from unsloth import FastLanguageModel
import torch

model_name = "Ozin11/deepse-phi-3-mini-time-prediction"
model, tokenizer = FastLanguageModel.from_pretrained(
    model_name = model_name,
    max_seq_length = 2048,
    load_in_4bit = True,
)
FastLanguageModel.for_inference(model)

test_task = 'Fix authentication bug in login module'
prompt = f'### Instruction:\nPrediksi durasi untuk tugas berikut.\n\n### Input:\n{test_task}\n\n### Response:\n'

inputs = tokenizer([prompt], return_tensors = 'pt').to('cuda')
outputs = model.generate(**inputs, max_new_tokens = 32)
response = tokenizer.batch_decode(outputs)[0]
print(response.split('### Response:')[-1].strip())

Limitations

Estimates are based on historical patterns and should be used as a reference only.