SFT-DecSelfMask-Llama-3.2-1B-Instruct

This model is a fine-tuned version of ferrazzipietro/DecSelfMask-Llama-3.2-1B-Instruct on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 4.7891
• F1 Micro: 0.0
• F1 Macro: 0.0
• F1 Weighted: 0.0
• Class/f1 Results Per Class: {}
• Items/f1 Scores Per Item: {'Cardiac tamponade': 0.0, 'Cardiovascular diseases (except for HF)': 0.0, 'TLoC during effort': 0.0, 'Level of consciousness': 0.0, 'Administration of oxygen/ventilation': 0.0, 'Situation description suggestive for situational sincope (coughing, straining, sudden abdominal pain, phlebotomy)': 0.0, 'Respiratory failure': 0.0, 'Blood pressure': 0.0, 'Thoracic ultrasound, any abnormalities': 0.0, 'Neurodegenerative disease': 0.0, 'Level of autonomy (mobility)': 0.0, 'Respiratory rate': 0.0, 'Blood in the stool': 0.0, 'Pulmonary embolism': 0.0, 'History of drug abuse': 0.0, 'Agitation [acute condition]': 0.0, 'Platelets': 0.0, 'History of recent trauma': 0.0, 'Presence of defibrillator': 0.0, 'Pulmonary scintigraphy, any abnormality': 0.0, 'Presence of pacemaker': 0.0, 'Antihypertensive therapy': 0.0, 'Cardiac ultrasound, any abnormality': 0.0, 'Hemorrhage': 0.0, 'Chest Rx, any abnormalities': 0.0, 'SpO2 in aa': 0.0, 'SARS-CoV-2 swab test': 0.0, 'Carotid sinus massage': 0.0, 'Acute coronary syndrome': 0.0, 'Speed with which the patient recovered consciousness': 0.0, 'Transaminase': 0.0, 'History of drug allergy': 0.0, 'Improvement of dyspnea': 0.0, 'Drooling during the episode': 0.0, 'Creatinine': 0.0, 'General condition deterioration': 0.0, 'Stiffness during the episode': 0.0, 'Presence of dyspnea': 0.0, 'Blood sodium': 0.0, 'PaC02': 0.0, "Improvement of patient's conditions": 0.0, 'Brain CT scan, any abnormality': 0.0, 'C-Reactive Protein': 0.0}

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0003
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-07 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 1

Training results

Framework versions

• PEFT 0.14.0
• Transformers 4.49.0.dev0
• Pytorch 2.5.1+cu124
• Datasets 3.6.0
• Tokenizers 0.21.0