, the smallest file is IQ1_S at 4.3 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
Parameters19.3B (approximate, mixture of experts)
Layers32
Trained context81,920 tokens
KV cache (fp16)128 KB per token

Source

Quantized bymradermacher
FamilyLlama
Licensellama3
Downloads (HF)106
Last updated2024-08-02

Frequently asked questions about Llama-3-3x8B-multilingual

How much VRAM does Llama-3-3x8B-multilingual need?

About 13.3 GB at Q4_K_M with an 8K context window. That is the 11.7 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-3-3x8B-multilingual on an 8 GB GPU?

Yes, with IQ2_S: the 6.1 GB file needs about 7.7 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Llama-3-3x8B-multilingual GGUF?

It depends on the quant: IQ1_S is 4.3 GB, IQ1_M is 4.7 GB, IQ2_XXS is 5.3 GB, IQ2_XS is 5.9 GB, IQ2_S is 6.1 GB, IQ2_M is 6.6 GB and 15 more.

What context length does Llama-3-3x8B-multilingual support?

The GGUF header lists a trained context length of 81,920 tokens (80K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.3 GB at 32K.

Where can I download Llama-3-3x8B-multilingual GGUF?

The files are on Hugging Face in the mradermacher/Llama-3-3x8B-multilingual-i1-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 Llama-3-3x8B-multilingual use?

Hugging Face lists the license as llama3. 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-3-3x8B-multilingual

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Llama models

All Llama →

Other medium models

All medium models →