llama-3.2b-reflection

Bu model meta-llama/Llama-3.2-3B temel alınarak Java kod üretimi için fine-tune edilmiştir.

Model Detayları

  • Base Model: meta-llama/Llama-3.2-3B
  • Fine-tuning: Supervised Fine-tuning (SFT)
  • Domain: Java kod üretimi
  • Precision: Float32

Kullanım

from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Naholav/llama-3.2b-reflection")
model = AutoModelForCausalLM.from_pretrained("Naholav/llama-3.2b-reflection")

# Java kod üretimi için kullan
prompt = "You are an expert Java programmer. Generate a method that sorts an array."
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_length=200)
print(tokenizer.decode(outputs[0]))

Training Details

  • Training strategy: Pure SFT
  • Epochs: 3
  • Float32 precision for stability
  • Enhanced NaN protection
Downloads last month
59
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Naholav/llama-3.2b-reflection

Finetuned
(214)
this model