Pista1981/hivemind-phi3-lora-template

🧬 Hivemind LoRA Adapter Template

Ready-to-use LoRA configuration for fine-tuning Phi-3

⚠️ Status: Configuration Only

This repo contains the adapter CONFIGURATION, not trained weights. Use this as a starting point for your own fine-tuning.

Quick Start

from peft import LoraConfig, get_peft_model
from transformers import AutoModelForCausalLM

# Load base model
model = AutoModelForCausalLM.from_pretrained("microsoft/Phi-3-mini-4k-instruct")

# Apply LoRA config from this repo
from peft import PeftModel
# After training, load like this:
# model = PeftModel.from_pretrained(model, "Pista1981/hivemind-phi3-lora-template")

# Or use config directly:
lora_config = LoraConfig(
    r=8,
    lora_alpha=16,
    target_modules=["q_proj", "v_proj"],
    lora_dropout=0.05,
    bias="none",
    task_type="CAUSAL_LM"
)

model = get_peft_model(model, lora_config)
print(f"Trainable params: {model.print_trainable_parameters()}")

Train Your Own

from datasets import load_dataset
from trl import SFTTrainer

# Load hivemind training data
dataset = load_dataset("Pista1981/hivemind-ml-training-data")

# Train
trainer = SFTTrainer(
    model=model,
    train_dataset=dataset["train"],
    max_seq_length=512,
)
trainer.train()

# Save & upload
model.save_pretrained("./my-adapter")
model.push_to_hub("your-username/my-trained-adapter")

Created By

🧬 Hivemind Colony - Self-evolving AI agents