GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
Ashura7
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
Ashura7
Model Tree
Qwen/Qwen2.5-Coder-1.5B-Instruct
Input Modalities
Output Modalities
Supported Functionality
| Metric | Baseline | Fine-tuned | Delta |
|---|---|---|---|
| Compile Rate | 0.0% | 20.0% | +20.0% |
| Pass Rate | 0.0% | 20.0% | +20.0% |
Evaluated on 100 competitive programming problems.
python
from transformers import AutoTokenizer, AutoModelForCausalLMfrom peft import PeftModelimport torch base_model_id = "Qwen/Qwen2.5-Coder-1.5B-Instruct"adapter_id = "Ashura7/cp-gpt-qwen2.5-coder-1.5b-lora" tokenizer = AutoTokenizer.from_pretrained(base_model_id)base = AutoModelForCausalLM.from_pretrained(base_model_id, torch_dtype=torch.float16, device_map="auto")model = PeftModel.from_pretrained(base, adapter_id) problem = "Given an array of integers, find the subarray with maximum sum." messages = [ {"role": "system", "content": "You are CP-GPT, an expert competitive programming assistant."}, {"role": "user", "content": f"Solve: {problem}"},] text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)inputs = tokenizer(text, return_tensors="pt").to(model.device) with torch.no_grad(): output = model.generate(**inputs, max_new_tokens=512, temperature=0.1) print(tokenizer.decode(output[0][inputs['input_ids'].shape[1]:], skip_special_tokens=True))| Param | Value |
|---|---|
| Base model | Qwen/Qwen2.5-Coder-1.5B-Instruct |
| LoRA r | 16 |
| LoRA alpha | 32 |
| Training examples | 950 |
| Epochs | 1 |
| Effective batch size | 16 |
| Learning rate | 0.0002 |
| Quantization | 4-bit (NF4) |
Part of CP-GPT — Phase 2 adds RAG (FAISS), FastAPI backend on Render, and Gradio UI on HF Spaces.
from transformers import AutoTokenizer, AutoModelForCausalLMfrom peft import PeftModelimport torch base_model_id = "Qwen/Qwen2.5-Coder-1.5B-Instruct"adapter_id = "Ashura7/cp-gpt-qwen2.5-coder-1.5b-lora" tokenizer = AutoTokenizer.from_pretrained(base_model_id)base = AutoModelForCausalLM.from_pretrained(base_model_id, torch_dtype=torch.float16, device_map="auto")model = PeftModel.from_pretrained(base, adapter_id) problem = "Given an array of integers, find the subarray with maximum sum." messages = [ {"role": "system", "content": "You are CP-GPT, an expert competitive programming assistant."}, {"role": "user", "content": f"Solve: {problem}"},] text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)inputs = tokenizer(text, return_tensors="pt").to(model.device) with torch.no_grad(): output = model.generate(**inputs, max_new_tokens=512, temperature=0.1) print(tokenizer.decode(output[0][inputs['input_ids'].shape[1]:], skip_special_tokens=True))