Transformers
English
Inference Endpoints
mradermacher's picture
auto-patch README.md
ccd78a8 verified
|
raw
history blame
6.28 kB
metadata
datasets:
  - jondurbin/airoboros-2.2.1
language:
  - en
library_name: transformers
license: other
license_link: https://huggingface.co/tiiuae/falcon-180B/raw/main/LICENSE.txt
license_name: falcon-180b-tii-license-1.0
quantized_by: mradermacher

About

static quants of https://huggingface.co/jondurbin/airoboros-180b-2.2.1

weighted/imatrix quants are available at https://huggingface.co/mradermacher/airoboros-180b-2.2.1-i1-GGUF

Usage

If you are unsure how to use GGUF files, refer to one of TheBloke's READMEs for more details, including on how to concatenate multi-part files.

Provided Quants

(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)

Link Type Size/GB Notes
PART 1 PART 2 Q2_K 67.8
PART 1 PART 2 IQ3_XS 76.3
PART 1 PART 2 IQ3_S 78.8 fast, beats Q3_K*
PART 1 PART 2 Q3_K_S 78.8
PART 1 PART 2 IQ3_M 82.4
PART 1 PART 2 Q3_K_M 86.5 lower quality
PART 1 PART 2 Q3_K_L 93.0
PART 1 PART 2 PART 3 IQ4_XS 98.3
PART 1 PART 2 PART 3 Q4_K_S 102.5 fast, medium quality
PART 1 PART 2 PART 3 Q4_K_M 109.8 fast, medium quality
PART 1 PART 2 PART 3 Q5_K_S 124.8
PART 1 PART 2 PART 3 Q5_K_M 132.0
PART 1 PART 2 PART 3 PART 4 Q6_K 148.5 very good quality
PART 1 PART 2 PART 3 PART 4 Q8_0 191.8 fast, best quality

Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):

image.png

And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9