LL

llamagnific-3-87b

LlamaGGUF by mradermacherUpdated 2024-07-17
Download on Hugging Face
Parameters86.8B
Context8K
Layers99
Architecturellama
Smallest quant32.2 GB
Quick answer

llamagnific-3-87b needs about 42.9 GB of memory at IQ3_M (a 39.1 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 9 quants, from 32.2 GB (Q2_K) to 49.5 GB (Q4_K_S).

llamagnific-3-87b is a 86.8B-parameter Llama model in GGUF format, quantized by mradermacher. Its trained context length is 8K tokens. The repository has 9 quant files, from 32.2 GB (Q2_K) to 49.5 GB (Q4_K_S).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K32.2 GB 36.1 GB
Too big Get file →
IQ3_XS35.9 GB 39.7 GB
Too big Get file →
Q3_K_S37.7 GB 41.5 GB
Too big Get file →
IQ3_S37.8 GB 41.7 GB
Too big Get file →
IQ3_M39.1 GB 42.9 GB
Too big Get file →
Q3_K_M42.0 GB 45.8 GB
Too big Get file →
Q3_K_L45.6 GB 49.5 GB
Too big Get file →
IQ4_XS47.0 GB 50.8 GB
Too big Get file →
Q4_K_S49.5 GB 53.3 GB
Too big Get file →

What GPU do you need for llamagnific-3-87b?

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: no quant fits at an 8K context.
  • 48 GB VRAM: the best quant that fits is Q3_K_M (42.0 GB file, about 45.8 GB in use).

Which llamagnific-3-87b quant should you choose?

IQ3_M (39.1 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.

If memory is tight, the smallest file is Q2_K at 32.2 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
Parameters86.8B
Layers99
Trained context8,192 tokens
KV cache (fp16)396 KB per token

Source

Quantized bymradermacher
FamilyLlama
Licensellama3
Downloads (HF)206
Last updated2024-07-17

Frequently asked questions about llamagnific-3-87b

How much VRAM does llamagnific-3-87b need?

About 42.9 GB at IQ3_M with an 8K context window. That is the 39.1 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 llamagnific-3-87b on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 36.1 GB, so you would need a larger GPU or CPU offloading.

What is the file size of llamagnific-3-87b GGUF?

It depends on the quant: Q2_K is 32.2 GB, IQ3_XS is 35.9 GB, Q3_K_S is 37.7 GB, IQ3_S is 37.8 GB, IQ3_M is 39.1 GB, Q3_K_M is 42.0 GB and 3 more.

What context length does llamagnific-3-87b support?

The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 3.3 GB at 8K and 3.3 GB at 8K.

Where can I download llamagnific-3-87b GGUF?

The files are on Hugging Face in the mradermacher/llamagnific-3-87b-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 llamagnific-3-87b 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 llamagnific-3-87b

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Llama models

All Llama →

Other large models

All large models →