mradermacher
commited on
Commit
•
5d18238
1
Parent(s):
1d8a2b3
auto-patch README.md
Browse files
README.md
CHANGED
@@ -1,6 +1,99 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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/langtech-dev/Salamandra-7b-RAG-v2
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: langtech-dev/Salamandra-7b-RAG-v2
|
3 |
+
language:
|
4 |
+
- bg
|
5 |
+
- ca
|
6 |
+
- code
|
7 |
+
- cs
|
8 |
+
- cy
|
9 |
+
- da
|
10 |
+
- de
|
11 |
+
- el
|
12 |
+
- en
|
13 |
+
- es
|
14 |
+
- et
|
15 |
+
- eu
|
16 |
+
- fi
|
17 |
+
- fr
|
18 |
+
- ga
|
19 |
+
- gl
|
20 |
+
- hr
|
21 |
+
- hu
|
22 |
+
- it
|
23 |
+
- lt
|
24 |
+
- lv
|
25 |
+
- mt
|
26 |
+
- nl
|
27 |
+
- nn
|
28 |
+
- \no
|
29 |
+
- oc
|
30 |
+
- pl
|
31 |
+
- pt
|
32 |
+
- ro
|
33 |
+
- ru
|
34 |
+
- sh
|
35 |
+
- sk
|
36 |
+
- sl
|
37 |
+
- sr
|
38 |
+
- sv
|
39 |
+
- uk
|
40 |
+
library_name: transformers
|
41 |
+
license: apache-2.0
|
42 |
+
quantized_by: mradermacher
|
43 |
+
---
|
44 |
+
## About
|
45 |
+
|
46 |
<!-- ### quantize_version: 2 -->
|
47 |
<!-- ### output_tensor_quantised: 1 -->
|
48 |
<!-- ### convert_type: hf -->
|
49 |
<!-- ### vocab_type: -->
|
50 |
<!-- ### tags: -->
|
51 |
static quants of https://huggingface.co/langtech-dev/Salamandra-7b-RAG-v2
|
52 |
+
|
53 |
+
<!-- provided-files -->
|
54 |
+
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.
|
55 |
+
## Usage
|
56 |
+
|
57 |
+
If you are unsure how to use GGUF files, refer to one of [TheBloke's
|
58 |
+
READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for
|
59 |
+
more details, including on how to concatenate multi-part files.
|
60 |
+
|
61 |
+
## Provided Quants
|
62 |
+
|
63 |
+
(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
|
64 |
+
|
65 |
+
| Link | Type | Size/GB | Notes |
|
66 |
+
|:-----|:-----|--------:|:------|
|
67 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.Q2_K.gguf) | Q2_K | 3.4 | |
|
68 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.Q3_K_S.gguf) | Q3_K_S | 3.9 | |
|
69 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.Q3_K_M.gguf) | Q3_K_M | 4.1 | lower quality |
|
70 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.Q3_K_L.gguf) | Q3_K_L | 4.4 | |
|
71 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.IQ4_XS.gguf) | IQ4_XS | 4.6 | |
|
72 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.Q4_K_S.gguf) | Q4_K_S | 4.8 | fast, recommended |
|
73 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.Q4_K_M.gguf) | Q4_K_M | 5.0 | fast, recommended |
|
74 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.Q5_K_S.gguf) | Q5_K_S | 5.6 | |
|
75 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.Q5_K_M.gguf) | Q5_K_M | 5.7 | |
|
76 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.Q6_K.gguf) | Q6_K | 6.5 | very good quality |
|
77 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.Q8_0.gguf) | Q8_0 | 8.4 | fast, best quality |
|
78 |
+
| [GGUF](https://huggingface.co/mradermacher/Salamandra-7b-RAG-v2-GGUF/resolve/main/Salamandra-7b-RAG-v2.f16.gguf) | f16 | 15.6 | 16 bpw, overkill |
|
79 |
+
|
80 |
+
Here is a handy graph by ikawrakow comparing some lower-quality quant
|
81 |
+
types (lower is better):
|
82 |
+
|
83 |
+
![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)
|
84 |
+
|
85 |
+
And here are Artefact2's thoughts on the matter:
|
86 |
+
https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9
|
87 |
+
|
88 |
+
## FAQ / Model Request
|
89 |
+
|
90 |
+
See https://huggingface.co/mradermacher/model_requests for some answers to
|
91 |
+
questions you might have and/or if you want some other model quantized.
|
92 |
+
|
93 |
+
## Thanks
|
94 |
+
|
95 |
+
I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
|
96 |
+
me use its servers and providing upgrades to my workstation to enable
|
97 |
+
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.
|
98 |
+
|
99 |
+
<!-- end -->
|