Transformers
GGUF
Inference Endpoints
imatrix
bloomz-mt-i1-GGUF / README.md
mradermacher's picture
auto-patch README.md
3f87f67 verified
metadata
base_model: bigscience/bloomz-mt
datasets:
  - bigscience/xP3mt
language:
  - ak
  - ar
  - as
  - bm
  - bn
  - ca
  - code
  - en
  - es
  - eu
  - fon
  - fr
  - gu
  - hi
  - id
  - ig
  - ki
  - kn
  - lg
  - ln
  - ml
  - mr
  - ne
  - nso
  - ny
  - or
  - pa
  - pt
  - rn
  - rw
  - sn
  - st
  - sw
  - ta
  - te
  - tn
  - ts
  - tum
  - tw
  - ur
  - vi
  - wo
  - xh
  - yo
  - zh
  - zu
library_name: transformers
license: bigscience-bloom-rail-1.0
quantized_by: mradermacher

About

weighted/imatrix quants of https://huggingface.co/bigscience/bloomz-mt

static quants are available at https://huggingface.co/mradermacher/bloomz-mt-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
GGUF i1-IQ1_S 39.3 for the desperate
GGUF i1-IQ1_M 42.9 mostly desperate
GGUF i1-IQ2_XXS 48.8
PART 1 PART 2 i1-IQ2_XS 54.0
PART 1 PART 2 i1-IQ2_S 57.0
PART 1 PART 2 i1-IQ2_M 61.8
PART 1 PART 2 i1-Q2_K 68.2 IQ3_XXS probably better
PART 1 PART 2 i1-IQ3_XXS 70.9 lower quality
PART 1 PART 2 i1-IQ3_XS 76.8
PART 1 PART 2 i1-IQ3_S 78.8 beats Q3_K*
PART 1 PART 2 i1-Q3_K_S 78.8 IQ3_XS probably better
PART 1 PART 2 i1-IQ3_M 87.4
PART 1 PART 2 i1-Q3_K_M 94.5 IQ3_S probably better
PART 1 PART 2 i1-IQ4_XS 96.7
PART 1 PART 2 PART 3 i1-Q4_0 102.7 fast, low quality
PART 1 PART 2 PART 3 i1-Q4_K_S 103.1 optimal size/speed/quality
PART 1 PART 2 PART 3 i1-Q3_K_L 103.1 IQ3_M probably better
PART 1 PART 2 PART 3 i1-Q4_K_M 114.8 fast, recommended
PART 1 PART 2 PART 3 i1-Q5_K_S 124.3
PART 1 PART 2 PART 3 i1-Q5_K_M 133.7
PART 1 PART 2 PART 3 i1-Q6_K 147.7 practically like static Q6_K

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

FAQ / Model Request

See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized.

Thanks

I thank my company, nethype GmbH, for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. Additional thanks to @nicoboss for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to.