Update README.md

README.md

@@ -19,10 +19,37 @@ PyAutoCode is a cut-down python autosuggestion built on **GPT-2** *(motivation:
 
 You can use my model too!. Here's a quick tour of how you can achieve this:
 
+Install transformers
+```sh
+$ pip install transformers
+```
+
+Call the API and get it to work!
 ```python
 from transformers import AutoTokenizer, AutoModelForCausalLM
 
 tokenizer = AutoTokenizer.from_pretrained("P0intMaN/PyAutoCode")
 
 model = AutoModelForCausalLM.from_pretrained("P0intMaN/PyAutoCode")
+
+# input: single line or multi-line. Highly recommended to use doc-strings.
+inp = """import pandas"""
+
+format_inp = inp.replace('\n', "<N>")
+tokenize_inp = tokenizer.encode(format_inp, return_tensors='pt')
+result = model.generate(tokenize_inp)
+
+decode_result = tokenizer.decode(result[0])
+format_result = decode_result.replace('<N>', "\n")
+
+# printing the result
+print(format_result)
+```
+
+Upon successful execution, the above should probably produce *(your results may vary when this model is fine-tuned)*
+```sh
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+
 ```