testmodel.py

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_path = "/workspace/distillation/gpt-oss-distilled/results/checkpoint-171"
model = AutoModelForCausalLM.from_pretrained(
    model_path,
    torch_dtype=torch.bfloat16,
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_path)

# Proper chat format
messages = [
    {"role": "user", "content": "Explain quantum computing in simple terms"}
]

# Use chat template if available
if hasattr(tokenizer, 'apply_chat_template'):
    text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
else:
    text = "User: Explain quantum computing in simple terms\nAssistant:"

inputs = tokenizer(text, return_tensors="pt").to(model.device)

# Better generation settings
with torch.no_grad():
    outputs = model.generate(
        **inputs,
        max_new_tokens=150,
        temperature=0.7,
        top_p=0.9,
        do_sample=True,
        repetition_penalty=1.1,  # Prevent repetition
        pad_token_id=tokenizer.eos_token_id,
        eos_token_id=tokenizer.eos_token_id,
        num_beams=1
        # no_repeat_ngram_size=3  # Prevent 3-gram repetition
    )

response = tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True)
print("Response:", response)