Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
tags:
|
6 |
+
- story
|
7 |
+
- roleplay
|
8 |
+
- creative
|
9 |
+
- rp
|
10 |
+
- fantasy
|
11 |
+
- story telling
|
12 |
+
- 32 bit upscale
|
13 |
+
- ultra high precision
|
14 |
+
- nsfw
|
15 |
+
---
|
16 |
+
[uploads in progress]
|
17 |
+
|
18 |
+
<B> Ultra High Quality - 20B Dark Forest Version 3.0 - 32 bit upscale - Imatrix Plus 2 </b>
|
19 |
+
|
20 |
+
Fully rebuilt from master files, including full merge(s) to maintain full 32 bit precision right
|
21 |
+
up until it is compressed into GGUF files which results on a top to bottom upgrade.
|
22 |
+
|
23 |
+
The result is superior performance in instruction following, reasoning, depth, nuance and emotion.
|
24 |
+
|
25 |
+
This is the Imatrix Plus 2 repo. Imatrix Plus 2 uses the Ultra Quality 32 bit upscale as a base with
|
26 |
+
applied in house datasets to further enhance the model's quality and performance.
|
27 |
+
|
28 |
+
NOTE: There are three original versions of "Dark Forest 20B", this is an upscale of the third version, with
|
29 |
+
links below to 1st and 2nd versions also upscaled.
|
30 |
+
|
31 |
+
<img src="dark-forest.jpg">
|
32 |
+
|
33 |
+
On average this means a q4km operates at Q6 levels and Q6 and Q8 exceeds original model full precision performance.
|
34 |
+
|
35 |
+
Perplexity drop (lower is better) is close to 10% (over 752 points for q4km) for all quants.
|
36 |
+
|
37 |
+
That means precision has been enhanced for all 20 billion parameters which affects "brain density" / "function",
|
38 |
+
instruction following and output quality.
|
39 |
+
|
40 |
+
Imatrix quants to follow shortly.
|
41 |
+
|
42 |
+
For more details, including a list of enhancements see our other 32 bit
|
43 |
+
upscale of "Space Whale 20B" rebuild here:
|
44 |
+
|
45 |
+
[ https://huggingface.co/DavidAU/Psyonic-Cetacean-Ultra-Quality-20b-GGUF ]
|
46 |
+
|
47 |
+
For Version 1 of Dark Forest Ultra Quality 32 bit upscale please go here:
|
48 |
+
|
49 |
+
[ https://huggingface.co/DavidAU/Dark-Forest-V1-Ultra-Quality-20b-GGUF ]
|
50 |
+
|
51 |
+
For Version 1 of Dark Forest Ultra Quality 32 bit upscale please go here:
|
52 |
+
|
53 |
+
[ https://huggingface.co/TeeZee/DarkForest-20B-v2.0 ]
|
54 |
+
|
55 |
+
Special thanks to "TEEZEE" for making a both fantasic models of "Dark Forest".
|
56 |
+
|
57 |
+
<b> Info from the original model card: </B>
|
58 |
+
|
59 |
+
Warning: This model can produce NSFW content!
|
60 |
+
|
61 |
+
Results:
|
62 |
+
|
63 |
+
- main difference to v1.0 - model has much better sense of humor.
|
64 |
+
- produces SFW nad NSFW content without issues, switches context seamlessly.
|
65 |
+
- good at following instructions.
|
66 |
+
- good at tracking multiple characters in one scene.
|
67 |
+
- very creative, scenarios produced are mature and complicated, model doesn't shy from writing about PTSD, menatal issues or complicated relationships.
|
68 |
+
- NSFW output is more creative and suprising than typical limaRP output.
|
69 |
+
- definitely for mature audiences, not only because of vivid NSFW content but also because of overall maturity of stories it produces.
|
70 |
+
- This is NOT Harry Potter level storytelling.
|
71 |
+
|
72 |
+
|
73 |
+
For original model spec and information please visit:
|
74 |
+
|
75 |
+
[ https://huggingface.co/TeeZee/DarkForest-20B-v3.0 ]
|