f

app.py

@@ -1,32 +1,29 @@
-import gradio as gr
-from transformers import AutoTokenizer, AutoConfig, AutoModelForSequenceClassification
+from transformers import AutoModelForCausalLM, AutoTokenizer
 
-# Load model and tokenizer
-model_name = "cheberle/autotrain-35swc-b4r9z"
-tokenizer = AutoTokenizer.from_pretrained(model_name)
+model_path = "cheberle/autotrain-35swc-b4r9z"
 
-# Explicitly define the model configuration if needed
-config = AutoConfig.from_pretrained(model_name)
-model = AutoModelForSequenceClassification.from_pretrained(model_name, config=config)
+# Load the tokenizer and model
+tokenizer = AutoTokenizer.from_pretrained(model_path)
+model = AutoModelForCausalLM.from_pretrained(
+    model_path,
+    device_map="auto",  # Auto-distributes model across available devices
+    torch_dtype="auto"  # Automatically selects the appropriate data type
+).eval()
 
-# Inference function
-def classify_text(input_text):
-    inputs = tokenizer(input_text, return_tensors="pt", padding=True, truncation=True)
-    outputs = model(**inputs)
-    probabilities = outputs.logits.softmax(dim=-1).tolist()[0]
-    labels = {i: f"Label {i}" for i in range(len(probabilities))}  # Define label mapping if needed
-    result = {labels[i]: prob for i, prob in enumerate(probabilities)}
-    return result
+# Prompt content
+messages = [{"role": "user", "content": "hi"}]
 
-# Gradio interface
-interface = gr.Interface(
-    fn=classify_text,
-    inputs="text",
-    outputs="label",
-    title="DeepSeek-R1 Text Classification",
-    description="Classify text inputs using the DeepSeek-R1 model."
+# Prepare input for the model
+input_ids = tokenizer.apply_chat_template(
+    conversation=messages, 
+    tokenize=True, 
+    add_generation_prompt=True, 
+    return_tensors='pt'
 )
 
-# Launch the app
-if __name__ == "__main__":
-    interface.launch()
+# Generate response
+output_ids = model.generate(input_ids.to('cuda'))  # Ensure the model uses the GPU if available
+response = tokenizer.decode(output_ids[0][input_ids.shape[1]:], skip_special_tokens=True)
+
+# Print response
+print(response)

requirements.txt

@@ -1,4 +1,4 @@
 huggingface_hub==0.25.2
-gradio
+transformers
 torch
-transformers
+accelerate