Update README.md
Browse files
README.md
CHANGED
|
@@ -4,7 +4,7 @@ license: cc-by-4.0
|
|
| 4 |
|
| 5 |
# #roleplay
|
| 6 |
|
| 7 |
-
GGUF-IQ-Imatrix quants for **cgato/L3-TheSpice-8b-v0.
|
| 8 |
|
| 9 |
> [!IMPORTANT]
|
| 10 |
> These quants have been done after the fixes from [llama.cpp/pull/6920](https://github.com/ggerganov/llama.cpp/pull/6920). <br>
|
|
@@ -15,72 +15,3 @@ GGUF-IQ-Imatrix quants for **cgato/L3-TheSpice-8b-v0.8.3**.
|
|
| 15 |
|
| 16 |
|
| 17 |
|
| 18 |
-
# Original model information by the author:
|
| 19 |
-
|
| 20 |
-
Now not overtrained and with the tokenizer fix to base llama3. Trained for 3 epochs.
|
| 21 |
-
|
| 22 |
-
The latest TheSpice, dipped in Mama Liz's LimaRP Oil.
|
| 23 |
-
I've focused on making the model more flexible and provide a more unique experience.
|
| 24 |
-
I'm still working on cleaning up my dataset, but I've shrunken it down a lot to focus on a "less is more" approach.
|
| 25 |
-
This is ultimate a return to form of the way I used to train Thespis, with more of a focus on a small hand edited dataset.
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
## Datasets Used
|
| 29 |
-
|
| 30 |
-
* Capybara
|
| 31 |
-
* Claude Multiround 30k
|
| 32 |
-
* Augmental
|
| 33 |
-
* ToxicQA
|
| 34 |
-
* Yahoo Answers
|
| 35 |
-
* Airoboros 3.1
|
| 36 |
-
* LimaRP
|
| 37 |
-
|
| 38 |
-
## Features ( Examples from 0.1.1 because I'm too lazy to take new screenshots. Its tested tho. )
|
| 39 |
-
|
| 40 |
-
Narration
|
| 41 |
-
|
| 42 |
-
If you request information on objects or characters in the scene, the model will narrate it to you. Most of the time, without moving the story forward.
|
| 43 |
-
|
| 44 |
-
# You can look at anything mostly as long as you end it with "What do I see?"
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
# You can also request to know what a character is thinking or planning.
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
# You can ask for a quick summary on the character as well.
|
| 53 |
-
|
| 54 |
-
|
| 55 |
-
|
| 56 |
-
# Before continuing the conversation as normal.
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
## Prompt Format: Chat ( The default Ooba template and Silly Tavern Template )
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
If you're using Ooba in verbose mode as a server, you can check if you're console is logging something that looks like this.
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
```
|
| 68 |
-
{System Prompt}
|
| 69 |
-
|
| 70 |
-
Username: {Input}
|
| 71 |
-
BotName: {Response}
|
| 72 |
-
Username: {Input}
|
| 73 |
-
BotName: {Response}
|
| 74 |
-
|
| 75 |
-
```
|
| 76 |
-
## Presets
|
| 77 |
-
|
| 78 |
-
All screenshots above were taken with the below SillyTavern Preset.
|
| 79 |
-
## Recommended Silly Tavern Preset -> (Temp: 1.25, MinP: 0.1, RepPen: 1.05)
|
| 80 |
-
This is a roughly equivalent Kobold Horde Preset.
|
| 81 |
-
## Recommended Kobold Horde Preset -> MinP
|
| 82 |
-
|
| 83 |
-
|
| 84 |
-
# Disclaimer
|
| 85 |
-
|
| 86 |
-
Please prompt responsibly and take anything outputted by any Language Model with a huge grain of salt. Thanks!
|
|
|
|
| 4 |
|
| 5 |
# #roleplay
|
| 6 |
|
| 7 |
+
GGUF-IQ-Imatrix quants for **cgato/L3-TheSpice-8b-v0.1.3**.
|
| 8 |
|
| 9 |
> [!IMPORTANT]
|
| 10 |
> These quants have been done after the fixes from [llama.cpp/pull/6920](https://github.com/ggerganov/llama.cpp/pull/6920). <br>
|
|
|
|
| 15 |
|
| 16 |
|
| 17 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|