mradermacher commited on
Commit
a5baed7
1 Parent(s): fa788df

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +59 -0
README.md CHANGED
@@ -1 +1,60 @@
 
 
 
 
 
 
 
 
 
 
1
  weighted/imatrix quants of https://huggingface.co/NeverSleep/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ exported_from: NeverSleep/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss
3
+ language:
4
+ - en
5
+ library_name: transformers
6
+ license: cc-by-nc-4.0
7
+ quantized_by: mradermacher
8
+ ---
9
+ ## About
10
+
11
  weighted/imatrix quants of https://huggingface.co/NeverSleep/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss
12
+
13
+
14
+ <!-- provided-files -->
15
+ static quants are available at https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-GGUF
16
+ ## Usage
17
+
18
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
19
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
20
+ more details, including on how to concatenate multi-part files.
21
+
22
+ ## Provided Quants
23
+
24
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
25
+
26
+ | Link | Type | Size/GB | Notes |
27
+ |:-----|:-----|--------:|:------|
28
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 13.1 | |
29
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-IQ2_M.gguf) | i1-IQ2_M | 16.0 | |
30
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-Q2_K.gguf) | i1-Q2_K | 17.8 | IQ3_XXS probably better |
31
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 18.8 | lower quality |
32
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-IQ3_XS.gguf) | i1-IQ3_XS | 19.8 | |
33
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-IQ3_S.gguf) | i1-IQ3_S | 20.9 | beats Q3_K* |
34
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-Q3_K_S.gguf) | i1-Q3_K_S | 20.9 | IQ3_XS probably better |
35
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-IQ3_M.gguf) | i1-IQ3_M | 21.9 | |
36
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-Q3_K_M.gguf) | i1-Q3_K_M | 23.1 | IQ3_S probably better |
37
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-Q3_K_L.gguf) | i1-Q3_K_L | 24.7 | IQ3_M probably better |
38
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-IQ4_XS.gguf) | i1-IQ4_XS | 25.6 | |
39
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-Q4_K_S.gguf) | i1-Q4_K_S | 27.3 | optimal size/speed/quality |
40
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-Q4_K_M.gguf) | i1-Q4_K_M | 29.0 | fast, recommended |
41
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-Q5_K_S.gguf) | i1-Q5_K_S | 32.7 | |
42
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-Q5_K_M.gguf) | i1-Q5_K_M | 33.7 | |
43
+ | [GGUF](https://huggingface.co/mradermacher/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss-i1-GGUF/resolve/main/Noromaid-v0.4-Mixtral-Instruct-8x7b-Zloss.i1-Q6_K.gguf) | i1-Q6_K | 38.9 | practically like static Q6_K |
44
+
45
+
46
+ Here is a handy graph by ikawrakow comparing some lower-quality quant
47
+ types (lower is better):
48
+
49
+ ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
50
+
51
+ And here are Artefact2's thoughts on the matter:
52
+ https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
53
+
54
+ ## Thanks
55
+
56
+ I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
57
+ me use its servers and providing upgrades to my workstation to enable
58
+ this work in my free time.
59
+
60
+ <!-- end -->