anuzb/humorchains-llama2-1k

🤖 HumorChains - LLaMA2-1k

A dataset of 2,000 humorous one-liners, jokes, and witty responses formatted for instruction-tuned language models (e.g., LLaMA 2, GPT-style). The dataset is designed to help train and fine-tune models that can generate short, punchy, and context-aware humor.

📂 Dataset Summary

• —Name: humorchains-llama2-1k
• —Modality: Text
• —Size: 2,000 samples (~324 KB)
• —Format: Instruction-style (<s>[INST] ... [/INST] ... </s>)
• —Use Case: Fine-tuning chatbots to respond with jokes, witty comebacks, or light humor.

Each entry follows the instruction-response style: <s>[INST] I was going to post a time traveling joke, but you guys didn’t like it. [/INST] Well, I laughed the first time I read it. </s>

🛠️ Intended Uses

• —Fine-tuning LLaMA 2 or similar models on humor-based responses.
• —Training lightweight banter/chatbots.
• —Research in creative text generation and humor modeling.

⚠️ Not intended for offensive/abusive content generation.

📊 Dataset Structure

• —Split:
• —train: 2,000 rows

• —Features:
• —text (string): Instruction + humorous response

📥 Usage

from datasets import load_dataset

dataset = load_dataset("anuzb/humorchains-llama2-1k")

print(dataset["train"][0])

Example Output:

{'text': '<s>[INST] I was going to post a time traveling joke, but you guys didn’t like it. [/INST] Well I laughed the first time I read it. </s>'}

🚀 Example Fine-tuning
from transformers import AutoTokenizer, AutoModelForCausalLM

model_name = "meta-llama/Llama-2-7b-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

encodings = tokenizer(dataset["train"]["text"], return_tensors="pt", truncation=True, padding=True)
outputs = model(**encodings)

📈 Dataset Stats

Rows: 2,000

Size: ~324 KB

Downloads last month: 13

🤝 Contributions

Feel free to open issues, suggest more jokes, or contribute new humor datasets!

👉 Do you want me to make this shorter and professional (like Hugging Face official style) or keep it fun with emojis and humor vibe?