HA

Hanames 90B L3.1

HanamesGGUF by RichardErkhovUpdated 2024-10-25
Download on Hugging Face
Parameters91.9B
Context128K
Layers105
Architecturellama
Smallest quant34.1 GB
Quick answer

Hanames 90B L3.1 needs about 59.4 GB of memory at Q4_K_M (a 55.3 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 22 quants, from 34.1 GB (Q2_K) to 97.7 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 14.1 GB.

Hanames 90B L3.1 is a 91.9B-parameter Hanames model in GGUF format, quantized by RichardErkhov. Its trained context length is 128K tokens. The repository has 22 quant files, from 34.1 GB (Q2_K) to 97.7 GB (Q8_0).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K34.1 GB 38.2 GB
Too big Get file →
IQ3_XS38.0 GB 42.0 GB
Too big Get file →
Q3_K_S39.9 GB 44.0 GB
Too big Get file →
IQ3_S40.0 GB (2 parts) 44.1 GB
Too big Get file →
IQ3_M41.4 GB (2 parts) 45.4 GB
Too big Get file →
Q3_K44.4 GB (2 parts) 48.5 GB
Too big Get file →
Q3_K_M44.4 GB (2 parts) 48.5 GB
Too big Get file →
Q3_K_L48.3 GB (2 parts) 52.4 GB
Too big Get file →
IQ4_XS49.8 GB (2 parts) 53.8 GB
Too big Get file →
Q4_052.0 GB (2 parts) 56.0 GB
Too big Get file →
Q4_K_S52.4 GB (2 parts) 56.4 GB
Too big Get file →
IQ4_NL52.5 GB (2 parts) 56.5 GB
Too big Get file →
Q4_K55.3 GB (2 parts) 59.4 GB
Too big Get file →
Q4_K_M55.3 GB (2 parts) 59.4 GB
Too big Get file →
Q4_157.7 GB (2 parts) 61.7 GB
Too big Get file →
Q5_063.4 GB (2 parts) 67.4 GB
Too big Get file →
Q5_K_S63.4 GB (2 parts) 67.4 GB
Too big Get file →
Q5_K65.1 GB (2 parts) 69.1 GB
Too big Get file →
Q5_K_M65.1 GB (2 parts) 69.1 GB
Too big Get file →
Q5_169.0 GB (2 parts) 73.1 GB
Too big Get file →
Q6_K75.4 GB (2 parts) 79.5 GB
Too big Get file →
Q8_097.7 GB (3 parts) 101.7 GB
Too big Get file →

What GPU do you need for Hanames 90B L3.1?

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 IQ3_M (41.4 GB file, about 45.4 GB in use).

Which Hanames 90B L3.1 quant should you choose?

Q4_K_M is the usual starting point for Hanames 90B L3.1: it balances file size and quality well. Here it is a 55.3 GB file.

The Q8_0 file is 97.7 GB, about 1.8 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 Q2_K at 34.1 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
Parameters91.9B
Layers105
Trained context131,072 tokens
KV cache (fp16)420 KB per token

Source

Quantized byRichardErkhov
FamilyHanames
Downloads (HF)10
Last updated2024-10-25

Frequently asked questions about Hanames 90B L3.1

How much VRAM does Hanames 90B L3.1 need?

About 59.4 GB at Q4_K_M with an 8K context window. That is the 55.3 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 Hanames 90B L3.1 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 38.2 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Hanames 90B L3.1 GGUF?

It depends on the quant: Q2_K is 34.1 GB, IQ3_XS is 38.0 GB, Q3_K_S is 39.9 GB, IQ3_S is 40.0 GB, IQ3_M is 41.4 GB, Q3_K is 44.4 GB and 16 more.

What context length does Hanames 90B L3.1 support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 3.5 GB at 8K and 14.1 GB at 32K.

Where can I download Hanames 90B L3.1 GGUF?

The files are on Hugging Face in the RichardErkhov/mergesloppa123123_-_Hanames-90B-L3.1-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 Hanames 90B L3.1 use?

Hugging Face lists the license as cc-by-nc-4.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Similar models

Browse all →

More from RichardErkhov

All by RichardErkhov →

More Hanames models

All Hanames →

Other large models

All large models →