mradermacher commited on
Commit
e596539
1 Parent(s): 67eaf4a

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +64 -0
README.md CHANGED
@@ -1,6 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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/abacusai/Smaug-Qwen2-72B-Instruct
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: abacusai/Smaug-Qwen2-72B-Instruct
3
+ language:
4
+ - en
5
+ library_name: transformers
6
+ license: other
7
+ license_link: https://huggingface.co/Qwen/Qwen2-72B-Instruct/blob/main/LICENSE
8
+ license_name: tongyi-qianwen
9
+ quantized_by: mradermacher
10
+ tags:
11
+ - chat
12
+ ---
13
+ ## About
14
+
15
  <!-- ### quantize_version: 2 -->
16
  <!-- ### output_tensor_quantised: 1 -->
17
  <!-- ### convert_type: hf -->
18
  <!-- ### vocab_type: -->
19
  <!-- ### tags: -->
20
  static quants of https://huggingface.co/abacusai/Smaug-Qwen2-72B-Instruct
21
+
22
+ <!-- provided-files -->
23
+ weighted/imatrix quants are available at https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-i1-GGUF
24
+ ## Usage
25
+
26
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
27
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
28
+ more details, including on how to concatenate multi-part files.
29
+
30
+ ## Provided Quants
31
+
32
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
33
+
34
+ | Link | Type | Size/GB | Notes |
35
+ |:-----|:-----|--------:|:------|
36
+ | [GGUF](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q2_K.gguf) | Q2_K | 29.9 | |
37
+ | [GGUF](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.IQ3_XS.gguf) | IQ3_XS | 32.9 | |
38
+ | [GGUF](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.IQ3_S.gguf) | IQ3_S | 34.6 | beats Q3_K* |
39
+ | [GGUF](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q3_K_S.gguf) | Q3_K_S | 34.6 | |
40
+ | [GGUF](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.IQ3_M.gguf) | IQ3_M | 35.6 | |
41
+ | [GGUF](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q3_K_M.gguf) | Q3_K_M | 37.8 | lower quality |
42
+ | [GGUF](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q3_K_L.gguf) | Q3_K_L | 39.6 | |
43
+ | [GGUF](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.IQ4_XS.gguf) | IQ4_XS | 40.3 | |
44
+ | [GGUF](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q4_K_S.gguf) | Q4_K_S | 44.0 | fast, recommended |
45
+ | [GGUF](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q4_K_M.gguf) | Q4_K_M | 47.5 | fast, recommended |
46
+ | [PART 1](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q5_K_S.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q5_K_S.gguf.part2of2) | Q5_K_S | 51.5 | |
47
+ | [PART 1](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q5_K_M.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q5_K_M.gguf.part2of2) | Q5_K_M | 54.5 | |
48
+ | [PART 1](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q6_K.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q6_K.gguf.part2of2) | Q6_K | 64.4 | very good quality |
49
+ | [PART 1](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q8_0.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/Smaug-Qwen2-72B-Instruct-GGUF/resolve/main/Smaug-Qwen2-72B-Instruct.Q8_0.gguf.part2of2) | Q8_0 | 77.4 | fast, best quality |
50
+
51
+ Here is a handy graph by ikawrakow comparing some lower-quality quant
52
+ types (lower is better):
53
+
54
+ ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
55
+
56
+ And here are Artefact2's thoughts on the matter:
57
+ https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
58
+
59
+ ## FAQ / Model Request
60
+
61
+ See https://huggingface.co/mradermacher/model_requests for some answers to
62
+ questions you might have and/or if you want some other model quantized.
63
+
64
+ ## Thanks
65
+
66
+ I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
67
+ me use its servers and providing upgrades to my workstation to enable
68
+ this work in my free time.
69
+
70
+ <!-- end -->