tpai-lora-cerebro

TPAI LoRA - Cerebro Interno

LoRA adapter for Qwen/Qwen2.5-0.5B-Instruct, trained as the "cerebro interno" (internal brain) for the TPAI (Teaching People AI) system.

Model Details

• Base Model: Qwen/Qwen2.5-0.5B-Instruct
• Adapter Type: LoRA (PEFT)
• LoRA Rank (r): 16
• LoRA Alpha: 32
• Target Modules: q_proj, k_proj, o_proj, v_proj
• Task Type: Causal Language Modeling

Training Details

• Training Samples: 1194+
• Epochs: 3
• Training Duration: ~8.5 hours
• Final Loss: 1.659
• Dataset: Multi-language programming assistant responses (TeachingPeopleAI/tpai-training-data)
• LoRA Dropout: 0.05

Usage

from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

# Load base model
base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-0.5B-Instruct")
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-0.5B-Instruct")

# Load LoRA adapter
model = PeftModel.from_pretrained(base_model, "Gil-tester/tpai-lora-cerebro")

# Generate response
input_text = "User: Explain quantum computing\nAssistant:"
inputs = tokenizer(input_text, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=100)
print(tokenizer.decode(outputs[0]))

Integration with TPAI

This model serves as the "cerebro interno" in the TPAI system, providing:

• Primary language generation
• Low-latency responses via local inference
• Cost-effective operation on HF Spaces

Framework Versions

• PEFT 0.18.1