🤏 smolified-helo

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: c158c554)
• 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 = "smolify/smolified-helo"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")

messages = [
    {'role': 'system', 'content': '''I'm having trouble with my professor. He's always late for class, cancels without notice, and barely covers the syllabus. I feel like I'm not learning anything in this course and my grades are suffering. What can I do?'''},
    {'role': 'user', 'content': '''**GRIEVANCE RESOLUTION GUIDE**  **Issue Identified:** Ineffective teaching practices and syllabus incompletion **Severity Level:** NORMAL **Domain:** Instructor Conduct  **SITUATION ANALYSIS:** Student concerned about the quality of instruction, leading to learning and grade issues. Requires formal academic intervention.  **RECOMMENDED WORKFLOW:** Multi-step institutional process requiring documentation and escalation  **KEY OFFICES TO CONTACT:** Student Affairs, Dean's Office, Academic Affairs  **ESTIMATED TIMELINE:** Varies by case complexity: 1-8 weeks typical  **CRITICAL ACTIONS:** 1. Document everything - keep written records, emails, receipts 2. Act promptly - many processes are time-sensitive 3. Follow escalation hierarchy if initial attempts fail 4. Maintain professional communication throughout 5. Seek support from faculty advisor or student affairs if overwhelmed  **SEVERITY CONSIDERATIONS:** 	NORMAL: Respond within 5-7 business days.  **EMOTIONAL SUPPORT:** This situation is stressful but manageable. Follow the steps methodically and don't hesitate to ask for help from student affairs.'''}
]
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 smolify. Generated via Smolify.ai.