mradermacher commited on
Commit
b59ead0
1 Parent(s): 06dee9b

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +60 -0
README.md CHANGED
@@ -1,5 +1,65 @@
 
 
 
 
 
 
 
 
 
 
1
  <!-- ### quantize_version: 2 -->
2
  <!-- ### output_tensor_quantised: 1 -->
3
  <!-- ### convert_type: hf -->
4
  <!-- ### vocab_type: -->
5
  static quants of https://huggingface.co/ontocord/phi-3-22b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: ontocord/phi-3-22b
3
+ language:
4
+ - en
5
+ library_name: transformers
6
+ license: mit
7
+ quantized_by: mradermacher
8
+ ---
9
+ ## About
10
+
11
  <!-- ### quantize_version: 2 -->
12
  <!-- ### output_tensor_quantised: 1 -->
13
  <!-- ### convert_type: hf -->
14
  <!-- ### vocab_type: -->
15
  static quants of https://huggingface.co/ontocord/phi-3-22b
16
+
17
+ <!-- provided-files -->
18
+ weighted/imatrix quants seem not to be available (by me) at this time. If they do not show up a week or so after the static ones, I have probably not planned for them. Feel free to request them by opening a Community Discussion.
19
+ ## Usage
20
+
21
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
22
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
23
+ more details, including on how to concatenate multi-part files.
24
+
25
+ ## Provided Quants
26
+
27
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
28
+
29
+ | Link | Type | Size/GB | Notes |
30
+ |:-----|:-----|--------:|:------|
31
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.Q2_K.gguf) | Q2_K | 8.2 | |
32
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.IQ3_XS.gguf) | IQ3_XS | 9.3 | |
33
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.IQ3_S.gguf) | IQ3_S | 9.7 | beats Q3_K* |
34
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.Q3_K_S.gguf) | Q3_K_S | 9.7 | |
35
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.IQ3_M.gguf) | IQ3_M | 10.3 | |
36
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.Q3_K_M.gguf) | Q3_K_M | 11.1 | lower quality |
37
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.Q3_K_L.gguf) | Q3_K_L | 12.0 | |
38
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.IQ4_XS.gguf) | IQ4_XS | 12.0 | |
39
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.Q4_K_S.gguf) | Q4_K_S | 12.7 | fast, recommended |
40
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.Q4_K_M.gguf) | Q4_K_M | 13.7 | fast, recommended |
41
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.Q5_K_S.gguf) | Q5_K_S | 15.3 | |
42
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.Q5_K_M.gguf) | Q5_K_M | 16.1 | |
43
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.Q6_K.gguf) | Q6_K | 18.3 | very good quality |
44
+ | [GGUF](https://huggingface.co/mradermacher/phi-3-22b-GGUF/resolve/main/phi-3-22b.Q8_0.gguf) | Q8_0 | 23.6 | fast, best quality |
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
+ ## FAQ / Model Request
55
+
56
+ See https://huggingface.co/mradermacher/model_requests for some answers to
57
+ questions you might have and/or if you want some other model quantized.
58
+
59
+ ## Thanks
60
+
61
+ I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
62
+ me use its servers and providing upgrades to my workstation to enable
63
+ this work in my free time.
64
+
65
+ <!-- end -->