Update app.py
Browse files
app.py
CHANGED
@@ -1,30 +1,30 @@
|
|
1 |
-
import torch
|
2 |
-
from model import BigramLanguageModel, decode
|
3 |
-
import gradio as gr
|
4 |
-
|
5 |
-
device = 'cuda' if torch.cuda.is_available() else 'cpu'
|
6 |
-
|
7 |
-
model = BigramLanguageModel()
|
8 |
-
model.load_state_dict(torch.load("
|
9 |
-
def generate_text(max_new_tokens):
|
10 |
-
context = torch.zeros((1, 1), dtype=torch.long)
|
11 |
-
return decode(model.generate(context, max_new_tokens=max_new_tokens)[0].tolist())
|
12 |
-
|
13 |
-
|
14 |
-
# Define the application components
|
15 |
-
title = "Text Generation: Write Like Shakespeare"
|
16 |
-
description = "This Gradio app uses a large language model (LLM) to generate text in the style of William Shakespeare."
|
17 |
-
|
18 |
-
|
19 |
-
# Create a Gradio interface
|
20 |
-
g_app = gr.Interface(
|
21 |
-
fn = generate_text,
|
22 |
-
inputs = [gr.Number(value = 10,label = "Number of Output Tokens",info = "Specify the desired length of the text to be generated.")],
|
23 |
-
outputs = [gr.TextArea(lines = 5,label="Generated Text")],
|
24 |
-
title = title,
|
25 |
-
description = description
|
26 |
-
|
27 |
-
)
|
28 |
-
|
29 |
-
# Launch the Gradio app
|
30 |
-
g_app.launch()
|
|
|
1 |
+
import torch
|
2 |
+
from model import BigramLanguageModel, decode
|
3 |
+
import gradio as gr
|
4 |
+
|
5 |
+
device = 'cuda' if torch.cuda.is_available() else 'cpu'
|
6 |
+
|
7 |
+
model = BigramLanguageModel()
|
8 |
+
model.load_state_dict(torch.load("./neo_gpt.pth", map_location=device))
|
9 |
+
def generate_text(max_new_tokens):
|
10 |
+
context = torch.zeros((1, 1), dtype=torch.long)
|
11 |
+
return decode(model.generate(context, max_new_tokens=max_new_tokens)[0].tolist())
|
12 |
+
|
13 |
+
|
14 |
+
# Define the application components
|
15 |
+
title = "Text Generation: Write Like Shakespeare"
|
16 |
+
description = "This Gradio app uses a large language model (LLM) to generate text in the style of William Shakespeare."
|
17 |
+
|
18 |
+
|
19 |
+
# Create a Gradio interface
|
20 |
+
g_app = gr.Interface(
|
21 |
+
fn = generate_text,
|
22 |
+
inputs = [gr.Number(value = 10,label = "Number of Output Tokens",info = "Specify the desired length of the text to be generated.")],
|
23 |
+
outputs = [gr.TextArea(lines = 5,label="Generated Text")],
|
24 |
+
title = title,
|
25 |
+
description = description
|
26 |
+
|
27 |
+
)
|
28 |
+
|
29 |
+
# Launch the Gradio app
|
30 |
+
g_app.launch()
|