bartowski commited on
Commit
a8ce121
1 Parent(s): 109ed41

measurement.json

Browse files
Files changed (2) hide show
  1. README.md +82 -0
  2. measurement.json +0 -0
README.md ADDED
@@ -0,0 +1,82 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ license_link: https://huggingface.co/microsoft/Phi-3-mini-4k-instruct/resolve/main/LICENSE
4
+
5
+ language:
6
+ - en
7
+ pipeline_tag: text-generation
8
+ tags:
9
+ - nlp
10
+ - code
11
+ inference:
12
+ parameters:
13
+ temperature: 0.0
14
+ widget:
15
+ - messages:
16
+ - role: user
17
+ content: Can you provide ways to eat combinations of bananas and dragonfruits?
18
+ quantized_by: bartowski
19
+ ---
20
+
21
+ ## Exllama v2 Quantizations of Phi-3.1-mini-4k-instruct
22
+
23
+ Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.1.6">turboderp's ExLlamaV2 v0.1.6</a> for quantization.
24
+
25
+ <b>The "main" branch only contains the measurement.json, download one of the other branches for the model (see below)</b>
26
+
27
+ Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
28
+
29
+ Conversion was done using the default calibration dataset.
30
+
31
+ Default arguments used except when the bits per weight is above 6.0, at that point the lm_head layer is quantized at 8 bits per weight instead of the default 6.
32
+
33
+ Original model: https://huggingface.co/microsoft/Phi-3-mini-4k-instruct
34
+
35
+
36
+ <a href="https://huggingface.co/bartowski/Phi-3.1-mini-4k-instruct-exl2/tree/8_0">8.0 bits per weight</a>
37
+
38
+ <a href="https://huggingface.co/bartowski/Phi-3.1-mini-4k-instruct-exl2/tree/6_5">6.5 bits per weight</a>
39
+
40
+ <a href="https://huggingface.co/bartowski/Phi-3.1-mini-4k-instruct-exl2/tree/5_0">5.0 bits per weight</a>
41
+
42
+ <a href="https://huggingface.co/bartowski/Phi-3.1-mini-4k-instruct-exl2/tree/4_25">4.25 bits per weight</a>
43
+
44
+ <a href="https://huggingface.co/bartowski/Phi-3.1-mini-4k-instruct-exl2/tree/3_5">3.5 bits per weight</a>
45
+
46
+
47
+ ## Download instructions
48
+
49
+ With git:
50
+
51
+ ```shell
52
+ git clone --single-branch --branch 6_5 https://huggingface.co/bartowski/Phi-3.1-mini-4k-instruct-exl2
53
+ ```
54
+
55
+ With huggingface hub (credit to TheBloke for instructions):
56
+
57
+ ```shell
58
+ pip3 install huggingface-hub
59
+ ```
60
+
61
+ To download the `main` (only useful if you only care about measurement.json) branch to a folder called `Phi-3.1-mini-4k-instruct-exl2`:
62
+
63
+ ```shell
64
+ mkdir Phi-3.1-mini-4k-instruct-exl2
65
+ huggingface-cli download bartowski/Phi-3.1-mini-4k-instruct-exl2 --local-dir Phi-3.1-mini-4k-instruct-exl2
66
+ ```
67
+
68
+ To download from a different branch, add the `--revision` parameter:
69
+
70
+ Linux:
71
+
72
+ ```shell
73
+ mkdir Phi-3.1-mini-4k-instruct-exl2-6_5
74
+ huggingface-cli download bartowski/Phi-3.1-mini-4k-instruct-exl2 --revision 6_5 --local-dir Phi-3.1-mini-4k-instruct-exl2-6_5
75
+ ```
76
+
77
+ Windows (which apparently doesn't like _ in folders sometimes?):
78
+
79
+ ```shell
80
+ mkdir Phi-3.1-mini-4k-instruct-exl2-6.5
81
+ huggingface-cli download bartowski/Phi-3.1-mini-4k-instruct-exl2 --revision 6_5 --local-dir Phi-3.1-mini-4k-instruct-exl2-6.5
82
+ ```
measurement.json ADDED
The diff for this file is too large to render. See raw diff