bartowski commited on
Commit
c4abd66
·
verified ·
1 Parent(s): 9ef3c67

measurement.json

Browse files
Files changed (2) hide show
  1. README.md +79 -0
  2. measurement.json +0 -0
README.md ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: llama2
3
+ library_name: transformers
4
+ tags:
5
+ - code
6
+ model-index:
7
+ - name: Code Millenials
8
+ results:
9
+ - task:
10
+ type: text-generation
11
+ dataset:
12
+ type: openai_humaneval
13
+ name: HumanEval
14
+ metrics:
15
+ - name: pass@1
16
+ type: pass@1
17
+ value: 0.8048
18
+ verified: false
19
+ quantized_by: bartowski
20
+ pipeline_tag: text-generation
21
+ ---
22
+
23
+ ## Exllama v2 Quantizations of code-millenials-34b
24
+
25
+ Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.12">turboderp's ExLlamaV2 v0.0.12</a> for quantization.
26
+
27
+ ## The "main" branch only contains the measurement.json, download one of the other branches for the model (see below)
28
+
29
+ Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
30
+
31
+ Conversion was done using the default calibration dataset.
32
+
33
+ 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.
34
+
35
+ Original model: https://huggingface.co/budecosystem/code-millenials-34b
36
+
37
+
38
+ <a href="https://huggingface.co/bartowski/code-millenials-34b-exl2/tree/6_5">6.5 bits per weight</a>
39
+ <a href="https://huggingface.co/bartowski/code-millenials-34b-exl2/tree/5_0">5.0 bits per weight</a>
40
+ <a href="https://huggingface.co/bartowski/code-millenials-34b-exl2/tree/4_25">4.25 bits per weight</a>
41
+ <a href="https://huggingface.co/bartowski/code-millenials-34b-exl2/tree/3_75">3.75 bits per weight</a>
42
+ <a href="https://huggingface.co/bartowski/code-millenials-34b-exl2/tree/3_0">3.0 bits per weight</a>
43
+
44
+ ## Download instructions
45
+
46
+ With git:
47
+
48
+ ```shell
49
+ git clone --single-branch --branch 6_5 https://huggingface.co/bartowski/code-millenials-34b-exl2
50
+ ```
51
+
52
+ With huggingface hub (credit to TheBloke for instructions):
53
+
54
+ ```shell
55
+ pip3 install huggingface-hub
56
+ ```
57
+
58
+ To download the `main` (only useful if you only care about measurement.json) branch to a folder called `code-millenials-34b-exl2`:
59
+
60
+ ```shell
61
+ mkdir code-millenials-34b-exl2
62
+ huggingface-cli download bartowski/code-millenials-34b-exl2 --local-dir code-millenials-34b-exl2 --local-dir-use-symlinks False
63
+ ```
64
+
65
+ To download from a different branch, add the `--revision` parameter:
66
+
67
+ Linux:
68
+
69
+ ```shell
70
+ mkdir code-millenials-34b-exl2-6_5
71
+ huggingface-cli download bartowski/code-millenials-34b-exl2 --revision 6_5 --local-dir code-millenials-34b-exl2-6_5 --local-dir-use-symlinks False
72
+ ```
73
+
74
+ Windows (which apparently doesn't like _ in folders sometimes?):
75
+
76
+ ```shell
77
+ mkdir code-millenials-34b-exl2-6.5
78
+ huggingface-cli download bartowski/code-millenials-34b-exl2 --revision 6_5 --local-dir code-millenials-34b-exl2-6.5 --local-dir-use-symlinks False
79
+ ```
measurement.json ADDED
The diff for this file is too large to render. See raw diff