Base Model: Mistral-7B-Instruct-v0.3
Fine-tuning Method: LoRA (Low-Rank Adaptation)
Training Data: ZENT platform conversations, documentation, and AI transmissions
Context Length: 8192 tokens
License: Apache 2.0

Intended Use

This model is designed for:

Powering AI agents on token launchpads
Crypto community chatbots
DeFi assistant applications
Blockchain education
Creating derivative AI agents

Usage

With Transformers

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "ZENTSPY/zent-agentic-7b"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

messages = [
    {"role": "system", "content": "You are ZENT AGENTIC, an autonomous AI agent for the ZENT Launchpad on Solana."},
    {"role": "user", "content": "How do I launch a token?"}
]

inputs = tokenizer.apply_chat_template(messages, return_tensors="pt")
outputs = model.generate(inputs, max_new_tokens=512)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

With Inference API

python

import requests

API_URL = "https://api-inference.huggingface.co/models/ZENTSPY/zent-agentic-7b"
headers = {"Authorization": "Bearer YOUR_HF_TOKEN"}

def query(payload):
    response = requests.post(API_URL, headers=headers, json=payload)
    return response.json()

output = query({
    "inputs": "What is ZENT Agentic Launchpad?",
})

With llama.cpp (GGUF)

bash

./main -m zent-agentic-7b.Q4_K_M.gguf \
  -p "You are ZENT AGENTIC. User: What is ZENT? Assistant:" \
  -n 256

Training Details

Training Data

Platform documentation and guides
User conversation examples
AI transmission content (23 types)
Quest and rewards information
Technical blockchain content

Training Hyperparameters

Learning Rate: 2e-5
Batch Size: 4
Gradient Accumulation: 4
Epochs: 3
LoRA Rank: 64
LoRA Alpha: 128
Target Modules: q_proj, k_proj, v_proj, o_proj

Hardware

GPU: NVIDIA A100 80GB
Training Time: ~4 hours

Evaluation

Metric	Score
ZENT Knowledge Accuracy	94.2%
Response Coherence	4.6/5.0
Personality Consistency	4.8/5.0
Helpfulness	4.5/5.0

Limitations

Knowledge cutoff based on training data
May hallucinate specific numbers/prices
Best used with retrieval augmentation for real-time data
Optimized for English only

Ethical Considerations

Not financial advice
Users should DYOR
Model may have biases from training data
Intended for educational/entertainment purposes

Citation

bibtex

@misc{zent-agentic-2024,
  author = {ZENTSPY},
  title = {ZENT AGENTIC: Fine-tuned LLM for Solana Token Launchpad},
  year = {2024},
  publisher = {Hugging Face},
  url = {https://huggingface.co/ZENTSPY/zent-agentic-7b}
}

Contact

For questions, issues, or collaborations:

Open an issue on GitHub
DM on Twitter @ZENTSPY
Join our community

Built with 💜 by ZENT Protocol

ZENTSPY/zent-agentic-7b

ZENT AGENTIC Model 🤖

Model Description

Model Details

Intended Use

Usage

With Transformers

With Inference API

With llama.cpp (GGUF)

Training Details

Training Data

Training Hyperparameters

Hardware

Evaluation

Limitations

Ethical Considerations

Citation

Links

Contact

ZENTSPY/
zent-agentic-7b