PEFT
Safetensors
Russian
qwen2
legal
contract-extraction
information-extraction
json-generation
lora
qwen
russian
enterprise
Instructions to use zieglerd/RussianConcsExt with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use zieglerd/RussianConcsExt with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-32B-Instruct") model = PeftModel.from_pretrained(base_model, "zieglerd/RussianConcsExt") - Notebooks
- Google Colab
- Kaggle
Commit History
-
Update README.md
Verified signatureda0bb95 -
Update README.md
Verified signatureac5c94b -
Update README.md
Verified signature5ae218e -
Update README.md
Verified signature02515de -
Upload folder using huggingface_hub
Verified signature4265e47 -
Upload folder using huggingface_hub
Verified signaturee15105e -
Update README.md
Dmitry ZieglerVerified signature###Descripton This LoRA adapter is fine-tuned for automated extraction of structured information from Russian-language contracts. The adapter extends the Qwen2.5-14B-Instruct model with specialized capabilities for enterprise legal document understanding. The model receives a full contract text in free-form format and generates a structured JSON response containing extracted contractual information, including document subject, financial conditions, VAT parameters, advance payment schedules, and specification data. The base model used for fine-tuning is Qwen2.5-14B-Instruct. On the Russian Purchase Contracts benchmark, the base model demonstrates approximately 85% accuracy on Russian contract understanding and extraction tasks. This LoRA adapter further specializes the model for real-world corporate contracts, improving robustness in complex document scenarios that are difficult for general-purpose language models. The model is designed to process unstructured Russian legal documents…3157115 -
initial commit
Dmitry ZieglerVerified signature4be373a