Update app.py

app.py

@@ -1,16 +1,28 @@
 import gradio as gr
 from optimum.intel import OVDiffusionPipeline
+from transformers import AutoTokenizer
 from threading import Lock
+import warnings
 
-# Load the pipeline globally (includes the correct tokenizer)
+# Suppress deprecation warnings (optional)
+warnings.filterwarnings("ignore", category=DeprecationWarning)
+
+# Load the pipeline globally
 model_id = "OpenVINO/FLUX.1-schnell-int4-ov"
 pipeline = OVDiffusionPipeline.from_pretrained(model_id, device="CPU")
+
+# Explicitly load and configure the tokenizer
+# FLUX.1-schnell uses a T5-based encoder, typically t5-v1_1-xl or similar
+tokenizer = AutoTokenizer.from_pretrained("google/t5-v1_1-xl", use_fast=True, add_prefix_space=True)
+pipeline.text_encoder_2.tokenizer = tokenizer  # Assign to the T5 encoder
+
 lock = Lock()
 
 # Define the image generation function
 def generate_image(prompt):
     with lock:
-        image = pipeline(prompt, num_inference_steps=4, guidance_scale=3.5).images[0]
+        # Reduce num_inference_steps for faster inference to avoid timeouts
+        image = pipeline(prompt, num_inference_steps=2, guidance_scale=3.5).images[0]
     return image
 
 # Create the Gradio interface
@@ -26,4 +38,4 @@ interface = gr.Interface(
 
 # Launch the interface
 if __name__ == "__main__":
-    interface.launch()
+    interface.launch(server_name="0.0.0.0", server_port=7860)  # Explicitly set for Spaces