GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
metinyurdev
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
metinyurdev
Model Tree
Qwen/Qwen2.5-14B-Instruct
Input Modalities
Output Modalities
Supported Functionality
⚠️ IMPORTANT COMPLIANCE NOTICE FOR EVALUATORS
This repository is strictly gated to protect institutional data privacy and comply with graduation project guidelines.
- Academic Evaluators (Thesis Committee / Professors): Please click the "Access Repository" button and submit a request specifying your academic title and affiliation. Permissions are reviewed and granted promptly.
- Industry Reviewers (AI Team Leads / Technical Recruiters): Access is granted exclusively for professional engineering evaluations. Please state your organization, company, or technical team when submitting your request.
- Frictionless Code Security: To comply with institutional framework rules, all underlying Python training blueprints, hyperparameter scheduling scripts, and proprietary dataset tokenization pipelines are kept strictly confidential. Only the final compiled model weights and configurations are accessible upon manual administrator approval.
The model's cognitive optimization was achieved using Parameter-Efficient Fine-Tuning (PEFT) via Low-Rank Adaptation (LoRA) abstractions, powered by the Unsloth framework on an NVIDIA RTX PRO 6000 Blackwell Server Edition.
The clinical domain adaptation followed a strict 3-Stage Fine-Tuning Pipeline:
#1800).Once your access request is Approved, you can load the base model and dynamically overlay these clinical adapters using standard transformers and peft syntax:
python
import torchfrom transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModelfrom huggingface_hub import notebook_login # Authenticate with your Read Tokennotebook_login() base_model_name = "Qwen/Qwen2.5-14B-Instruct"adapter_model_name = "metinyurdev/clinical-asistant-qwen2.5-14b-lora" tokenizer = AutoTokenizer.from_pretrained(base_model_name)model = AutoModelForCausalLM.from_pretrained( base_model_name, torch_dtype=torch.bfloat16, device_map="auto") # Overlay the gated clinical LoRA weightsmodel = PeftModel.from_pretrained(model, adapter_model_name)print("🎉 Clinical Stage 3 Adapters successfully merged in-memory!")For direct institutional validation, deep-technical inquiries, or professional evaluation regarding this graduation thesis framework, please initiate an official request via the manual access panel above.
import torchfrom transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModelfrom huggingface_hub import notebook_login # Authenticate with your Read Tokennotebook_login() base_model_name = "Qwen/Qwen2.5-14B-Instruct"adapter_model_name = "metinyurdev/clinical-asistant-qwen2.5-14b-lora" tokenizer = AutoTokenizer.from_pretrained(base_model_name)model = AutoModelForCausalLM.from_pretrained( base_model_name, torch_dtype=torch.bfloat16, device_map="auto") # Overlay the gated clinical LoRA weightsmodel = PeftModel.from_pretrained(model, adapter_model_name)print("🎉 Clinical Stage 3 Adapters successfully merged in-memory!")