replace app

main.py

@@ -1,30 +1,26 @@
+from transformers import pipeline
 import gradio as gr
 import torch
-import requests
-from torchvision import transforms
+import logging
 
-model = torch.hub.load("pytorch/vision:v0.6.0", "resnet18", pretrained=True).eval()
-response = requests.get("https://git.io/JJkYN")
-labels = response.text.split("\n")
+logging.basicConfig(level=logging.INFO)
+logging.info(f"CUDA available: {torch.cuda.is_available()}, CUDA version: {torch.version.cuda}")
 
+def decomp_create_prompt(input_data: str) -> str:
+    before = "# This is the decompiled code:\n"
+    after = "\n# What is the source code?\n"
+    prompt = before + input_data.strip() + after
 
-def predict(inp):
-    inp = transforms.ToTensor()(inp).unsqueeze(0)
-    with torch.no_grad():
-        prediction = torch.nn.functional.softmax(model(inp)[0], dim=0)
-        confidences = {labels[i]: float(prediction[i]) for i in range(1000)}
-    return confidences
+    return prompt
 
 
-def run():
-    demo = gr.Interface(
-        fn=predict,
-        inputs=gr.Image(type="pil"),
-        outputs=gr.Label(num_top_classes=3),
-    )
+pipe = pipeline(model="ejschwartz/decaf-v1-22b-4bit", return_full_text=False)
+pipe.model.to("cuda")
 
-    demo.launch(server_name="0.0.0.0", server_port=7860)
+def generate(text):
+    print(f"Generating text... {text}")
+    prompt = decomp_create_prompt(text)
+    return pipe(prompt, max_new_tokens=2000)[0]['generated_text']
 
-
-if __name__ == "__main__":
-    run()
+demo = gr.Interface(fn=generate, inputs="text", outputs="text")
+demo.launch()