Update README.md

README.md

@@ -1,25 +1,23 @@
----
-license: mit
-library_name: transformers
-tags: [custom, transformer, scratch, chatbot]
-language: en
----
+# Dave – Fully Custom AI Model
 
-# Dave
+**Model type:** Fully Custom Transformer from Scratch  
+**Purpose:** Conversational AI with a unique personality: fun, smart, coding-savvy, sometimes sad or depressed, talks about drugs and alcohol, lost his wife, happy about his kid, sometimes cringe, close friends, swears often.
 
-**A fully custom AI trained from scratch.**
+**Training:**  
+- 3000 generated prompts and responses  
+- Custom tokenizer and transformer trained from scratch  
+- Trained entirely in Google Colab  
+- Saved as `pytorch_model.bin`, `config.json`, `tokenizer.json`
 
-Dave is a 100% custom-built transformer AI, trained entirely from scratch without pretrained models.
-It has 3,000 generated prompts and responses.
+**Example usage:**
 
-## Usage
 ```python
-from transformers import PreTrainedTokenizerFast
-import torch
-from your_model_file import SimpleTransformer
+from transformers import AutoTokenizer, AutoModelForCausalLM
 
-tokenizer = PreTrainedTokenizerFast(tokenizer_file='./dave/tokenizer.json')
-model = SimpleTransformer(vocab_size=5000, d_model=128, nhead=4, num_layers=4)
-model.load_state_dict(torch.load('./dave/pytorch_model.bin'))
-model.eval()
-```
+tokenizer = AutoTokenizer.from_pretrained("DSDUDEd/Dave")
+model = AutoModelForCausalLM.from_pretrained("DSDUDEd/Dave")
+
+prompt = "Write a Python function to reverse a string."
+inputs = tokenizer(prompt, return_tensors="pt")
+outputs = model.generate(**inputs, max_new_tokens=64)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))