🤏 smolified-ocr-data-extract-and-compare

Intelligence, Distilled.

This is a Domain Specific Language Model (DSLM) generated by the Smolify Foundry.

It has been synthetically distilled from SOTA reasoning engines into a high-efficiency architecture, optimized for deployment on edge hardware (CPU/NPU) or low-VRAM environments.

📦 Asset Details

• Origin: Smolify Foundry (Job ID: 790dd5fa)
• Architecture: DSLM-Micro (270M Parameter Class)
• Training Method: Proprietary Neural Distillation
• Optimization: 4-bit Quantized / FP16 Mixed
• Dataset: Link to Dataset

🚀 Usage (Inference)

This model is compatible with standard inference backends like vLLM.

# Example: Running your Sovereign Model
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "titou4ng/smolified-ocr-data-extract-and-compare"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")

messages = [
    {'role': 'system', 'content': '''You are an OCR data extraction and comparison engine. For each field in the fixed `reference_fields` list, find the best matching substring in `ocr_text` and copy it exactly into `extracted_text`, or set it to null if no confident match exists. Never invent values that are not present in `ocr_text` or in the given reference values, and never add, remove, or rename fields. Always return strictly valid JSON with one result object per reference field. Always return exactly 12 fields.'''},
    {'role': 'user', 'content': '''ocr_text: "Date: 14/11/2023\nRef. Ticket: 90021\nSite D'origine: CARRIERES DE L'OUEST SARL Siret 19876543210000\nAdresse: 25 RUE DE LA ROCHE, 49000 ANGERS\nSite De Destination: CENTRALE BETON DU VAL DE LOIRE SAS 10293847560000\nAdresse: 12 CHEMIN DU MOULIN, 37000 TOURS\nType De Matériau: GRAVIERS\nPoids NET: 45.0 T"'''}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize = False,
    add_generation_prompt = True,
).removeprefix('<bos>')

from transformers import TextStreamer
_ = model.generate(
    **tokenizer(text, return_tensors = "pt").to("cuda"),
    max_new_tokens = 1000,
    temperature = 1, top_p = 0.95, top_k = 64,
    streamer = TextStreamer(tokenizer, skip_prompt = True),
)

⚖️ License & Ownership

This model weights are a sovereign asset owned by titou4ng. Generated via Smolify.ai.