exported_from: SparseLLM/ReluLLaMA-70B
language:
- en
library_name: transformers
license: llama2
quantized_by: mradermacher
About
weighted/imatrix quants of https://huggingface.co/SparseLLM/ReluLLaMA-70B
static quants are available at https://huggingface.co/mradermacher/ReluLLaMA-70B-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-IQ2_M | 23.3 | |
| GGUF | i1-Q2_K | 25.6 | IQ3_XXS probably better |
| GGUF | i1-IQ3_XXS | 26.7 | lower quality |
| GGUF | i1-IQ3_XS | 28.4 | |
| GGUF | i1-IQ3_S | 30.0 | beats Q3_K* |
| GGUF | i1-Q3_K_S | 30.0 | IQ3_XS probably better |
| GGUF | i1-IQ3_M | 31.0 | |
| GGUF | i1-Q3_K_M | 33.4 | IQ3_S probably better |
| GGUF | i1-Q3_K_L | 36.2 | IQ3_M probably better |
| GGUF | i1-IQ4_XS | 36.9 | |
| GGUF | i1-Q4_0 | 39.1 | fast, low quality |
| GGUF | i1-Q4_K_S | 39.3 | optimal size/speed/quality |
| GGUF | i1-Q4_K_M | 41.5 | fast, recommended |
| GGUF | i1-Q5_K_S | 47.6 | |
| GGUF | i1-Q5_K_M | 48.9 | |
| PART 1 PART 2 | i1-Q6_K | 56.7 | practically like static Q6_K |
Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):
And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
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.