Hugging Face's logo

mradermacher
/
Llama-3-13B-Instruct-v0.1-i1-GGUF

Transformers
GGUF
PyTorch
English
mergekit
Merge
facebook
meta
llama
llama-3
Inference Endpoints
imatrix
conversational
Model card Files Community
mradermacher commited on
Commit
153f57a
•
1 Parent(s): a53615f

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +83 -0
README.md CHANGED
@@ -1,6 +1,89 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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/MaziyarPanahi/Llama-3-13B-Instruct-v0.1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: MaziyarPanahi/Llama-3-13B-Instruct-v0.1
3
+ language:
4
+ - en
5
+ library_name: transformers
6
+ license: other
7
+ license_link: LICENSE
8
+ license_name: llama3
9
+ model_creator: MaziyarPanahi
10
+ model_name: Llama-3-13B-Instruct-v0.1
11
+ quantized_by: mradermacher
12
+ tags:
13
+ - mergekit
14
+ - merge
15
+ - facebook
16
+ - meta
17
+ - pytorch
18
+ - llama
19
+ - llama-3
20
+ ---
21
+ ## About
22
+
23
  <!-- ### quantize_version: 2 -->
24
  <!-- ### output_tensor_quantised: 1 -->
25
  <!-- ### convert_type: hf -->
26
  <!-- ### vocab_type: -->
27
  <!-- ### tags: nicoboss -->
28
  weighted/imatrix quants of https://huggingface.co/MaziyarPanahi/Llama-3-13B-Instruct-v0.1
29
+
30
+ <!-- provided-files -->
31
+ static quants are available at https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-GGUF
32
+ ## Usage
33
+
34
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
35
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
36
+ more details, including on how to concatenate multi-part files.
37
+
38
+ ## Provided Quants
39
+
40
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
41
+
42
+ | Link | Type | Size/GB | Notes |
43
+ |:-----|:-----|--------:|:------|
44
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ1_S.gguf) | i1-IQ1_S | 3.2 | for the desperate |
45
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ1_M.gguf) | i1-IQ1_M | 3.5 | mostly desperate |
46
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 3.9 | |
47
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ2_XS.gguf) | i1-IQ2_XS | 4.3 | |
48
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ2_S.gguf) | i1-IQ2_S | 4.5 | |
49
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ2_M.gguf) | i1-IQ2_M | 4.8 | |
50
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q2_K_S.gguf) | i1-Q2_K_S | 4.9 | very low quality |
51
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q2_K.gguf) | i1-Q2_K | 5.2 | IQ3_XXS probably better |
52
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 5.4 | lower quality |
53
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ3_XS.gguf) | i1-IQ3_XS | 5.8 | |
54
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q3_K_S.gguf) | i1-Q3_K_S | 6.0 | IQ3_XS probably better |
55
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ3_S.gguf) | i1-IQ3_S | 6.0 | beats Q3_K* |
56
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ3_M.gguf) | i1-IQ3_M | 6.2 | |
57
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q3_K_M.gguf) | i1-Q3_K_M | 6.6 | IQ3_S probably better |
58
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q3_K_L.gguf) | i1-Q3_K_L | 7.2 | IQ3_M probably better |
59
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-IQ4_XS.gguf) | i1-IQ4_XS | 7.3 | |
60
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q4_0_4_4.gguf) | i1-Q4_0_4_4 | 7.7 | fast on arm, low quality |
61
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q4_0_4_8.gguf) | i1-Q4_0_4_8 | 7.7 | fast on arm+i8mm, low quality |
62
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q4_0_8_8.gguf) | i1-Q4_0_8_8 | 7.7 | fast on arm+sve, low quality |
63
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q4_0.gguf) | i1-Q4_0 | 7.7 | fast, low quality |
64
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q4_K_S.gguf) | i1-Q4_K_S | 7.8 | optimal size/speed/quality |
65
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q4_K_M.gguf) | i1-Q4_K_M | 8.2 | fast, recommended |
66
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q5_K_S.gguf) | i1-Q5_K_S | 9.3 | |
67
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q5_K_M.gguf) | i1-Q5_K_M | 9.5 | |
68
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3-13B-Instruct-v0.1-i1-GGUF/resolve/main/Llama-3-13B-Instruct-v0.1.i1-Q6_K.gguf) | i1-Q6_K | 11.0 | practically like static Q6_K |
69
+
70
+ Here is a handy graph by ikawrakow comparing some lower-quality quant
71
+ types (lower is better):
72
+
73
+ ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
74
+
75
+ And here are Artefact2's thoughts on the matter:
76
+ https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
77
+
78
+ ## FAQ / Model Request
79
+
80
+ See https://huggingface.co/mradermacher/model_requests for some answers to
81
+ questions you might have and/or if you want some other model quantized.
82
+
83
+ ## Thanks
84
+
85
+ I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
86
+ me use its servers and providing upgrades to my workstation to enable
87
+ 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.
88
+
89
+ <!-- end -->