mradermacher commited on
Commit
99198bb
1 Parent(s): 5e42579

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +62 -0
README.md CHANGED
@@ -1,6 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  <!-- ### quantize_version: 2 -->
2
  <!-- ### output_tensor_quantised: 1 -->
3
  <!-- ### convert_type: hf -->
4
  <!-- ### vocab_type: -->
5
  <!-- ### tags: -->
6
  static quants of https://huggingface.co/Qwen/Qwen2-57B-A14B
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: Qwen/Qwen2-57B-A14B
3
+ language:
4
+ - en
5
+ library_name: transformers
6
+ license: apache-2.0
7
+ quantized_by: mradermacher
8
+ tags:
9
+ - pretrained
10
+ - moe
11
+ ---
12
+ ## About
13
+
14
  <!-- ### quantize_version: 2 -->
15
  <!-- ### output_tensor_quantised: 1 -->
16
  <!-- ### convert_type: hf -->
17
  <!-- ### vocab_type: -->
18
  <!-- ### tags: -->
19
  static quants of https://huggingface.co/Qwen/Qwen2-57B-A14B
20
+
21
+ <!-- provided-files -->
22
+ weighted/imatrix quants are available at https://huggingface.co/mradermacher/Qwen2-57B-A14B-i1-GGUF
23
+ ## Usage
24
+
25
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
26
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
27
+ more details, including on how to concatenate multi-part files.
28
+
29
+ ## Provided Quants
30
+
31
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
32
+
33
+ | Link | Type | Size/GB | Notes |
34
+ |:-----|:-----|--------:|:------|
35
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q2_K.gguf) | Q2_K | 21.2 | |
36
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.IQ3_XS.gguf) | IQ3_XS | 23.7 | |
37
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q3_K_S.gguf) | Q3_K_S | 25.0 | |
38
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.IQ3_S.gguf) | IQ3_S | 25.0 | beats Q3_K* |
39
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.IQ3_M.gguf) | IQ3_M | 25.3 | |
40
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q3_K_M.gguf) | Q3_K_M | 27.6 | lower quality |
41
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q3_K_L.gguf) | Q3_K_L | 29.9 | |
42
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q4_K_S.gguf) | Q4_K_S | 32.8 | fast, recommended |
43
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q4_K_M.gguf) | Q4_K_M | 35.0 | fast, recommended |
44
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q5_K_S.gguf) | Q5_K_S | 39.7 | |
45
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q5_K_M.gguf) | Q5_K_M | 40.9 | |
46
+ | [GGUF](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q6_K.gguf) | Q6_K | 47.2 | very good quality |
47
+ | [PART 1](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q8_0.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/Qwen2-57B-A14B-GGUF/resolve/main/Qwen2-57B-A14B.Q8_0.gguf.part2of2) | Q8_0 | 61.1 | fast, best quality |
48
+
49
+ Here is a handy graph by ikawrakow comparing some lower-quality quant
50
+ types (lower is better):
51
+
52
+ ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
53
+
54
+ And here are Artefact2's thoughts on the matter:
55
+ https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
56
+
57
+ ## FAQ / Model Request
58
+
59
+ See https://huggingface.co/mradermacher/model_requests for some answers to
60
+ questions you might have and/or if you want some other model quantized.
61
+
62
+ ## Thanks
63
+
64
+ I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
65
+ me use its servers and providing upgrades to my workstation to enable
66
+ this work in my free time.
67
+
68
+ <!-- end -->