Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cc-by-nc-4.0
|
3 |
+
tags:
|
4 |
+
- not-for-all-audiences
|
5 |
+
- nsfw
|
6 |
+
---
|
7 |
+
|
8 |
+
This quant was made for [infermatic.ai](https://infermatic.ai/)
|
9 |
+
|
10 |
+
Dynamic FP8 quant of [Lumimaid-v0.2-70B-FP8-Dynamic](https://huggingface.co/NeverSleep/Lumimaid-v0.2-70B?not-for-all-audiences=true) made with AutoFP8.
|
11 |
+
|
12 |
+
## Lumimaid 0.2
|
13 |
+
<img src="https://cdn-uploads.huggingface.co/production/uploads/63ab1241ad514ca8d1430003/HY1KTq6FMAm-CwmY8-ndO.png" alt="Image" style="display: block; margin-left: auto; margin-right: auto; width: 65%;">
|
14 |
+
<div style="text-align: center; font-size: 30px;">
|
15 |
+
<a href="https://huggingface.co/NeverSleep/Lumimaid-v0.2-8B">8b</a> -
|
16 |
+
<a href="https://huggingface.co/NeverSleep/Lumimaid-v0.2-12B">12b</a> -
|
17 |
+
<a href="https://huggingface.co/NeverSleep/Lumimaid-v0.2-70B">[70b]</a> -
|
18 |
+
<a href="https://huggingface.co/NeverSleep/Lumimaid-v0.2-123B">123b</a>
|
19 |
+
</div>
|
20 |
+
|
21 |
+
### This model is based on: [Meta-Llama-3.1-70B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct)
|
22 |
+
Wandb: https://wandb.ai/undis95/Lumi-Llama-3-1-70B?nw=nwuserundis95
|
23 |
+
|
24 |
+
Lumimaid 0.1 -> 0.2 is a HUGE step up dataset wise.
|
25 |
+
|
26 |
+
As some people have told us our models are sloppy, Ikari decided to say fuck it and literally nuke all chats out with most slop.
|
27 |
+
|
28 |
+
Our dataset stayed the same since day one, we added data over time, cleaned them, and repeat. After not releasing model for a while because we were never satisfied, we think it's time to come back!
|
29 |
+
|
30 |
+
## Prompt template: Llama-3-Instruct
|
31 |
+
|
32 |
+
```
|
33 |
+
<|begin_of_text|><|start_header_id|>system<|end_header_id|>
|
34 |
+
|
35 |
+
{system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|>
|
36 |
+
|
37 |
+
{input}<|eot_id|><|start_header_id|>assistant<|end_header_id|>
|
38 |
+
|
39 |
+
{output}<|eot_id|>
|
40 |
+
```
|
41 |
+
|
42 |
+
## Credits:
|
43 |
+
- Undi
|
44 |
+
- IkariDev
|
45 |
+
|
46 |
+
## Training data we used to make our dataset:
|
47 |
+
|
48 |
+
- [Epiculous/Gnosis](https://huggingface.co/Epiculous/Gnosis)
|
49 |
+
- [ChaoticNeutrals/Luminous_Opus](https://huggingface.co/datasets/ChaoticNeutrals/Luminous_Opus)
|
50 |
+
- [ChaoticNeutrals/Synthetic-Dark-RP](https://huggingface.co/datasets/ChaoticNeutrals/Synthetic-Dark-RP)
|
51 |
+
- [ChaoticNeutrals/Synthetic-RP](https://huggingface.co/datasets/ChaoticNeutrals/Synthetic-RP)
|
52 |
+
- [Gryphe/Sonnet3.5-SlimOrcaDedupCleaned](https://huggingface.co/datasets/Gryphe/Sonnet3.5-SlimOrcaDedupCleaned)
|
53 |
+
- [Gryphe/Opus-WritingPrompts](https://huggingface.co/datasets/Gryphe/Opus-WritingPrompts)
|
54 |
+
- [meseca/writing-opus-6k](https://huggingface.co/datasets/meseca/writing-opus-6k)
|
55 |
+
- [meseca/opus-instruct-9k](https://huggingface.co/datasets/meseca/opus-instruct-9k)
|
56 |
+
- [PJMixers/grimulkan_theory-of-mind-ShareGPT](https://huggingface.co/datasets/PJMixers/grimulkan_theory-of-mind-ShareGPT)
|
57 |
+
- [NobodyExistsOnTheInternet/ToxicQAFinal](https://huggingface.co/datasets/NobodyExistsOnTheInternet/ToxicQAFinal)
|
58 |
+
- [Undi95/toxic-dpo-v0.1-sharegpt](https://huggingface.co/datasets/Undi95/toxic-dpo-v0.1-sharegpt)
|
59 |
+
- [cgato/SlimOrcaDedupCleaned](https://huggingface.co/datasets/cgato/SlimOrcaDedupCleaned)
|
60 |
+
- [kalomaze/Opus_Instruct_25k](https://huggingface.co/datasets/kalomaze/Opus_Instruct_25k)
|
61 |
+
- [Doctor-Shotgun/no-robots-sharegpt](https://huggingface.co/datasets/Doctor-Shotgun/no-robots-sharegpt)
|
62 |
+
- [Norquinal/claude_multiround_chat_30k](https://huggingface.co/datasets/Norquinal/claude_multiround_chat_30k)
|
63 |
+
- [nothingiisreal/Claude-3-Opus-Instruct-15K](https://huggingface.co/datasets/nothingiisreal/Claude-3-Opus-Instruct-15K)
|
64 |
+
- All the Aesirs dataset, cleaned, unslopped
|
65 |
+
- All le luminae dataset, cleaned, unslopped
|
66 |
+
- Small part of Airoboros reduced
|
67 |
+
|
68 |
+
We sadly didn't find the sources of the following, DM us if you recognize your set !
|
69 |
+
|
70 |
+
- Opus_Instruct-v2-6.5K-Filtered-v2-sharegpt
|
71 |
+
- claude_sharegpt_trimmed
|
72 |
+
- CapybaraPure_Decontaminated-ShareGPT_reduced
|
73 |
+
|
74 |
+
## Datasets credits:
|
75 |
+
- Epiculous
|
76 |
+
- ChaoticNeutrals
|
77 |
+
- Gryphe
|
78 |
+
- meseca
|
79 |
+
- PJMixers
|
80 |
+
- NobodyExistsOnTheInternet
|
81 |
+
- cgato
|
82 |
+
- kalomaze
|
83 |
+
- Doctor-Shotgun
|
84 |
+
- Norquinal
|
85 |
+
- nothingiisreal
|
86 |
+
|
87 |
+
## Others
|
88 |
+
|
89 |
+
Undi: If you want to support us, you can [here](https://ko-fi.com/undiai).
|
90 |
+
|
91 |
+
IkariDev: Visit my [retro/neocities style website](https://ikaridevgit.github.io/) please kek
|