TheDrummer
commited on
Commit
•
fae0881
1
Parent(s):
fbf1f05
Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: other
|
3 |
+
---
|
4 |
+
# Join our Discord! https://discord.gg/Nbv9pQ88Xb
|
5 |
+
---
|
6 |
+
|
7 |
+
[BeaverAI](https://huggingface.co/BeaverAI) & [Steelskull](https://huggingface.co/Steelskull) proudly present...
|
8 |
+
|
9 |
+
# Theia 21B v1 🧘
|
10 |
+
|
11 |
+
#### An upscaled NeMo with half its layers trained on my special sauce 💦
|
12 |
+
|
13 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/qMWMRWJTuKvDgIwiPvr-5.png)
|
14 |
+
|
15 |
+
*According to the giant impact hypothesis, Theia orbited the Sun, nearly along the orbit of the proto-Earth, by staying close to one or the other of the Sun-Earth system's two more stable Lagrangian points.*
|
16 |
+
|
17 |
+
![image/gif](https://cdn-uploads.huggingface.co/production/uploads/65f2fd1c25b848bd061b5c2e/a41bOgBNuF714OWlPcCeY.gif)
|
18 |
+
|
19 |
+
*Theia was eventually perturbed away from that relationship by the gravitational influence of Jupiter, Venus, or both, resulting in a collision between Theia and Earth.*
|
20 |
+
|
21 |
+
## Description
|
22 |
+
|
23 |
+
- This is a finetune of [Steelskull's NeMoria 21B](https://huggingface.co/TheSkullery/NeMoria-21b) model.
|
24 |
+
- You can check out his official merges here: https://huggingface.co/Steelskull.
|
25 |
+
- The finetuning process involved completion, RP, and instruct training.
|
26 |
+
- Testers have reported enhanced capabilities and richer prose.
|
27 |
+
|
28 |
+
## Links
|
29 |
+
|
30 |
+
- Original: https://huggingface.co/TheDrummer/Theia-21B-v1
|
31 |
+
- GGUF: https://huggingface.co/TheDrummer/Theia-21B-v1-GGUF
|
32 |
+
|
33 |
+
## Usage
|
34 |
+
|
35 |
+
- Mistral format
|
36 |
+
- Make sure that there is a space between `[/INST]` and `{{char}}:` like `[/INST] {{char}}:`.
|
37 |
+
- `[/INST]{{char}}:` (no space between) has a huge negative impact on performance.
|