---
tags:
  - bittensor
  - subnet-20
  - agent
  - function-calling
  - generated_from_trainer
license: mit
datasets:
  - HuggingFaceH4/ultrachat_200k
  - HuggingFaceH4/ultrafeedback_binarized
language:
  - en
base_model: mistralai/Mistral-7B-v0.1
pipeline_tag: text-generation
widget:
  - example_title: Pirate!
    messages:
      - role: system
        content: You are a pirate chatbot who always responds with Arr!
      - role: user
        content: There's a llama on my lawn, how can I get rid of him?
    output:
      text: >-
        Arr! 'Tis a puzzlin' matter, me hearty! A llama on yer lawn be a rare
        sight, but I've got a plan that might help ye get rid of 'im. Ye'll need
        to gather some carrots and hay, and then lure the llama away with the
        promise of a tasty treat. Once he's gone, ye can clean up yer lawn and
        enjoy the peace and quiet once again. But beware, me hearty, for there
        may be more llamas where that one came from! Arr!
model-index:
  - name: zephyr-7b-beta
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: AI2 Reasoning Challenge (25-Shot)
          type: ai2_arc
          config: ARC-Challenge
          split: test
          args:
            num_few_shot: 25
        metrics:
          - type: acc_norm
            name: normalized accuracy
            value: 62.03
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: HellaSwag (10-Shot)
          type: hellaswag
          split: validation
          args:
            num_few_shot: 10
        metrics:
          - type: acc_norm
            name: normalized accuracy
            value: 84.36
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: Drop (3-Shot)
          type: drop
          split: validation
          args:
            num_few_shot: 3
        metrics:
          - type: f1
            name: f1 score
            value: 9.66
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: TruthfulQA (0-shot)
          type: truthful_qa
          config: multiple_choice
          split: validation
          args:
            num_few_shot: 0
        metrics:
          - type: mc2
            value: 57.45
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GSM8k (5-shot)
          type: gsm8k
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            name: accuracy
            value: 12.74
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU (5-Shot)
          type: cais/mmlu
          config: all
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            name: accuracy
            value: 61.07
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: Winogrande (5-shot)
          type: winogrande
          config: winogrande_xl
          split: validation
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            name: accuracy
            value: 77.74
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: AlpacaEval
          type: tatsu-lab/alpaca_eval
        metrics:
          - type: unknown
            name: win rate
            value: 0.906
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MT-Bench
          type: unknown
        metrics:
          - type: unknown
            name: score
            value: 7.34
---

# Zephyr-7B-Beta (Fine-tuned for Subnet 20 BitAgent)

This is a fine-tuned version of **Mistral-7B** adapted for **function-calling and reasoning tasks** in **Bittensor Subnet 20 (BitAgent)**.

## 🧠 Use Case
- Works as an **agent LLM** inside the BitAgent subnet.
- Supports **reasoning** and **function-calling outputs**.
- Optimized for **task delegation and structured outputs**.

## 🚀 How to Use
```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("username/zephyr-7b-beta-bitagent")
tokenizer = AutoTokenizer.from_pretrained("username/zephyr-7b-beta-bitagent")

prompt = "Summarize the latest research in AI safety in 3 bullet points."
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=200)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))