mradermacher commited on
Commit
cf8617c
·
verified ·
1 Parent(s): 2efba1a

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +44 -1
README.md CHANGED
@@ -1 +1,44 @@
1
- weighted/imatrix quants of https://huggingface.co/Dracones/perky-70b-v0.1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ library_name: transformers
5
+ license: llama2
6
+ quantized_by: mradermacher
7
+ tags:
8
+ - not-for-all-audiences
9
+ ---
10
+ ## About
11
+
12
+ weighted/imatrix quants of https://huggingface.co/Dracones/perky-70b-v0.1
13
+
14
+ <!-- provided-files -->
15
+ ## Usage
16
+
17
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
18
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
19
+ more details, including on how to concatenate multi-part files.
20
+
21
+ ## Provided Quants
22
+
23
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
24
+
25
+ | Link | Type | Size/GB | Notes |
26
+ |:-----|:-----|--------:|:------|
27
+ | [GGUF](https://huggingface.co/mradermacher/perky-70b-v0.1-i1-GGUF/resolve/main/perky-70b-v0.1.i1-Q2_K.gguf) | i1-Q2_K | 25.9 | IQ3_XXS probably better |
28
+ | [GGUF](https://huggingface.co/mradermacher/perky-70b-v0.1-i1-GGUF/resolve/main/perky-70b-v0.1.i1-Q3_K_S.gguf) | i1-Q3_K_S | 30.3 | IQ3_XS probably better |
29
+ | [GGUF](https://huggingface.co/mradermacher/perky-70b-v0.1-i1-GGUF/resolve/main/perky-70b-v0.1.i1-Q3_K_M.gguf) | i1-Q3_K_M | 33.7 | IQ3_S probably better |
30
+ | [GGUF](https://huggingface.co/mradermacher/perky-70b-v0.1-i1-GGUF/resolve/main/perky-70b-v0.1.i1-Q3_K_L.gguf) | i1-Q3_K_L | 36.6 | IQ3_M probably better |
31
+ | [GGUF](https://huggingface.co/mradermacher/perky-70b-v0.1-i1-GGUF/resolve/main/perky-70b-v0.1.i1-Q4_K_S.gguf) | i1-Q4_K_S | 39.7 | almost as good as Q4_K_M |
32
+ | [GGUF](https://huggingface.co/mradermacher/perky-70b-v0.1-i1-GGUF/resolve/main/perky-70b-v0.1.i1-Q4_K_M.gguf) | i1-Q4_K_M | 41.8 | fast, medium quality |
33
+ | [GGUF](https://huggingface.co/mradermacher/perky-70b-v0.1-i1-GGUF/resolve/main/perky-70b-v0.1.i1-Q5_K_S.gguf) | i1-Q5_K_S | 47.9 | |
34
+
35
+
36
+ Here is a handy graph by ikawrakow comparing some lower-quality quant
37
+ types (lower is better):
38
+
39
+ ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
40
+
41
+ And here are Artefact2's thoughts on the matter:
42
+ https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
43
+
44
+ <!-- end -->