ColQwen2.5 XTR Pruned
Fine-tuned with XTR Loss + LoRA + Token Pruning
Training Info
- Epoch: 1
- Global Step: 136
- Final Loss: 0.1882
- LoRA rank: 32
- LoRA alpha: 64
- Pruning ratio: 0.65
Usage
from peft import PeftModel
from colpali_engine.models import ColQwen2, ColQwen2Processor
# Load base model
base_model = ColQwen2.from_pretrained("vidore/colqwen2-v0.1")
# Load LoRA adapter
model = PeftModel.from_pretrained(base_model, "DungND1107/colqwen2_5_xtr_pruned")
# Load processor
processor = ColQwen2Processor.from_pretrained("DungND1107/colqwen2_5_xtr_pruned")
- Downloads last month
- 13
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support