GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
art0xx
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
art0xx
Model Tree
unsloth/DeepSeek-R1-Distill-Qwen-14B-unsloth-bnb-4bit
Input Modalities
Output Modalities
Supported Functionality
| Attribute | Value |
|---|---|
| Base Model | unsloth/DeepSeek-R1-Distill-Qwen-14B-unsloth-bnb-4bit |
| Method | QLoRA (4-bit) |
| Size | 275 MB |
| Training | 1.9 hours on RTX 5090 |
| Loss | 0.78 |
| Dataset | 4,086 LOTL Q&A examples |
python
from transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModel BASE = "unsloth/DeepSeek-R1-Distill-Qwen-14B-unsloth-bnb-4bit"ADAPTER = "art0xx/lotl-deepseek-14b" model = AutoModelForCausalLM.from_pretrained(BASE, load_in_4bit=True)model = PeftModel.from_pretrained(model, ADAPTER)tokenizer = AutoTokenizer.from_pretrained(BASE) messages = [ {"role": "system", "content": "You are a LOTL cybersecurity assistant."}, {"role": "user", "content": "How do attackers use certutil?"},]inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True)outputs = model.generate(inputs, max_new_tokens=1024)print(tokenizer.decode(outputs[0], skip_special_tokens=True))