Spaces:
Runtime error
Runtime error
Commit
•
cb7b577
1
Parent(s):
d37bd96
Upload config.yaml
Browse files- config.yaml +37 -0
config.yaml
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
scraping_url:
|
| 2 |
+
anorganik:
|
| 3 |
+
- "https://pestisida.id/pupuk_app/pendaftaran_anorganik.php?pendaftaran_pupuk_anorganPage="
|
| 4 |
+
organik:
|
| 5 |
+
- "https://pestisida.id/pupuk_app/pendaftaran_organik.php?pendaftaran_pupuk_organikPage="
|
| 6 |
+
|
| 7 |
+
parameters:
|
| 8 |
+
similarity_threshold: 80
|
| 9 |
+
model_types:
|
| 10 |
+
bert: "indobenchmark/indobert-base-p2"
|
| 11 |
+
roberta: "indolem/indobert-base-uncased"
|
| 12 |
+
training_args:
|
| 13 |
+
test_size: 0.3
|
| 14 |
+
random_state: 42
|
| 15 |
+
model_args:
|
| 16 |
+
num_train_epochs: 3
|
| 17 |
+
train_batch_size: 128
|
| 18 |
+
eval_batch_size: 128
|
| 19 |
+
overwrite_output_dir: True
|
| 20 |
+
fp16: True
|
| 21 |
+
do_lower_case: True
|
| 22 |
+
class_names:
|
| 23 |
+
- "Fertilizer"
|
| 24 |
+
- "Pesticide"
|
| 25 |
+
- "Others"
|
| 26 |
+
|
| 27 |
+
excluded_words:
|
| 28 |
+
- "kg"
|
| 29 |
+
- "ml"
|
| 30 |
+
- "gr"
|
| 31 |
+
- "l"
|
| 32 |
+
- "g"
|
| 33 |
+
- "gram"
|
| 34 |
+
- "grm"
|
| 35 |
+
- "ltr"
|
| 36 |
+
- "lt"
|
| 37 |
+
- "liter"
|