KamilHugsFaces/t5-gemma-2-reasoning-query-match-v4

Fine-tuned T5-Gemma-2 model for entailment classification.

Training Details

• Base model: google/t5gemma-2-4b-4b
• Training variant: query_match_v1
• Epochs: 3
• Batch size: 1
• Learning rate: 1e-05
• Run name: query_match_v1_20260114_231334

Training Data

• Training examples: 215
• Validation examples: 46
• Test examples: 49
• Class weights: {'true': 1.0, 'false': 4.0}

Evaluation Results

Test Set Performance

• F1 Score: 0.7424
• F1 (False class): 0.4615
• Accuracy: 0.7143
• Precision (False): 0.3529
• Recall (False): 0.6667

Usage

from transformers import AutoModelForSeq2SeqLM, AutoTokenizer

model = AutoModelForSeq2SeqLM.from_pretrained("KamilHugsFaces/t5-gemma-2-reasoning-query-match-v4")
tokenizer = AutoTokenizer.from_pretrained("KamilHugsFaces/t5-gemma-2-reasoning-query-match-v4")

# Format input
input_text = "entailment: [Your claim and evidence here]"
inputs = tokenizer(input_text, return_tensors="pt", max_length=250, truncation=True)

# Generate prediction
outputs = model.generate(**inputs, max_new_tokens=8)
prediction = tokenizer.decode(outputs[0], skip_special_tokens=True)
# Output: "true" or "false"

Training Configuration

{ "variant_name": "query_match_v1", "run_name": "query_match_v1_20260114_231334", "num_epochs": 3, "batch_size": 1, "learning_rate": 1e-05, "warmup_steps": 100, "model_name": "google/t5gemma-2-4b-4b", "class_weights": { "true": 1.0, "false": 4.0 }, "use_confidence_weighting": true, "confidence_weight_alpha": 2, "train_size": 215, "val_size": 46, "test_size": 49 }

Framework

• Transformers: 5.0.0.dev0
• PyTorch: 2.9.1+cu128
• Trained on: Modal (A100 GPU)