GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
alperiox
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
alperiox
Model Tree
Qwen/Qwen2.5-1.5B-Instruct
Input Modalities
Output Modalities
Supported Functionality
| Model | Parse | Fonksiyon Doğruluğu | Gerekli Argümanlar | Şema Uyumu |
|---|---|---|---|---|
| Qwen2.5 1.5B (base) | 92% | 54% | 52% | 44% |
| Qwen2.5 1.5B (fine-tuned) | 100% | 88% | 88% | 84% |
Fine-tuning ile fonksiyon doğruluğu %54'ten %88'e yükselmiştir (+34 puan).
read_text_file, read_multiple_files, write_file, edit_file, create_directory, list_directory, directory_tree, move_file, search_files, get_file_info
issue_read, issue_write, list_issues, search_issues, add_issue_comment, pull_request_read, create_pull_request, update_pull_request, merge_pull_request, list_pull_requests, create_branch, get_file_contents, create_or_update_file, list_commits, list_branches, search_code, get_repository_tree, actions_list,
python
from unsloth import FastModelfrom unsloth.chat_templates import get_chat_templatefrom peft import PeftModelbase_model, tokenizer = FastModel.from_pretrained("unsloth/Qwen2.5-1.5B-Instruct",max_seq_length=4096,load_in_4bit=True,)tokenizer = get_chat_template(tokenizer, chat_template="qwen-2.5")model = PeftModel.from_pretrained(base_model, "alperiox/qwen25-1.5b-turkish-fc")messages = [{"role": "system", "content": SYSTEM_PROMPT},{"role": "user", "content": "main branch'ten feature/login adında yeni bir branch oluştur"},]# ... generate response
Bu model, Python Ankara topluluğunun "Küçük Dil Modellerini Türkçe İçin Fine-Tuning" atölyesi kapsamında geliştirilmiştir.
Eğitim verisi: alperiox/turkish-devtools-function-calling
get_job_logs