aashish1904
commited on
Commit
•
e80580a
1
Parent(s):
ab09ffc
Upload README.md with huggingface_hub
Browse files
README.md
ADDED
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
---
|
3 |
+
|
4 |
+
license: apache-2.0
|
5 |
+
datasets:
|
6 |
+
- Epiculous/SynthRP-Gens-v1-Filtered-n-Cleaned
|
7 |
+
- Epiculous/Synthstruct-Gens-v1-Filtered-n-Cleaned
|
8 |
+
language:
|
9 |
+
- en
|
10 |
+
- fr
|
11 |
+
- de
|
12 |
+
- es
|
13 |
+
- it
|
14 |
+
- pt
|
15 |
+
- ru
|
16 |
+
- zh
|
17 |
+
- ja
|
18 |
+
pipeline_tag: text-generation
|
19 |
+
|
20 |
+
---
|
21 |
+
|
22 |
+
![](https://lh7-rt.googleusercontent.com/docsz/AD_4nXeiuCm7c8lEwEJuRey9kiVZsRn2W-b4pWlu3-X534V3YmVuVc2ZL-NXg2RkzSOOS2JXGHutDuyyNAUtdJI65jGTo8jT9Y99tMi4H4MqL44Uc5QKG77B0d6-JfIkZHFaUA71-RtjyYZWVIhqsNZcx8-OMaA?key=xt3VSDoCbmTY7o-cwwOFwQ)
|
23 |
+
|
24 |
+
# QuantFactory/Azure_Dusk-v0.1-GGUF
|
25 |
+
This is quantized version of [Epiculous/Azure_Dusk-v0.1](https://huggingface.co/Epiculous/Azure_Dusk-v0.1) created using llama.cpp
|
26 |
+
|
27 |
+
# Original Model Card
|
28 |
+
|
29 |
+
|
30 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/64adfd277b5ff762771e4571/l0b889iXmZy-dz-Yg8e1i.png)
|
31 |
+
|
32 |
+
Flipping the training process that created Crimson Dawn on it's head, I present to you, Azure Dusk! While both models are built using [Mistral-Nemo-Base-2407](https://huggingface.co/mistralai/Mistral-Nemo-Base-2407); Azure Dusk's training methodology was instruct first, then RP dataset applied after, however, the end goal reamains the same AI should not be a boring bland generic assistant, but something that you can connect with on a more personal level. Something that can be interesting in a Roleplay, but useful as an assistant too.
|
33 |
+
|
34 |
+
# Quants!
|
35 |
+
<strong>full</strong> / [exl2](https://huggingface.co/Epiculous/Azure_Dusk-v0.1-Exl2) / [gguf](https://huggingface.co/Epiculous/Azure_Dusk-v0.1-GGUF)
|
36 |
+
|
37 |
+
## Prompting
|
38 |
+
Azure Dusk was trained with the Mistral Instruct template, therefore it should be prompted in a similar way that you would prompt any other mistral based model.
|
39 |
+
|
40 |
+
```
|
41 |
+
"<s>[INST] Prompt goes here [/INST]<\s>"
|
42 |
+
```
|
43 |
+
### Context and Instruct
|
44 |
+
[Magnum-123B-Context.json](https://files.catbox.moe/rkyqwg.json) <br/>
|
45 |
+
[Magnum-123B-Instruct.json](https://files.catbox.moe/obb5oe.json) <br/>
|
46 |
+
*** NOTE *** <br/>
|
47 |
+
There have been reports of the quantized model misbehaving with the mistral prompt, if you are seeing issues it may be worth trying ChatML Context and Instruct templates.
|
48 |
+
If you are using GGUF I strongly advise using ChatML, for some reason that quantization performs better using ChatML.
|
49 |
+
### Current Top Sampler Settings
|
50 |
+
[Violet_Twilight-Nitral-Special](https://files.catbox.moe/ot54u3.json)- Considered the best settings! <br/>
|
51 |
+
[Crimson_Dawn-Nitral-Special](https://files.catbox.moe/8xjxht.json) <br/>
|
52 |
+
[Crimson_Dawn-Magnum-Style](https://files.catbox.moe/lc59dn.json)
|
53 |
+
|
54 |
+
### Tokenizer
|
55 |
+
If you are using SillyTavern, please set the tokenizer to API (WebUI/ koboldcpp)
|
56 |
+
|
57 |
+
## Training
|
58 |
+
Training was done twice over 2 epochs each on two 2x [NVIDIA A6000 GPUs](https://www.nvidia.com/en-us/design-visualization/rtx-a6000/) using LoRA. A two-phased approach was used in which the base model was trained 2 epochs on Instruct data, the LoRA was then applied to base. Finally, the new modified base was trained 2 epochs on RP, and the new RP LoRA was applied to the modified base, resulting in what you see here.
|
59 |
+
|
60 |
+
[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
|
61 |
+
|
62 |
+
## Special Thanks
|
63 |
+
Special thanks to my friends over at Anthracite! Without their help and Kalomaze starting the synthetic data script, none of this would have been possible.
|
64 |
+
Also want to thank my friends in The Chaotic Neutrals for their friendship, support, and guidance.
|