Spaces:
Running
Running
jeanpoll commited on
Commit Β·
61e5adb
1
Parent(s): c564c57
update with new model
Browse files- email_parser/_models_signatures.py +8 -6
- email_parser/config.ini +1 -1
- email_parser/models/model_signature_lstm_v10/variables/variables.data-00000-of-00001 +0 -0
- email_parser/models/model_signature_lstm_v10/variables/variables.index +0 -0
- email_parser/models/{model_signature_lstm_v10 β model_signature_lstm_v16}/keras_metadata.pb +2 -2
- email_parser/models/{model_signature_lstm_v10 β model_signature_lstm_v16}/minmax_scaler.p +0 -0
- email_parser/models/{model_signature_lstm_v10 β model_signature_lstm_v16}/saved_model.pb +2 -2
- email_parser/models/{model_signature_lstm_v10 β model_signature_lstm_v16}/standard_scaler.p +0 -0
- email_parser/models/model_signature_lstm_v16/variables/variables.data-00000-of-00001 +0 -0
- email_parser/models/model_signature_lstm_v16/variables/variables.index +0 -0
email_parser/_models_signatures.py
CHANGED
|
@@ -81,15 +81,17 @@ def f_retrieve_entities_for_line(df_ner, start=0, end=1e12):
|
|
| 81 |
df = df_ner.query(f"""(start>= {start} and end <= {end}) or (start<={start} and end>={end})""")
|
| 82 |
return df
|
| 83 |
|
|
|
|
| 84 |
embedder_model = SentenceTransformer("distiluse-base-multilingual-cased-v1")
|
| 85 |
|
|
|
|
| 86 |
def f_create_embedding_inv_dist_feature(text1, text2):
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
| 90 |
-
|
| 91 |
-
|
| 92 |
-
|
| 93 |
|
| 94 |
|
| 95 |
def f_create_email_lines_features(text, df_ner=None, position_offset=0):
|
|
|
|
| 81 |
df = df_ner.query(f"""(start>= {start} and end <= {end}) or (start<={start} and end>={end})""")
|
| 82 |
return df
|
| 83 |
|
| 84 |
+
|
| 85 |
embedder_model = SentenceTransformer("distiluse-base-multilingual-cased-v1")
|
| 86 |
|
| 87 |
+
|
| 88 |
def f_create_embedding_inv_dist_feature(text1, text2):
|
| 89 |
+
""" Computing distance between two texts based on their embedding
|
| 90 |
+
provided by the SentenceTransformer above"""
|
| 91 |
+
embedding_merci = embedder_model.encode(text1)
|
| 92 |
+
embedding_line = embedder_model.encode(text2)
|
| 93 |
+
dist = distance.cosine(embedding_merci, embedding_line)
|
| 94 |
+
return min(5, 1 / (dist + 0.0001))
|
| 95 |
|
| 96 |
|
| 97 |
def f_create_email_lines_features(text, df_ner=None, position_offset=0):
|
email_parser/config.ini
CHANGED
|
@@ -3,5 +3,5 @@ ner_model_fr = Jean-Baptiste/camembert-ner-with-dates
|
|
| 3 |
ner_model_en = Jean-Baptiste/roberta-large-ner-english
|
| 4 |
device = -1
|
| 5 |
default_lang = en
|
| 6 |
-
name_model_signature =
|
| 7 |
path_models = models
|
|
|
|
| 3 |
ner_model_en = Jean-Baptiste/roberta-large-ner-english
|
| 4 |
device = -1
|
| 5 |
default_lang = en
|
| 6 |
+
name_model_signature = model_signature_lstm_v16
|
| 7 |
path_models = models
|
email_parser/models/model_signature_lstm_v10/variables/variables.data-00000-of-00001
DELETED
|
Binary file (116 kB)
|
|
|
email_parser/models/model_signature_lstm_v10/variables/variables.index
DELETED
|
Binary file (3.48 kB)
|
|
|
email_parser/models/{model_signature_lstm_v10 β model_signature_lstm_v16}/keras_metadata.pb
RENAMED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:73a9ab3d0262141747b5aafa348c57e1a5813e03cf1e5c291162b37cdf846143
|
| 3 |
+
size 22142
|
email_parser/models/{model_signature_lstm_v10 β model_signature_lstm_v16}/minmax_scaler.p
RENAMED
|
File without changes
|
email_parser/models/{model_signature_lstm_v10 β model_signature_lstm_v16}/saved_model.pb
RENAMED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:70ff3069a27b192313bbdd9685f106fc46ef2082eab6479697f575bad555eb4f
|
| 3 |
+
size 5296390
|
email_parser/models/{model_signature_lstm_v10 β model_signature_lstm_v16}/standard_scaler.p
RENAMED
|
File without changes
|
email_parser/models/model_signature_lstm_v16/variables/variables.data-00000-of-00001
ADDED
|
Binary file (117 kB). View file
|
|
|
email_parser/models/model_signature_lstm_v16/variables/variables.index
ADDED
|
Binary file (3.51 kB). View file
|
|
|