mradermacher commited on
Commit
7373a0f
1 Parent(s): 516d95a

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +63 -0
README.md CHANGED
@@ -1,3 +1,66 @@
 
 
 
 
 
 
 
 
 
1
  <!-- ### convert_type: -->
2
  <!-- ### vocab_type: -->
3
  weighted/imatrix quants of https://huggingface.co/arlineka/CatNyanster-34b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ exported_from: arlineka/CatNyanster-34b
3
+ language:
4
+ - en
5
+ library_name: transformers
6
+ quantized_by: mradermacher
7
+ ---
8
+ ## About
9
+
10
  <!-- ### convert_type: -->
11
  <!-- ### vocab_type: -->
12
  weighted/imatrix quants of https://huggingface.co/arlineka/CatNyanster-34b
13
+
14
+
15
+ <!-- provided-files -->
16
+ static quants are available at https://huggingface.co/mradermacher/CatNyanster-34b-GGUF
17
+ ## Usage
18
+
19
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
20
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
21
+ more details, including on how to concatenate multi-part files.
22
+
23
+ ## Provided Quants
24
+
25
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
26
+
27
+ | Link | Type | Size/GB | Notes |
28
+ |:-----|:-----|--------:|:------|
29
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ1_S.gguf) | i1-IQ1_S | 8.2 | for the desperate |
30
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ1_M.gguf) | i1-IQ1_M | 8.9 | for the desperate |
31
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 10.0 | |
32
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ2_XS.gguf) | i1-IQ2_XS | 11.0 | |
33
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ2_S.gguf) | i1-IQ2_S | 11.6 | |
34
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ2_M.gguf) | i1-IQ2_M | 12.5 | |
35
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-Q2_K.gguf) | i1-Q2_K | 13.5 | IQ3_XXS probably better |
36
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 14.0 | lower quality |
37
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ3_XS.gguf) | i1-IQ3_XS | 14.9 | |
38
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-Q3_K_S.gguf) | i1-Q3_K_S | 15.6 | IQ3_XS probably better |
39
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ3_S.gguf) | i1-IQ3_S | 15.7 | beats Q3_K* |
40
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ3_M.gguf) | i1-IQ3_M | 16.2 | |
41
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-Q3_K_M.gguf) | i1-Q3_K_M | 17.3 | IQ3_S probably better |
42
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-Q3_K_L.gguf) | i1-Q3_K_L | 18.8 | IQ3_M probably better |
43
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-IQ4_XS.gguf) | i1-IQ4_XS | 19.1 | |
44
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-Q4_0.gguf) | i1-Q4_0 | 20.2 | fast, low quality |
45
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-Q4_K_S.gguf) | i1-Q4_K_S | 20.2 | optimal size/speed/quality |
46
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-Q4_K_M.gguf) | i1-Q4_K_M | 21.3 | fast, recommended |
47
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-Q5_K_S.gguf) | i1-Q5_K_S | 24.3 | |
48
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-Q5_K_M.gguf) | i1-Q5_K_M | 25.0 | |
49
+ | [GGUF](https://huggingface.co/mradermacher/CatNyanster-34b-i1-GGUF/resolve/main/CatNyanster-34b.i1-Q6_K.gguf) | i1-Q6_K | 28.9 | practically like static Q6_K |
50
+
51
+
52
+ Here is a handy graph by ikawrakow comparing some lower-quality quant
53
+ types (lower is better):
54
+
55
+ ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
56
+
57
+ And here are Artefact2's thoughts on the matter:
58
+ https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
59
+
60
+ ## Thanks
61
+
62
+ I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
63
+ me use its servers and providing upgrades to my workstation to enable
64
+ this work in my free time.
65
+
66
+ <!-- end -->