bartowski commited on
Commit
bf0328a
1 Parent(s): f1c1527

measurement.json

Browse files
Files changed (2) hide show
  1. README.md +106 -0
  2. measurement.json +0 -0
README.md ADDED
@@ -0,0 +1,106 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ license_name: internlm2-20b
4
+ license_link: https://huggingface.co/internlm/internlm2-20b#open-source-license
5
+ base_model: internlm/internlm2-20b
6
+ datasets:
7
+ - ai2_arc
8
+ - allenai/ultrafeedback_binarized_cleaned
9
+ - argilla/distilabel-intel-orca-dpo-pairs
10
+ - jondurbin/airoboros-3.2
11
+ - codeparrot/apps
12
+ - facebook/belebele
13
+ - bluemoon-fandom-1-1-rp-cleaned
14
+ - boolq
15
+ - camel-ai/biology
16
+ - camel-ai/chemistry
17
+ - camel-ai/math
18
+ - camel-ai/physics
19
+ - jondurbin/contextual-dpo-v0.1
20
+ - jondurbin/gutenberg-dpo-v0.1
21
+ - jondurbin/py-dpo-v0.1
22
+ - jondurbin/truthy-dpo-v0.1
23
+ - LDJnr/Capybara
24
+ - jondurbin/cinematika-v0.1
25
+ - WizardLM/WizardLM_evol_instruct_70k
26
+ - glaiveai/glaive-function-calling-v2
27
+ - jondurbin/gutenberg-dpo-v0.1
28
+ - grimulkan/LimaRP-augmented
29
+ - lmsys/lmsys-chat-1m
30
+ - ParisNeo/lollms_aware_dataset
31
+ - TIGER-Lab/MathInstruct
32
+ - Muennighoff/natural-instructions
33
+ - openbookqa
34
+ - kingbri/PIPPA-shareGPT
35
+ - piqa
36
+ - Vezora/Tested-22k-Python-Alpaca
37
+ - ropes
38
+ - cakiki/rosetta-code
39
+ - Open-Orca/SlimOrca
40
+ - b-mc2/sql-create-context
41
+ - squad_v2
42
+ - mattpscott/airoboros-summarization
43
+ - migtissera/Synthia-v1.3
44
+ - unalignment/toxic-dpo-v0.2
45
+ - WhiteRabbitNeo/WRN-Chapter-1
46
+ - WhiteRabbitNeo/WRN-Chapter-2
47
+ - winogrande
48
+ quantized_by: bartowski
49
+ pipeline_tag: text-generation
50
+ ---
51
+
52
+ ## Exllama v2 Quantizations of bagel-20b-v04-llama
53
+
54
+ Using <a href="https://github.com/turboderp/exllamav2/releases/tag/v0.0.13">turboderp's ExLlamaV2 v0.0.13</a> for quantization.
55
+
56
+ <b>The "main" branch only contains the measurement.json, download one of the other branches for the model (see below)</b>
57
+
58
+ Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
59
+
60
+ Original model: https://huggingface.co/jondurbin/bagel-20b-v04-llama
61
+
62
+ | Branch | Bits | lm_head bits | VRAM (4k) | VRAM (16k) | VRAM (32k) | Description |
63
+ | ------ | ---- | ------------ | ---- | ---- | ---- | ----------- |
64
+ | [6_5](https://huggingface.co/bartowski/bagel-20b-v04-llama-exl2/tree/6_5) | 6.5 | 8.0 | 19.6 GB | 21.0 GB | 23.0 GB | Near unquantized performance at vastly reduced size, **recommended**. |
65
+ | [4_25](https://huggingface.co/bartowski/bagel-20b-v04-llama-exl2/tree/4_25) | 4.25 | 6.0 | 13.8 GB | 15.2 GB | 17.2 GB | GPTQ equivalent bits per weight, slightly higher quality. |
66
+ | [3_5](https://huggingface.co/bartowski/bagel-20b-v04-llama-exl2/tree/3_5) | 3.5 | 6.0 | 12.4 GB | 13.8 GB | 15.8 GB | Lower quality, only use if you have to. |
67
+ | [3_0](https://huggingface.co/bartowski/bagel-20b-v04-llama-exl2/tree/3_0) | 3.0 | 6.0 | 11.1 GB | 12.5 GB | 15.5 GB | Very low quality. Usable on 12GB. |
68
+
69
+ ## Download instructions
70
+
71
+ With git:
72
+
73
+ ```shell
74
+ git clone --single-branch --branch 6_5 https://huggingface.co/bartowski/bagel-20b-v04-llama-exl2 bagel-20b-v04-llama-exl2-6_5
75
+ ```
76
+
77
+ With huggingface hub (credit to TheBloke for instructions):
78
+
79
+ ```shell
80
+ pip3 install huggingface-hub
81
+ ```
82
+
83
+ To download the `main` (only useful if you only care about measurement.json) branch to a folder called `bagel-20b-v04-llama-exl2`:
84
+
85
+ ```shell
86
+ mkdir bagel-20b-v04-llama-exl2
87
+ huggingface-cli download bartowski/bagel-20b-v04-llama-exl2 --local-dir bagel-20b-v04-llama-exl2 --local-dir-use-symlinks False
88
+ ```
89
+
90
+ To download from a different branch, add the `--revision` parameter:
91
+
92
+ Linux:
93
+
94
+ ```shell
95
+ mkdir bagel-20b-v04-llama-exl2-6_5
96
+ huggingface-cli download bartowski/bagel-20b-v04-llama-exl2 --revision 6_5 --local-dir bagel-20b-v04-llama-exl2-6_5 --local-dir-use-symlinks False
97
+ ```
98
+
99
+ Windows (which apparently doesn't like _ in folders sometimes?):
100
+
101
+ ```shell
102
+ mkdir bagel-20b-v04-llama-exl2-6.5
103
+ huggingface-cli download bartowski/bagel-20b-v04-llama-exl2 --revision 6_5 --local-dir bagel-20b-v04-llama-exl2-6.5 --local-dir-use-symlinks False
104
+ ```
105
+
106
+ 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