Mr-FineTuner/Test_02_noFinetune_Mistral_myValidator

Non-Fine-Tuned Mistral-7B CEFR Evaluation

This repository contains the evaluation results of the base unsloth/mistral-7b-instruct-v0.3-bnb-4bit model for CEFR-level sentence generation, without fine-tuning, as part of an ablation study. The model is evaluated using a fine-tuned classifier from Mr-FineTuner/Skripsi_validator_best_model.

• Base Model: unsloth/mistral-7b-instruct-v0.3-bnb-4bit
• Evaluation Details:
  • Dataset: Rebalanced test dataset (test_merged_output.txt), which was also used to train and evaluate the classifier, potentially introducing bias.
  • No fine-tuning performed; base model used directly.
  • Classifier: MLP classifier trained on train_merged_output.txt, dev_merged_output.txt, and test_merged_output.txt for CEFR level prediction.
• Evaluation Metrics (Exact Matches):
  • CEFR Classifier Accuracy: 0.333
  • Precision (Macro): 0.139
  • Recall (Macro): 0.333
  • F1-Score (Macro): 0.194
• Evaluation Metrics (Within ±1 Level):
  • CEFR Classifier Accuracy: 0.667
  • Precision (Macro): 0.556
  • Recall (Macro): 0.667
  • F1-Score (Macro): 0.583
• Other Metrics:
  • Perplexity: 953.237
  • Diversity (Unique Sentences): 0.100
  • Inference Time (ms): 2352.285
  • Model Size (GB): 7.0
  • Robustness (F1): 0.185
• 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=0, FP=0, FN=10, TN=50
  • B1: TP=10, FP=10, FN=0, TN=40
  • B2: TP=10, FP=20, FN=0, TN=30
  • C1: TP=0, FP=0, FN=10, TN=50
  • C2: TP=0, FP=10, FN=10, TN=40
• Per-Class Confusion Metrics (Within ±1 Level):
  • A1: TP=0, FP=0, FN=10, TN=50
  • A2: TP=10, FP=0, FN=0, TN=50
  • B1: TP=10, FP=0, FN=0, TN=50
  • B2: TP=10, FP=20, FN=0, TN=30
  • C1: TP=10, FP=0, FN=0, TN=50
  • C2: TP=0, FP=0, FN=10, TN=50
• Note on Bias:
  • The test dataset used for evaluation (test_merged_output.txt) was part of the training and evaluation data for the classifier (Mr-FineTuner/Skripsi_validator_best_model). This may lead to inflated performance metrics due to the classifier's familiarity with the dataset. For a more robust evaluation, a new dataset not used in classifier training is recommended.
• Usage:

  from transformers import AutoModelForCausalLM, AutoTokenizer
  
  model = AutoModelForCausalLM.from_pretrained("unsloth/mistral-7b-instruct-v0.3-bnb-4bit")
  tokenizer = AutoTokenizer.from_pretrained("unsloth/mistral-7b-instruct-v0.3-bnb-4bit")
  
  # Example inference
  prompt = "[INST] Generate a CEFR B1 level sentence. [/INST]"
  inputs = tokenizer(prompt, return_tensors="pt")
  outputs = model.generate(**inputs, max_length=50)
  print(tokenizer.decode(outputs[0], skip_special_tokens=True))
  

Uploaded using huggingface_hub.