Text Generation
English
AI hybrid software
AI AutoCorrect Software
AI AutoCreative Enhancement
AI Auto Low Quant Enhancement and Optimize
AUTO Reconsider Systems
Live Streaming Auto Enhancement in Real Time
ALL GGUF enhanced
ALL EXL2 enhanced
ALL HQQ enhanced
ALL Full source enhanced
ALL GPTQ enhanced
ALL AWQ enhanced
creative
creative writing
fiction writing
plot generation
sub-plot generation
story generation
scene continue
storytelling
fiction story
science fiction
romance
all genres
story
writing
vivid writing
fiction
sillytavern enhancement
horror
llama 2
llama 3
llama 3.1
llama 3.2
llama 3.3
Qwen
command-r
mistral
mistral large
deepseek
gemma
phi
Update README.md
Browse files
README.md
CHANGED
@@ -79,12 +79,24 @@ of the negatives such as special settings, parameters and/or samplers to "reign"
|
|
79 |
Class 2, 3 and 4 models have incredible creative "chops" but the downside is they can be very difficult to use - AutoCorrect engine
|
80 |
is the automated, hands off answer to this issue.
|
81 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
82 |
And... it also allows the release of stronger models at class 5 and above - yet allow users (including myself) - to be able to
|
83 |
use these models under all conditions WITHOUT the need to manually stop, fix, or steer generation NOR set parameters
|
84 |
and/or samplers to "reign" in the bad behaviour of these models.
|
85 |
|
86 |
Yep - have your cake and eat it too.
|
87 |
|
|
|
|
|
88 |
Because many of the issues with "creative" / "unstable" models are the same as using low and ultra low quants (Q2k, IQ1s,IQ2s) this
|
89 |
engine also fixes on the fly many of these issues too, resulting in the ability to use models at these ultra low quant levels with enhanced results.
|
90 |
|
|
|
79 |
Class 2, 3 and 4 models have incredible creative "chops" but the downside is they can be very difficult to use - AutoCorrect engine
|
80 |
is the automated, hands off answer to this issue.
|
81 |
|
82 |
+
In otherwords the following models (and all their brothers and sisters) will now operate perfectly, under all conditions:
|
83 |
+
|
84 |
+
https://huggingface.co/DavidAU/L3-DARKEST-PLANET-16.5B-GGUF
|
85 |
+
|
86 |
+
https://huggingface.co/DavidAU/MN-DARKEST-UNIVERSE-29B-GGUF
|
87 |
+
|
88 |
+
https://huggingface.co/DavidAU/L3-Stheno-Maid-Blackroot-Grand-HORROR-16B-GGUF
|
89 |
+
|
90 |
+
And many more...
|
91 |
+
|
92 |
And... it also allows the release of stronger models at class 5 and above - yet allow users (including myself) - to be able to
|
93 |
use these models under all conditions WITHOUT the need to manually stop, fix, or steer generation NOR set parameters
|
94 |
and/or samplers to "reign" in the bad behaviour of these models.
|
95 |
|
96 |
Yep - have your cake and eat it too.
|
97 |
|
98 |
+
And all models - my repo and others - will all operate better.
|
99 |
+
|
100 |
Because many of the issues with "creative" / "unstable" models are the same as using low and ultra low quants (Q2k, IQ1s,IQ2s) this
|
101 |
engine also fixes on the fly many of these issues too, resulting in the ability to use models at these ultra low quant levels with enhanced results.
|
102 |
|