mradermacher commited on
Commit
6df91a8
1 Parent(s): f337784

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +59 -0
README.md CHANGED
@@ -1 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  static quants of https://huggingface.co/sophosympatheia/Midnight-Miqu-70B-v1.5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: []
3
+ language:
4
+ - en
5
+ library_name: transformers
6
+ license: other
7
+ quantized_by: mradermacher
8
+ tags:
9
+ - mergekit
10
+ - merge
11
+ ---
12
+ ## About
13
+
14
  static quants of https://huggingface.co/sophosympatheia/Midnight-Miqu-70B-v1.5
15
+
16
+
17
+ <!-- provided-files -->
18
+ ## Usage
19
+
20
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
21
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
22
+ more details, including on how to concatenate multi-part files.
23
+
24
+ ## Provided Quants
25
+
26
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
27
+
28
+ | Link | Type | Size/GB | Notes |
29
+ |:-----|:-----|--------:|:------|
30
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q2_K.gguf) | Q2_K | 25.9 | |
31
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.IQ3_XS.gguf) | IQ3_XS | 28.6 | |
32
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.IQ3_S.gguf) | IQ3_S | 30.3 | fast, beats Q3_K* |
33
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q3_K_S.gguf) | Q3_K_S | 30.3 | |
34
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.IQ3_M.gguf) | IQ3_M | 31.4 | |
35
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q3_K_M.gguf) | Q3_K_M | 33.7 | lower quality |
36
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q3_K_L.gguf) | Q3_K_L | 36.6 | |
37
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.IQ4_XS.gguf) | IQ4_XS | 37.6 | |
38
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q4_K_S.gguf) | Q4_K_S | 39.7 | fast, medium quality |
39
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q4_K_M.gguf) | Q4_K_M | 41.8 | fast, medium quality |
40
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q5_K_S.gguf) | Q5_K_S | 47.9 | |
41
+ | [GGUF](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q5_K_M.gguf) | Q5_K_M | 49.2 | |
42
+ | [PART 1](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q6_K.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q6_K.gguf.part2of2) | Q6_K | 57.0 | very good quality |
43
+ | [PART 1](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q8_0.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/Midnight-Miqu-70B-v1.5-GGUF/resolve/main/Midnight-Miqu-70B-v1.5.Q8_0.gguf.part2of2) | Q8_0 | 73.6 | fast, best quality |
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 -->