File size: 1,311 Bytes
fd9a78c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 |
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)
|