Model Card for Mistral-7B for Story Generation

Model Description

This model is a fine-tuned Mistral-7B model on stories from the WritingPrompts dataset.

Language(s) (NLP): English
Finetuned from model: m-elio/Mistral-BookCorpus-Gutenberg
Dataset used for fine-tuning: WritingPrompts

Example of Usage

import torch

from transformers import AutoModelForCausalLM, AutoTokenizer
from transformers.trainer_utils import set_seed

set_seed(42)

model_id = "m-elio/Mistral-BookCorpus-Gutenberg-Writing-Prompts"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")

instruction_text = "Write a story for the writing prompt provided as input"
input_text = "A story about a dancer who tries to win the National championship."

prompt = "Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n" \
        f"### Instruction:\nWrite a story for the writing prompt provided as input\n\n" \
        f"### Input:\n{input_text}\n\n" \
        f"### Answer:\n"

input_ids = tokenizer(prompt, return_tensors="pt").input_ids
outputs = model.generate(input_ids=input_ids, top_k=0, top_p=0.92, do_sample=True, max_new_tokens=2048)

print(tokenizer.batch_decode(outputs.detach().cpu().numpy()[:, input_ids.shape[1]:], skip_special_tokens=True)[0])