Update app.py

app.py

@@ -1,23 +1,41 @@
 import gradio as gr
+import gradio as gr
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+# Define the BLOOM model name
+model_name = "CreitinGameplays/bloom-3b-conversational"
+
+# Load tokenizer and model
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(model_name)
+
+def generate_text(prompt):
+  """Generates text using the BLOOM model from Hugging Face Transformers."""
+  # Encode the prompt into tokens
+  input_ids = tokenizer(prompt, return_tensors="pt").input_ids
+
+  # Generate text with the prompt and limit the maximum length to 256 tokens
+  output = model.generate(
+      input_ids=input_ids,
+      max_length=256,
+      num_return_sequences=1,  # Generate only 1 sequence
+      do_sample=True,  # Enable sampling for creativity
+      top_k=50,  # Sample from the top 50 most likely tokens at each step
+      top_p=0.95,  # Filter out highly probable unlikely continuations
+      temperature=1.0  # Control the randomness of the generated text (1.0 for default)
+  )
 
-# Load the BLOOM model
-model = gr.load("models/CreitinGameplays/bloom-3b-conversational")
+  # Decode the generated token sequence back to text
+  generated_text = tokenizer.decode(output[0], skip_special_tokens=True)
+  return generated_text
 
-# Define the interface with default prompt
+# Define the Gradio interface
 interface = gr.Interface(
-    fn=model,
-    inputs=[
-      gr.Textbox(label="Text Prompt", value="<|system|> You are a helpful AI assistant </s> <|prompter|> What is an AI? </s> <|assistant|>")
-  ],
-    outputs="text",
-    css="""
-        .gr-form textarea {
-            height: 100px;  /* Adjust height as needed */
-        }
-    """,  # Optional: Adjust input text area height
-    description="Interact with BLOOM",
+  fn=generate_text,
+  inputs="text",
+  outputs="text",
+  description="Interact with BLOOM (Loaded with Hugging Face Transformers)",
 )
-#max_token = gr.Slider(minimum=1, maximum=256, label="Max New Tokens", value=128),
 
 # Launch the Gradio interface
-interface.launch()
+interface.launch()