Views
No views yet
1from transformers import AutoModelForCausalLM, AutoTokenizer
2import torch
3
4model_id = "maru-miya/dpo-qwen-cot-merged_v27"
5
6tokenizer = AutoTokenizer.from_pretrained(model_id)
7model = AutoModelForCausalLM.from_pretrained(
8 model_id,
9 torch_dtype=torch.float16,
10 device_map="auto"
11)
12
13# Test inference
14prompt = "Your question here"
15inputs = tokenizer.apply_chat_template([{ "role": "user", "content": prompt }], tokenize=True, add_generation_prompt=True, return_tensors="pt").to("cuda")
16outputs = model.generate(**inputs, max_new_tokens=512)
17print(tokenizer.decode(outputs[0]))