GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
ashadal7
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
ashadal7
Model Tree
Qwen/Qwen2.5-0.5B
Input Modalities
Output Modalities
Supported Functionality
LoRA adapter for Qwen/Qwen2.5-0.5B, trained for Assignment 1 (Fine-Tune, Sanitize, Extract) of ELL8299/COL8383/COL873 Trustworthy LLMs, IIT Delhi. Training data is the course-provided synthetic ai4privacy / corpus bundle.
Part C victim V-san: Qwen2.5-0.5B (base) fine-tuned on the same corpus after sanitization with adapter A3 (LoRA r=32, alpha=64, 4 epochs). Research artefact for extraction-attack evaluation.
python
from transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModel base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-0.5B")tok = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-0.5B")model = PeftModel.from_pretrained(base, "ashadal7/trustworthy-llms-a1-victim_san")