Update README.md

README.md

@@ -4,6 +4,21 @@ license: mit
 
 # PyAutoCode: GPT-2 based Python auto-code.
 
-PyAutoCode is a cut-down python autosuggestion built on **GPT-2** (motivation: GPyT) model. This baby model is not **"fine-tuned"** yet therefore, I highly recommend not to use it in a production environment or incorporate PyAutoCode in any of your projects. It has been trained on **112GB** of Python data sourced from the best crowdsource platform ever -- **GitHub**.
+PyAutoCode is a cut-down python autosuggestion built on **GPT-2** *(motivation: GPyT)* model. This baby model *(trained only up to 3 epochs)* is not **"fine-tuned"** yet therefore, I highly recommend not to use it in a production environment or incorporate PyAutoCode in any of your projects. It has been trained on **112GB** of Python data sourced from the best crowdsource platform ever -- **GitHub**.
 
+*NOTE: Increased training and fine tuning would be highly appreciated and I firmly believe that it would improve the ability of PyAutoCode significantly.*
 
+## Some Model Features
+
+- Built on *GPT-2*
+- Tokenized with *ByteLevelBPETokenizer*
+- Data Sourced from *GitHub (almost 5 consecutive days of latest Python repositories)*
+- Makes use of *GPTLMHeadModel* and *DataCollatorForLanguageModelling* for training
+
+## Usage
+
+You can use my model too!. Here's a quick tour of how you can achieve this:
+
+```python
+
+```