mradermacher commited on
Commit
220ca87
1 Parent(s): 82932ba

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +72 -0
README.md CHANGED
@@ -1,6 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
  <!-- ### quantize_version: 2 -->
2
  <!-- ### output_tensor_quantised: 1 -->
3
  <!-- ### convert_type: hf -->
4
  <!-- ### vocab_type: -->
5
  <!-- ### tags: nicoboss -->
6
  weighted/imatrix quants of https://huggingface.co/nerdigent/L3.1-BlueSkyy-8B
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: nerdigent/L3.1-BlueSkyy-8B
3
+ language:
4
+ - en
5
+ library_name: transformers
6
+ quantized_by: mradermacher
7
+ tags:
8
+ - mergekit
9
+ - merge
10
+ ---
11
+ ## About
12
+
13
  <!-- ### quantize_version: 2 -->
14
  <!-- ### output_tensor_quantised: 1 -->
15
  <!-- ### convert_type: hf -->
16
  <!-- ### vocab_type: -->
17
  <!-- ### tags: nicoboss -->
18
  weighted/imatrix quants of https://huggingface.co/nerdigent/L3.1-BlueSkyy-8B
19
+
20
+ <!-- provided-files -->
21
+ static quants are available at https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-GGUF
22
+ ## Usage
23
+
24
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
25
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
26
+ more details, including on how to concatenate multi-part files.
27
+
28
+ ## Provided Quants
29
+
30
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
31
+
32
+ | Link | Type | Size/GB | Notes |
33
+ |:-----|:-----|--------:|:------|
34
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ1_S.gguf) | i1-IQ1_S | 2.1 | for the desperate |
35
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ1_M.gguf) | i1-IQ1_M | 2.3 | mostly desperate |
36
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 2.5 | |
37
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ2_XS.gguf) | i1-IQ2_XS | 2.7 | |
38
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ2_S.gguf) | i1-IQ2_S | 2.9 | |
39
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ2_M.gguf) | i1-IQ2_M | 3.0 | |
40
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q2_K.gguf) | i1-Q2_K | 3.3 | IQ3_XXS probably better |
41
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 3.4 | lower quality |
42
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ3_XS.gguf) | i1-IQ3_XS | 3.6 | |
43
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q3_K_S.gguf) | i1-Q3_K_S | 3.8 | IQ3_XS probably better |
44
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ3_S.gguf) | i1-IQ3_S | 3.8 | beats Q3_K* |
45
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ3_M.gguf) | i1-IQ3_M | 3.9 | |
46
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q3_K_M.gguf) | i1-Q3_K_M | 4.1 | IQ3_S probably better |
47
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q3_K_L.gguf) | i1-Q3_K_L | 4.4 | IQ3_M probably better |
48
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-IQ4_XS.gguf) | i1-IQ4_XS | 4.5 | |
49
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q4_0_4_4.gguf) | i1-Q4_0_4_4 | 4.8 | fast on arm, low quality |
50
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q4_0_4_8.gguf) | i1-Q4_0_4_8 | 4.8 | fast on arm+i8mm, low quality |
51
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q4_0_8_8.gguf) | i1-Q4_0_8_8 | 4.8 | fast on arm+sve, low quality |
52
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q4_0.gguf) | i1-Q4_0 | 4.8 | fast, low quality |
53
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q4_K_S.gguf) | i1-Q4_K_S | 4.8 | optimal size/speed/quality |
54
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q4_K_M.gguf) | i1-Q4_K_M | 5.0 | fast, recommended |
55
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q5_K_S.gguf) | i1-Q5_K_S | 5.7 | |
56
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q5_K_M.gguf) | i1-Q5_K_M | 5.8 | |
57
+ | [GGUF](https://huggingface.co/mradermacher/L3.1-BlueSkyy-8B-i1-GGUF/resolve/main/L3.1-BlueSkyy-8B.i1-Q6_K.gguf) | i1-Q6_K | 6.7 | practically like static Q6_K |
58
+
59
+ Here is a handy graph by ikawrakow comparing some lower-quality quant
60
+ types (lower is better):
61
+
62
+ ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
63
+
64
+ And here are Artefact2's thoughts on the matter:
65
+ https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
66
+
67
+ ## FAQ / Model Request
68
+
69
+ See https://huggingface.co/mradermacher/model_requests for some answers to
70
+ questions you might have and/or if you want some other model quantized.
71
+
72
+ ## Thanks
73
+
74
+ I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
75
+ me use its servers and providing upgrades to my workstation to enable
76
+ this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to.
77
+
78
+ <!-- end -->