Spaces:
Running
Running
paper_preview_demo add an example & stream the result
Browse files- apps/paper_preview.py +26 -1
apps/paper_preview.py
CHANGED
@@ -40,6 +40,9 @@ def paper_preview_demo(client):
|
|
40 |
except Exception:
|
41 |
yield traceback.format_exc()
|
42 |
|
|
|
|
|
|
|
43 |
with gr.Row():
|
44 |
with gr.Column():
|
45 |
title = gr.Textbox(label="论文标题")
|
@@ -53,8 +56,30 @@ def paper_preview_demo(client):
|
|
53 |
|
54 |
with gr.Column():
|
55 |
outputs = gr.Textbox(label="速览内容", lines=5)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
56 |
|
57 |
submit.click(
|
58 |
preview, inputs=[title, abstract, temperature, top_p], outputs=outputs
|
59 |
)
|
60 |
-
clear.click(
|
|
|
40 |
except Exception:
|
41 |
yield traceback.format_exc()
|
42 |
|
43 |
+
def clear_data():
|
44 |
+
return None, None
|
45 |
+
|
46 |
with gr.Row():
|
47 |
with gr.Column():
|
48 |
title = gr.Textbox(label="论文标题")
|
|
|
56 |
|
57 |
with gr.Column():
|
58 |
outputs = gr.Textbox(label="速览内容", lines=5)
|
59 |
+
gr.Examples(
|
60 |
+
[
|
61 |
+
[
|
62 |
+
"GLM: General Language Model Pretraining with Autoregressive Blank Infilling",
|
63 |
+
"There have been various types of pretraining architectures including autoencoding models "
|
64 |
+
"(e.g., BERT), autoregressive models (e.g., GPT), and encoder-decoder models (e.g., T5). "
|
65 |
+
"However, none of the pretraining frameworks performs the best for all tasks of three main "
|
66 |
+
"categories including natural language understanding (NLU), unconditional generation, and "
|
67 |
+
"conditional generation. We propose a General Language Model (GLM) based on autoregressive "
|
68 |
+
"blank infilling to address this challenge. GLM improves blank filling pretraining by adding 2D"
|
69 |
+
" positional encodings and allowing an arbitrary order to predict spans, which results in "
|
70 |
+
"performance gains over BERT and T5 on NLU tasks. Meanwhile, GLM can be pretrained for "
|
71 |
+
"different types of tasks by varying the number and lengths of blanks. On a wide range of tasks"
|
72 |
+
" across NLU, conditional and unconditional generation, GLM outperforms BERT, T5, and GPT given"
|
73 |
+
" the same model sizes and data, and achieves the best performance from a single pretrained "
|
74 |
+
"model with 1.25x parameters of BERT Large , demonstrating its generalizability to different"
|
75 |
+
" downstream tasks.",
|
76 |
+
]
|
77 |
+
],
|
78 |
+
[title, abstract],
|
79 |
+
label="样例",
|
80 |
+
)
|
81 |
|
82 |
submit.click(
|
83 |
preview, inputs=[title, abstract, temperature, top_p], outputs=outputs
|
84 |
)
|
85 |
+
clear.click(clear_data, inputs=None, outputs=[title, abstract])
|