AI Emergency Kit - crisis-agent-v1

AI Emergency Kit - Your intelligent crisis response assistant. Fine-tuned Mistral-7B model that provides structured, actionable guidance during emergency situations.

Model Details

• Base Model: unsloth/Mistral-7B-Instruct-v0.2
• Fine-tuning Method: LoRA (Low-Rank Adaptation)
• Model Type: merged model
• Training Dataset: ianktoo/crisis-response-training

Usage

Using Unsloth

from unsloth import FastLanguageModel

model, tokenizer = FastLanguageModel.from_pretrained(
    "ianktoo/crisis-agent-v1",
    max_seq_length=2048,
    dtype=None,
    load_in_4bit=True,
)

# Format prompt
prompt = "<s>[INST] Your crisis scenario here [/INST]"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

# Generate response
outputs = model.generate(**inputs, max_new_tokens=512, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)

Using Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("ianktoo/crisis-agent-v1")
tokenizer = AutoTokenizer.from_pretrained("ianktoo/crisis-agent-v1")

# Use model for inference

About AI Emergency Kit

AI Emergency Kit is designed to be your reliable AI companion during crisis situations. It provides structured, JSON-formatted responses with actionable guidance, resource recommendations, and step-by-step instructions to help navigate emergency scenarios.

Training Details

This model was fine-tuned using the crisis-agent fine-tuning pipeline to create the AI Emergency Kit.

Limitations

• Model is trained on synthetic crisis scenarios
• Responses should be validated by human experts
• Not intended for real-time emergency response without human oversight

Citation

If you use this model, please cite:

@software{crisis_agent_model,
  title = {Crisis Agent Model},
  author = {Ian Too},
  year = {2026},
  url = {https://huggingface.co/ianktoo/crisis-agent-v1}
}