Hugging Face's logo

mradermacher
/
X-ALMA-13B-Pretrain-i1-GGUF

Transformers
GGUF
Inference Endpoints
imatrix
conversational
Model card Files Community
mradermacher commited on
Commit
dcb1e2e
·
verified ·
1 Parent(s): 62df52c

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +123 -0
README.md CHANGED
@@ -1,6 +1,129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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/haoranxu/X-ALMA-13B-Pretrain
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: haoranxu/X-ALMA-13B-Pretrain
3
+ datasets:
4
+ - oscar-corpus/OSCAR-2301
5
+ - allenai/nllb
6
+ - Helsinki-NLP/opus-100
7
+ language:
8
+ - en
9
+ - da
10
+ - nl
11
+ - de
12
+ - is
13
+ - no
14
+ - sc
15
+ - af
16
+ - ca
17
+ - ro
18
+ - gl
19
+ - it
20
+ - pt
21
+ - es
22
+ - bg
23
+ - mk
24
+ - sr
25
+ - uk
26
+ - ru
27
+ - id
28
+ - ms
29
+ - th
30
+ - vi
31
+ - mg
32
+ - fr
33
+ - hu
34
+ - el
35
+ - cs
36
+ - pl
37
+ - lt
38
+ - lv
39
+ - ka
40
+ - zh
41
+ - ja
42
+ - ko
43
+ - fi
44
+ - et
45
+ - gu
46
+ - hi
47
+ - mr
48
+ - ne
49
+ - ur
50
+ - az
51
+ - kk
52
+ - ky
53
+ - tr
54
+ - uz
55
+ - ar
56
+ - he
57
+ - fa
58
+ library_name: transformers
59
+ license: mit
60
+ quantized_by: mradermacher
61
+ ---
62
+ ## About
63
+
64
  <!-- ### quantize_version: 2 -->
65
  <!-- ### output_tensor_quantised: 1 -->
66
  <!-- ### convert_type: hf -->
67
  <!-- ### vocab_type: -->
68
  <!-- ### tags: nicoboss -->
69
  weighted/imatrix quants of https://huggingface.co/haoranxu/X-ALMA-13B-Pretrain
70
+
71
+ <!-- provided-files -->
72
+ static quants are available at https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-GGUF
73
+ ## Usage
74
+
75
+ If you are unsure how to use GGUF files, refer to one of [TheBloke's
76
+ READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
77
+ more details, including on how to concatenate multi-part files.
78
+
79
+ ## Provided Quants
80
+
81
+ (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
82
+
83
+ | Link | Type | Size/GB | Notes |
84
+ |:-----|:-----|--------:|:------|
85
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ1_S.gguf) | i1-IQ1_S | 3.0 | for the desperate |
86
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ1_M.gguf) | i1-IQ1_M | 3.2 | mostly desperate |
87
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 3.6 | |
88
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ2_XS.gguf) | i1-IQ2_XS | 4.0 | |
89
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ2_S.gguf) | i1-IQ2_S | 4.3 | |
90
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q2_K_S.gguf) | i1-Q2_K_S | 4.5 | very low quality |
91
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ2_M.gguf) | i1-IQ2_M | 4.6 | |
92
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q2_K.gguf) | i1-Q2_K | 5.0 | IQ3_XXS probably better |
93
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 5.1 | lower quality |
94
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ3_XS.gguf) | i1-IQ3_XS | 5.5 | |
95
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ3_S.gguf) | i1-IQ3_S | 5.8 | beats Q3_K* |
96
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q3_K_S.gguf) | i1-Q3_K_S | 5.8 | IQ3_XS probably better |
97
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ3_M.gguf) | i1-IQ3_M | 6.1 | |
98
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q3_K_M.gguf) | i1-Q3_K_M | 6.4 | IQ3_S probably better |
99
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q3_K_L.gguf) | i1-Q3_K_L | 7.0 | IQ3_M probably better |
100
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ4_XS.gguf) | i1-IQ4_XS | 7.1 | |
101
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-IQ4_NL.gguf) | i1-IQ4_NL | 7.5 | prefer IQ4_XS |
102
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q4_0.gguf) | i1-Q4_0 | 7.5 | fast, low quality |
103
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q4_K_S.gguf) | i1-Q4_K_S | 7.5 | optimal size/speed/quality |
104
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q4_K_M.gguf) | i1-Q4_K_M | 8.0 | fast, recommended |
105
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q4_1.gguf) | i1-Q4_1 | 8.3 | |
106
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q5_K_S.gguf) | i1-Q5_K_S | 9.1 | |
107
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q5_K_M.gguf) | i1-Q5_K_M | 9.3 | |
108
+ | [GGUF](https://huggingface.co/mradermacher/X-ALMA-13B-Pretrain-i1-GGUF/resolve/main/X-ALMA-13B-Pretrain.i1-Q6_K.gguf) | i1-Q6_K | 10.8 | practically like static Q6_K |
109
+
110
+ Here is a handy graph by ikawrakow comparing some lower-quality quant
111
+ types (lower is better):
112
+
113
+ ![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
114
+
115
+ And here are Artefact2's thoughts on the matter:
116
+ https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
117
+
118
+ ## FAQ / Model Request
119
+
120
+ See https://huggingface.co/mradermacher/model_requests for some answers to
121
+ questions you might have and/or if you want some other model quantized.
122
+
123
+ ## Thanks
124
+
125
+ I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
126
+ me use its servers and providing upgrades to my workstation to enable
127
+ 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.
128
+
129
+ <!-- end -->