Update app.py
Browse files
app.py
CHANGED
@@ -44,14 +44,24 @@ def proc_submission(
|
|
44 |
str in HTML format, string of the summary, str of compression rate in %
|
45 |
"""
|
46 |
|
47 |
-
|
48 |
-
"length_penalty":
|
49 |
"repetition_penalty": 3.5,
|
50 |
"no_repeat_ngram_size": 3,
|
51 |
"encoder_no_repeat_ngram_size": 4,
|
52 |
"num_beams": int(num_beams),
|
53 |
"min_length": 11,
|
54 |
-
"max_length":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
55 |
"early_stopping": True,
|
56 |
}
|
57 |
st = time.perf_counter()
|
@@ -75,7 +85,7 @@ def proc_submission(
|
|
75 |
model_led_tldr if (model_type == "LED") else model_tldr,
|
76 |
tokenizer_led_tldr if (model_type == "LED") else tokenizer_tldr,
|
77 |
batch_length=token_batch_length,
|
78 |
-
**
|
79 |
)
|
80 |
|
81 |
else:
|
@@ -84,7 +94,7 @@ def proc_submission(
|
|
84 |
model_led_det if (model_type == "LED") else model_det,
|
85 |
tokenizer_led_det if (model_type == "LED") else tokenizer_det,
|
86 |
batch_length=token_batch_length,
|
87 |
-
**
|
88 |
)
|
89 |
#_summaries = summarize_via_tokenbatches(
|
90 |
#tr_in,
|
|
|
44 |
str in HTML format, string of the summary, str of compression rate in %
|
45 |
"""
|
46 |
|
47 |
+
settings_tldr = {
|
48 |
+
"length_penalty": 0.6,
|
49 |
"repetition_penalty": 3.5,
|
50 |
"no_repeat_ngram_size": 3,
|
51 |
"encoder_no_repeat_ngram_size": 4,
|
52 |
"num_beams": int(num_beams),
|
53 |
"min_length": 11,
|
54 |
+
"max_length": 62,
|
55 |
+
"early_stopping": True,
|
56 |
+
}
|
57 |
+
settings_det = {
|
58 |
+
"length_penalty": float(length_penalty),#2.0 if (model_type == "LED") else 0.8,
|
59 |
+
"repetition_penalty": 3.5,
|
60 |
+
"no_repeat_ngram_size": 3,
|
61 |
+
"encoder_no_repeat_ngram_size": 4,
|
62 |
+
"num_beams": int(num_beams),
|
63 |
+
"min_length": 100,
|
64 |
+
"max_length": int(token_batch_length // 4) if (token_batch_length <500) else 512,
|
65 |
"early_stopping": True,
|
66 |
}
|
67 |
st = time.perf_counter()
|
|
|
85 |
model_led_tldr if (model_type == "LED") else model_tldr,
|
86 |
tokenizer_led_tldr if (model_type == "LED") else tokenizer_tldr,
|
87 |
batch_length=token_batch_length,
|
88 |
+
**settings_tldr,
|
89 |
)
|
90 |
|
91 |
else:
|
|
|
94 |
model_led_det if (model_type == "LED") else model_det,
|
95 |
tokenizer_led_det if (model_type == "LED") else tokenizer_det,
|
96 |
batch_length=token_batch_length,
|
97 |
+
**settings_det,
|
98 |
)
|
99 |
#_summaries = summarize_via_tokenbatches(
|
100 |
#tr_in,
|