Update README.md

README.md

@@ -3,65 +3,53 @@ license: apache-2.0
 language:
 - en
 base_model:
-- mistralai/Mistral-Small-24B-Instruct-2501
+- meta-llama/Llama-3.2-3B-Instruct
 tags:
 - not-for-all-audiences
 ---
 
 GGUF:
 
-https://huggingface.co/mradermacher/CardProjector-24B-v1-GGUF
-https://huggingface.co/mradermacher/CardProjector-24B-v1-i1-GGUF
-https://huggingface.co/AlexBefest/CardProjector-24B-v1-GGUF
-
 
 [WoonaAi](https://huggingface.co/WoonaAI) presents...
 
-# CardProjector-24B-v1
+# CardProjector-3B-v1.1
 
-![image/png](https://cdn-uploads.huggingface.co/production/uploads/6336c5b3e3ac69e6a90581da/geRIv7Q6jbNixWbRnrHx8.png)
 
-## Overview
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6336c5b3e3ac69e6a90581da/lYuI5jEFDrDSxBkFffjjq.png)
 
-CardProjector-24B-v1 is a specialized language model derived from Mistral-Small-24B-Instruct-2501, fine-tuned to generate character cards for **SillyTavern** in the **chara_card_v2** specification. This model is designed to assist creators and roleplayers by automating the process of crafting detailed and well-structured character cards, ensuring compatibility with SillyTavern's format.
 
----
+## What's new in v1.1?
 
-## Samples:
+- Fixed endless loops
+- The dataset has been significantly redesigned, which made it possible to achieve quite high quality on the 3B model (in some cases, even almost similar quality to the CardProjector 24B v1).
+- Significantly improved the stability of the model behavior
+- Added the ability to generate character cards using a Chain of Thoughts (for the CardProjector-R1-preview-8B-v1.1), which helped to achieve a higher generation quality than CardProjector 24B v1 in most cases. Read more [here](https://huggingface.co/AlexBefest/CardProjector-R1-preview-8B-v1.1).
 
-![image/png](https://cdn-uploads.huggingface.co/production/uploads/6336c5b3e3ac69e6a90581da/1L3X7Std5loXePHea_kEi.png)
+## Overview
 
-![image/png](https://cdn-uploads.huggingface.co/production/uploads/6336c5b3e3ac69e6a90581da/NFqZ6Dc8MZh4Tb6aa-q-B.png)
+CardProjector is a specialized series of language models, fine-tuned to generate character cards for **SillyTavern** in the **chara_card_v2** specification. These models are designed to assist creators and roleplayers by automating the process of crafting detailed and well-structured character cards, ensuring compatibility with SillyTavern's format.
 
-![image/png](https://cdn-uploads.huggingface.co/production/uploads/6336c5b3e3ac69e6a90581da/8-jBpTJ3q3dlz_pAzwDfU.png)
+---
+
+## Sample:
 
-[More samples in json](https://huggingface.co/AlexBefest/CardProjector-24B-v1/tree/main/Samples)
 
 ---
 
 ## Parameters
 
-- **Chat Template**: Mistral v7 Tekken
+- **Chat Template**: Llama-3
 - **Sampler parameters**:
 ```plain text
-Temperature: 0.3-0.7
-Rep. pen.: 1.1-1.2 (Recommended!)
-Top P: 1
-Min P: 0.05-0.1
-Max output: 4096
+Temperature: 0.7-0.85
+Rep. pen.: 1.1
 ```
 - **System prompt (recommended)**:
 ```plain text
 You are an AI assistant helping the user generate character cards for a role-playing game for SillyTavern, using the strict SillyTavern json format in the chara_card_v2 specification. 
 ```
 
-## Notes
-
-**This model has significant problems with infinite loops during generation. This issue will be fixed in future versions.**
-
-This model is ideal for the basis of creating your own card, it supports multi-turn conservation, so you can work with the card directly in the chat. It is recommended to use a UI that does not replace Placeholder Tags ({{user}} {{char}}) for ease of working with the card (For example, KoboldUI Classic in Editing Mode)
-Sometimes there is a problem with endless generation loops, so I strongly recommend setting the parameter rep pen 1.1-1.2.
-
 ## Safety
 
 Wha?