Update README.md

README.md

@@ -0,0 +1,60 @@
+**EXAMPLE USAGE**
+
+```
+# Install required packages if needed
+# !pip install transformers torch unsloth
+
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from unsloth.chat_templates import get_chat_template
+from unsloth import FastLanguageModel
+import torch
+
+# Load the electrical engineering model
+model_name = "neuralnets/electrical_engg_model"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(model_name)
+
+# Apply the chat template to format inputs correctly
+tokenizer = get_chat_template(
+    tokenizer,
+    chat_template = "llama-3.1",
+)
+
+# Enable faster inference using Unsloth
+model = FastLanguageModel.for_inference(model)
+
+# Move model to GPU if available (or specify your device)
+device = "cuda" if torch.cuda.is_available() else "cpu"
+model = model.to(device)
+
+# Create an electrical engineering related query
+messages = [
+    {"role": "user", "content": "Explain the working principle of a three-phase induction motor."},
+]
+
+# Format the input using the chat template
+inputs = tokenizer.apply_chat_template(
+    messages,
+    tokenize = True,
+    add_generation_prompt = True,  # Required for generation
+    return_tensors = "pt",
+).to(device)
+
+# Set up text streaming for real-time output
+from transformers import TextStreamer
+text_streamer = TextStreamer(tokenizer, skip_prompt = True)
+
+# Generate response
+outputs = model.generate(
+    input_ids = inputs, 
+    streamer = text_streamer,
+    max_new_tokens = 512,
+    use_cache = True, 
+    temperature = 0.7,  # Adjust temperature for creativity vs precision
+    min_p = 0.05       # Nucleus sampling parameter
+)
+
+# If you want to capture the full response as a string
+full_response = tokenizer.decode(outputs[0], skip_special_tokens=True)
+
+```