bartowski commited on
Commit
29d5072
1 Parent(s): c5655a9

measurement.json

Browse files
Files changed (2) hide show
  1. README.md +64 -0
  2. measurement.json +0 -0
README.md ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: unknown
3
+ tags:
4
+ - merge
5
+ quantized_by: bartowski
6
+ pipeline_tag: text-generation
7
+ ---
8
+
9
+ ## Exllama v2 Quantizations of Everyone-LLM-7b-Base
10
+
11
+ Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.13">turboderp's ExLlamaV2 v0.0.13</a> for quantization.
12
+
13
+ <b>The "main" branch only contains the measurement.json, download one of the other branches for the model (see below)</b>
14
+
15
+ Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
16
+
17
+ Original model: https://huggingface.co/rombodawg/Everyone-LLM-7b-Base
18
+
19
+ | Branch | Bits | lm_head bits | VRAM (4k) | VRAM (16k) | VRAM (32k) | Description |
20
+ | ----- | ---- | ------- | ------ | ------ | ------ | ------------ |
21
+ | [8_0](https://huggingface.co/bartowski/Everyone-LLM-7b-Base-exl2/tree/8_0) | 8.0 | 8.0 | 8.4 GB | 9.8 GB | 11.8 GB | Maximum quality that ExLlamaV2 can produce, near unquantized performance. |
22
+ | [6_5](https://huggingface.co/bartowski/Everyone-LLM-7b-Base-exl2/tree/6_5) | 6.5 | 8.0 | 7.2 GB | 8.6 GB | 10.6 GB | Very similar to 8.0, good tradeoff of size vs performance, **recommended**. |
23
+ | [5_0](https://huggingface.co/bartowski/Everyone-LLM-7b-Base-exl2/tree/5_0) | 5.0 | 6.0 | 6.0 GB | 7.4 GB | 9.4 GB | Slightly lower quality vs 6.5, but usable on 8GB cards. |
24
+ | [4_25](https://huggingface.co/bartowski/Everyone-LLM-7b-Base-exl2/tree/4_25) | 4.25 | 6.0 | 5.3 GB | 6.7 GB | 8.7 GB | GPTQ equivalent bits per weight, slightly higher quality. |
25
+ | [3_5](https://huggingface.co/bartowski/Everyone-LLM-7b-Base-exl2/tree/3_5) | 3.5 | 6.0 | 4.7 GB | 6.1 GB | 8.1 GB | Lower quality, only use if you have to. |
26
+
27
+ ## Download instructions
28
+
29
+ With git:
30
+
31
+ ```shell
32
+ git clone --single-branch --branch 6_5 https://huggingface.co/bartowski/Everyone-LLM-7b-Base-exl2 Everyone-LLM-7b-Base-exl2-6_5
33
+ ```
34
+
35
+ With huggingface hub (credit to TheBloke for instructions):
36
+
37
+ ```shell
38
+ pip3 install huggingface-hub
39
+ ```
40
+
41
+ To download the `main` (only useful if you only care about measurement.json) branch to a folder called `Everyone-LLM-7b-Base-exl2`:
42
+
43
+ ```shell
44
+ mkdir Everyone-LLM-7b-Base-exl2
45
+ huggingface-cli download bartowski/Everyone-LLM-7b-Base-exl2 --local-dir Everyone-LLM-7b-Base-exl2 --local-dir-use-symlinks False
46
+ ```
47
+
48
+ To download from a different branch, add the `--revision` parameter:
49
+
50
+ Linux:
51
+
52
+ ```shell
53
+ mkdir Everyone-LLM-7b-Base-exl2-6_5
54
+ huggingface-cli download bartowski/Everyone-LLM-7b-Base-exl2 --revision 6_5 --local-dir Everyone-LLM-7b-Base-exl2-6_5 --local-dir-use-symlinks False
55
+ ```
56
+
57
+ Windows (which apparently doesn't like _ in folders sometimes?):
58
+
59
+ ```shell
60
+ mkdir Everyone-LLM-7b-Base-exl2-6.5
61
+ huggingface-cli download bartowski/Everyone-LLM-7b-Base-exl2 --revision 6_5 --local-dir Everyone-LLM-7b-Base-exl2-6.5 --local-dir-use-symlinks False
62
+ ```
63
+
64
+ Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
measurement.json ADDED
The diff for this file is too large to render. See raw diff