Hugging Face's logo

mradermacher
/
airoboros-dpo-110b-3.3-i1-GGUF

Transformers
GGUF
English
Inference Endpoints
imatrix
conversational
Model card Files Community
mradermacher commited on
Commit
1449e3c
•
1 Parent(s): 568be1b

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +90 -0
README.md CHANGED
@@ -1,6 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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/jondurbin/airoboros-dpo-110b-3.3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: jondurbin/airoboros-dpo-110b-3.3
3
+ datasets:
4
+ - jondurbin/airoboros-3.2
5
+ - bluemoon-fandom-1-1-rp-cleaned
6
+ - boolq
7
+ - LDJnr/Capybara
8
+ - jondurbin/cinematika-v0.1
9
+ - glaiveai/glaive-function-calling-v2
10
+ - grimulkan/LimaRP-augmented
11
+ - piqa
12
+ - Vezora/Tested-22k-Python-Alpaca
13
+ - mattpscott/airoboros-summarization
14
+ - unalignment/toxic-dpo-v0.2
15
+ - allenai/ultrafeedback_binarized_cleaned
16
+ - argilla/distilabel-intel-orca-dpo-pairs
17
+ - jondurbin/airoboros-3.2
18
+ - jondurbin/contextual-dpo-v0.1
19
+ - jondurbin/gutenberg-dpo-v0.1
20
+ - jondurbin/py-dpo-v0.1
21
+ - jondurbin/truthy-dpo-v0.1
22
+ - jondurbin/gutenberg-dpo-v0.1
23
+ - lmsys/lmsys-chat-1m
24
+ language:
25
+ - en
26
+ library_name: transformers
27
+ license: other
28
+ license_link: https://huggingface.co/Qwen/Qwen1.5-110B/blob/main/LICENSE
29
+ license_name: tongyi-qianwen
30
+ quantized_by: mradermacher
31
+ ---
32
+ ## About
33
+
34
  <!-- ### quantize_version: 2 -->
35
  <!-- ### output_tensor_quantised: 1 -->
36
  <!-- ### convert_type: hf -->
37
  <!-- ### vocab_type: -->
38
  <!-- ### tags: nicoboss -->
39
  weighted/imatrix quants of https://huggingface.co/jondurbin/airoboros-dpo-110b-3.3
40
+
41
+ <!-- provided-files -->
42
+ static quants are available at https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-GGUF
43
+ ## Usage
44
+
45
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
46
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
47
+ more details, including on how to concatenate multi-part files.
48
+
49
+ ## Provided Quants
50
+
51
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
52
+
53
+ | Link | Type | Size/GB | Notes |
54
+ |:-----|:-----|--------:|:------|
55
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ1_S.gguf) | i1-IQ1_S | 23.7 | for the desperate |
56
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ1_M.gguf) | i1-IQ1_M | 26.0 | mostly desperate |
57
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 29.9 | |
58
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ2_XS.gguf) | i1-IQ2_XS | 33.1 | |
59
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ2_S.gguf) | i1-IQ2_S | 34.4 | |
60
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ2_M.gguf) | i1-IQ2_M | 37.5 | |
61
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q2_K.gguf) | i1-Q2_K | 41.3 | IQ3_XXS probably better |
62
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 43.2 | lower quality |
63
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ3_XS.gguf) | i1-IQ3_XS | 46.0 | |
64
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ3_S.gguf) | i1-IQ3_S | 48.6 | beats Q3_K* |
65
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q3_K_S.gguf) | i1-Q3_K_S | 48.6 | IQ3_XS probably better |
66
+ | [GGUF](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ3_M.gguf) | i1-IQ3_M | 49.8 | |
67
+ | [PART 1](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q3_K_M.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q3_K_M.gguf.part2of2) | i1-Q3_K_M | 53.8 | IQ3_S probably better |
68
+ | [PART 1](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q3_K_L.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q3_K_L.gguf.part2of2) | i1-Q3_K_L | 58.2 | IQ3_M probably better |
69
+ | [PART 1](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ4_XS.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-IQ4_XS.gguf.part2of2) | i1-IQ4_XS | 59.7 | |
70
+ | [PART 1](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q4_0.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q4_0.gguf.part2of2) | i1-Q4_0 | 63.2 | fast, low quality |
71
+ | [PART 1](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q4_K_S.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q4_K_S.gguf.part2of2) | i1-Q4_K_S | 63.6 | optimal size/speed/quality |
72
+ | [PART 1](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q4_K_M.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q4_K_M.gguf.part2of2) | i1-Q4_K_M | 67.3 | fast, recommended |
73
+ | [PART 1](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q5_K_S.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q5_K_S.gguf.part2of2) | i1-Q5_K_S | 76.7 | |
74
+ | [PART 1](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q5_K_M.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q5_K_M.gguf.part2of2) | i1-Q5_K_M | 78.9 | |
75
+ | [PART 1](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q6_K.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/airoboros-dpo-110b-3.3-i1-GGUF/resolve/main/airoboros-dpo-110b-3.3.i1-Q6_K.gguf.part2of2) | i1-Q6_K | 91.3 | practically like static Q6_K |
76
+
77
+ Here is a handy graph by ikawrakow comparing some lower-quality quant
78
+ types (lower is better):
79
+
80
+ ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
81
+
82
+ And here are Artefact2's thoughts on the matter:
83
+ https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
84
+
85
+ ## FAQ / Model Request
86
+
87
+ See https://huggingface.co/mradermacher/model_requests for some answers to
88
+ questions you might have and/or if you want some other model quantized.
89
+
90
+ ## Thanks
91
+
92
+ I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
93
+ me use its servers and providing upgrades to my workstation to enable
94
+ 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.
95
+
96
+ <!-- end -->