LL

Grafted Llama2 2x70B

LlamaGGUF by mradermacherUpdated 2024-11-20
Download on Hugging Face
Parameters125B
Context4K
Layers80
Architecturellama
Smallest quant25.9 GB
Quick answer

Grafted Llama2 2x70B needs about 39.9 GB of memory at IQ2_XS (a 36.7 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 8 quants, from 25.9 GB (IQ1_S) to 48.2 GB (IQ3_XXS).

Grafted Llama2 2x70B is a 125B-parameter Llama model in GGUF format, quantized by mradermacher. Its trained context length is 4K tokens. The repository has 8 quant files, from 25.9 GB (IQ1_S) to 48.2 GB (IQ3_XXS).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S25.9 GB 27.7 GB
Too big Get file →
IQ1_M28.5 GB 30.4 GB
Too big Get file →
IQ2_XXS33.0 GB 34.8 GB
Too big Get file →
IQ2_XS36.7 GB 38.6 GB
Too big Get file →
IQ2_S38.0 GB 39.8 GB
Too big Get file →
IQ2_M41.5 GB 43.4 GB
Too big Get file →
Q2_K45.9 GB 47.7 GB
Too big Get file →
IQ3_XXS48.2 GB 50.0 GB
Too big Get file →

What GPU do you need for Grafted Llama2 2x70B?

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: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: no quant fits at an 8K context.
  • 32 GB VRAM: the best quant that fits is IQ1_M (28.5 GB file, about 31.7 GB in use).
  • 48 GB VRAM: the best quant that fits is IQ2_M (41.5 GB file, about 44.7 GB in use).

Which Grafted Llama2 2x70B quant should you choose?

IQ2_XS (36.7 GB) is the middle option in this repository. 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.

If memory is tight, the smallest file is IQ1_S at 25.9 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
Parameters125B (approximate, mixture of experts)
Layers80
Trained context4,096 tokens
KV cache (fp16)320 KB per token

Source

Quantized bymradermacher
FamilyLlama
Licensellama2
Downloads (HF)66
Last updated2024-11-20

Frequently asked questions about Grafted Llama2 2x70B

How much VRAM does Grafted Llama2 2x70B need?

About 39.9 GB at IQ2_XS with an 8K context window. That is the 36.7 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 Grafted Llama2 2x70B on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 29.0 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Grafted Llama2 2x70B GGUF?

It depends on the quant: IQ1_S is 25.9 GB, IQ1_M is 28.5 GB, IQ2_XXS is 33.0 GB, IQ2_XS is 36.7 GB, IQ2_S is 38.0 GB, IQ2_M is 41.5 GB and 2 more.

What context length does Grafted Llama2 2x70B support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 2.7 GB at 8K and 1.3 GB at 4K.

Where can I download Grafted Llama2 2x70B GGUF?

The files are on Hugging Face in the mradermacher/Grafted-Llama2-2x70B-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 Grafted Llama2 2x70B use?

Hugging Face lists the license as llama2. 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 Grafted-Llama2-2x70B

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Llama models

All Llama →

Other large models

All large models →