Update app.py

app.py

@@ -1,3 +1,4 @@
+import gradio as gr
 import torch
 import numpy as np
 from model import Transformer
@@ -33,21 +34,20 @@ m = model.to(DEVICE)
 m = torch.load("base_mode.pth", map_location=torch.device(DEVICE))
 m.eval()
 
-print(
-    "Model with {:.2f}M parameters".format(sum(p.numel() for p in m.parameters()) / 1e6)
-)
+#print(
+#    "Model with {:.2f}M parameters".format(sum(p.numel() for p in m.parameters()) / 1e6)
+#)
+def model_generate(text):
+    # generate some output based on the context
+    #context = torch.tensor(np.array(encode("Hello! My name is ", tokenizer)))
+    #context = torch.zeros((1, 1), dtype=torch.long, device=DEVICE)
+    text_input = str(input())
+    context_np = np.array(encode(text_input, tokenizer))
+    context_np = np.array([context_np])
+    context = torch.from_numpy(context_np)
+    #print(context)
+
+    return decode(enc_sec=m.generate(idx=context, max_new_tokens=100, block_size=BLOCK_SIZE)[0], tokenizer=tokenizer)
 
-# generate some output based on the context
-#context = torch.tensor(np.array(encode("Hello! My name is ", tokenizer)))
-#context = torch.zeros((1, 1), dtype=torch.long, device=DEVICE)
-text_input = str(input())
-context_np = np.array(encode(text_input, tokenizer))
-context_np = np.array([context_np])
-context = torch.from_numpy(context_np)
-print(context)
-print(
-    decode(
-        enc_sec=m.generate(idx=context, max_new_tokens=100, block_size=BLOCK_SIZE)[0],
-        tokenizer=tokenizer,
-    )
-)
+iface = gr.Interface(fn=model_generate, inputs="text", outputs="text")
+iface.launch()