Fine-Tuned LoRA Adapters for Mixtral-8x7B CEFR Model

• Base Model: unsloth/noSynthetic-mixtral_3epoch_02dropout_base
• Fine-Tuning: LoRA with SMOTE-balanced dataset
• Training Details:
  • Dataset: CEFR-level sentences with SMOTE and undersampling for balance
  • LoRA Parameters: r=32, lora_alpha=32, lora_dropout=0.2
  • Training Args: learning_rate=1e-5, batch_size=8, epochs=3, cosine scheduler
  • Optimizer: adamw_8bit
  • Early Stopping: Patience=2, threshold=0.01
• Evaluation Metrics (Exact Matches):
  • CEFR Classifier Accuracy: 0.167
  • Precision (Macro): 0.042
  • Recall (Macro): 0.167
  • F1-Score (Macro): 0.067
• Evaluation Metrics (Within ±1 Level):
  • CEFR Classifier Accuracy: 0.500
  • Precision (Macro): 0.306
  • Recall (Macro): 0.500
  • F1-Score (Macro): 0.361
• Other Metrics:
  • Perplexity: 9.119
  • Diversity (Unique Sentences): 0.033
  • Inference Time (ms): 1043.625
  • Model Size (GB): 28.0 (base model + LoRA adapters)
  • Robustness (F1): 0.063
• Confusion Matrix (Exact Matches):
  • CSV: confusion_matrix_exact.csv
  • Image: confusion_matrix_exact.png
• Confusion Matrix (Within ±1 Level):
  • CSV: confusion_matrix_within1.csv
  • Image: confusion_matrix_within1.png
• Per-Class Confusion Metrics (Exact Matches):
  • A1: TP=0, FP=0, FN=10, TN=50
  • A2: TP=10, FP=30, FN=0, TN=20
  • B1: TP=0, FP=20, FN=10, TN=30
  • B2: TP=0, FP=0, FN=10, TN=50
  • C1: TP=0, FP=0, FN=10, TN=50
  • C2: TP=0, FP=0, FN=10, TN=50
• Per-Class Confusion Metrics (Within ±1 Level):
  • A1: TP=10, FP=0, FN=0, TN=50
  • A2: TP=10, FP=10, FN=0, TN=40
  • B1: TP=10, FP=20, FN=0, TN=30
  • B2: TP=0, FP=0, FN=10, TN=50
  • C1: TP=0, FP=0, FN=10, TN=50
  • C2: TP=0, FP=0, FN=10, TN=50
• Usage:
  python

  1from transformers import AutoModelForCausalLM, AutoTokenizer 2from peft import PeftModel 3 4# Load base model 5base_model = AutoModelForCausalLM.from_pretrained( 6 "unsloth/noSynthetic-mixtral_3epoch_02dropout_base", 7 quantization_config=BitsAndBytesConfig(load_in_4bit=True), 8 device_map="auto" 9) 10tokenizer = AutoTokenizer.from_pretrained("unsloth/noSynthetic-mixtral_3epoch_02dropout_base") 11 12# Load LoRA adapters 13model = PeftModel.from_pretrained(base_model, "Mr-FineTuner/noSynthetic-mixtral_3epoch_02dropout_lora") 14 15# Example inference 16prompt = "<s>[INST] Generate a CEFR B1 level sentence. [/INST]" 17inputs = tokenizer(prompt, return_tensors="pt").to(model.device) 18outputs = model.generate(**inputs, max_length=50) 19print(tokenizer.decode(outputs[0], skip_special_tokens=True))