Instructions to use CasperYL/pascal-unsloth-mistral-lora-en with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use CasperYL/pascal-unsloth-mistral-lora-en with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("unsloth/mistral-7b-instruct-v0.3-bnb-4bit") model = PeftModel.from_pretrained(base_model, "CasperYL/pascal-unsloth-mistral-lora-en") - Notebooks
- Google Colab
- Kaggle
- Local Apps
- Unsloth Studio new
How to use CasperYL/pascal-unsloth-mistral-lora-en with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for CasperYL/pascal-unsloth-mistral-lora-en to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for CasperYL/pascal-unsloth-mistral-lora-en to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for CasperYL/pascal-unsloth-mistral-lora-en to start chatting
Load model with FastModel
pip install unsloth from unsloth import FastModel model, tokenizer = FastModel.from_pretrained( model_name="CasperYL/pascal-unsloth-mistral-lora-en", max_seq_length=2048, )
Pascal Mistral LoRA EN
A small LoRA adapter fine-tuned to imitate the English "deep thought" style of Pascal, the relaxed sea otter philosopher from Animal Crossing.
This is a fan/research project for local experimentation with character-style fine-tuning, maaan. It is not affiliated with Nintendo, Animal Crossing, or any official rights holder.
Base Model
This adapter was trained on:
unsloth/mistral-7b-instruct-v0.3-bnb-4bit
This is a LoRA adapter only. It is not a standalone model.
Intended Use
Best for short, single-turn Pascal-style responses:
- casual island chatter
- surreal little observations
- "deep thought" style quote generation
- local LoRA inference experiments
Limitations
- Works best in single-turn mode.
- Multi-turn chat can drift or become repetitive.
- It may produce incorrect, strange, or non-canonical lines.
- It imitates style; it does not retrieve an authoritative quote database.
- Ordinary direct questions may produce more generic chatbot answers.
Example Prompt
Player: I found something shiny near the shore.
Pascal:
Loading Example
from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer
adapter_id = "CasperYL/pascal-unsloth-mistral-lora-en"
tokenizer = AutoTokenizer.from_pretrained(adapter_id)
base_model = AutoModelForCausalLM.from_pretrained(
"unsloth/mistral-7b-instruct-v0.3-bnb-4bit",
device_map="auto",
)
model = PeftModel.from_pretrained(base_model, adapter_id)
Training Data
The adapter was trained from a small, single-turn Alpaca-style dataset built from Pascal "deep thought" quotes and hand-written player prompts.
License / Rights Note
This is a fan-made adapter for research and personal experimentation. Animal Crossing and Pascal are properties of their respective rights holders. Please respect the base model license and any applicable rights around source material.
- Downloads last month
- 73
Model tree for CasperYL/pascal-unsloth-mistral-lora-en
Base model
mistralai/Mistral-7B-v0.3