Commit
·
88ce039
1
Parent(s):
48a4d29
Upload BertForSequenceClassification
Browse files- config.json +6 -1
- modeling_bert.py +6 -11
- pytorch_model.bin +1 -1
config.json
CHANGED
@@ -1,7 +1,11 @@
|
|
1 |
{
|
|
|
|
|
|
|
2 |
"attention_probs_dropout_prob": 0.1,
|
3 |
"auto_map": {
|
4 |
-
"AutoConfig": "configuration_bert.BertConfig"
|
|
|
5 |
},
|
6 |
"classifier_dropout": null,
|
7 |
"hidden_act": "gelu",
|
@@ -16,6 +20,7 @@
|
|
16 |
"num_hidden_layers": 12,
|
17 |
"pad_token_id": 0,
|
18 |
"position_embedding_type": "absolute",
|
|
|
19 |
"transformers_version": "4.33.3",
|
20 |
"type_vocab_size": 2,
|
21 |
"use_cache": true,
|
|
|
1 |
{
|
2 |
+
"architectures": [
|
3 |
+
"BertForSequenceClassification"
|
4 |
+
],
|
5 |
"attention_probs_dropout_prob": 0.1,
|
6 |
"auto_map": {
|
7 |
+
"AutoConfig": "configuration_bert.BertConfig",
|
8 |
+
"AutoModelForSequenceClassification": "modeling_bert.BertForSequenceClassification"
|
9 |
},
|
10 |
"classifier_dropout": null,
|
11 |
"hidden_act": "gelu",
|
|
|
20 |
"num_hidden_layers": 12,
|
21 |
"pad_token_id": 0,
|
22 |
"position_embedding_type": "absolute",
|
23 |
+
"torch_dtype": "float32",
|
24 |
"transformers_version": "4.33.3",
|
25 |
"type_vocab_size": 2,
|
26 |
"use_cache": true,
|
modeling_bert.py
CHANGED
@@ -19,16 +19,12 @@ from transformers.modeling_outputs import (
|
|
19 |
SequenceClassifierOutput
|
20 |
)
|
21 |
|
22 |
-
from .configuration_bert import
|
23 |
|
24 |
|
25 |
class BertPreTrainedModel(PreTrainedModel):
|
26 |
-
"""
|
27 |
-
An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
|
28 |
-
models.
|
29 |
-
"""
|
30 |
|
31 |
-
config_class =
|
32 |
load_tf_weights = load_tf_weights_in_bert
|
33 |
base_model_prefix = "bert"
|
34 |
supports_gradient_checkpointing = True
|
@@ -50,7 +46,7 @@ class BertPreTrainedModel(PreTrainedModel):
|
|
50 |
module.weight.data.fill_(1.0)
|
51 |
|
52 |
|
53 |
-
class
|
54 |
|
55 |
def __init__(self, config):
|
56 |
super().__init__()
|
@@ -68,7 +64,7 @@ class BertClsPooler(nn.Module):
|
|
68 |
|
69 |
class BertModel(BertPreTrainedModel):
|
70 |
|
71 |
-
config_class =
|
72 |
|
73 |
def __init__(self, config, add_pooling_layer=True):
|
74 |
super().__init__(config)
|
@@ -77,7 +73,7 @@ class BertModel(BertPreTrainedModel):
|
|
77 |
self.embeddings = BertEmbeddings(config)
|
78 |
self.encoder = BertEncoder(config)
|
79 |
|
80 |
-
self.pooler =
|
81 |
|
82 |
# Initialize weights and apply final processing
|
83 |
self.post_init()
|
@@ -201,7 +197,7 @@ class BertModel(BertPreTrainedModel):
|
|
201 |
|
202 |
class BertForSequenceClassification(BertPreTrainedModel):
|
203 |
|
204 |
-
config_class =
|
205 |
|
206 |
def __init__(self, config):
|
207 |
super().__init__(config)
|
@@ -290,4 +286,3 @@ class BertForSequenceClassification(BertPreTrainedModel):
|
|
290 |
hidden_states=outputs.hidden_states,
|
291 |
attentions=outputs.attentions,
|
292 |
)
|
293 |
-
|
|
|
19 |
SequenceClassifierOutput
|
20 |
)
|
21 |
|
22 |
+
from .configuration_bert import BertConfig
|
23 |
|
24 |
|
25 |
class BertPreTrainedModel(PreTrainedModel):
|
|
|
|
|
|
|
|
|
26 |
|
27 |
+
config_class = BertConfig
|
28 |
load_tf_weights = load_tf_weights_in_bert
|
29 |
base_model_prefix = "bert"
|
30 |
supports_gradient_checkpointing = True
|
|
|
46 |
module.weight.data.fill_(1.0)
|
47 |
|
48 |
|
49 |
+
class BertPooler(nn.Module):
|
50 |
|
51 |
def __init__(self, config):
|
52 |
super().__init__()
|
|
|
64 |
|
65 |
class BertModel(BertPreTrainedModel):
|
66 |
|
67 |
+
config_class = BertConfig
|
68 |
|
69 |
def __init__(self, config, add_pooling_layer=True):
|
70 |
super().__init__(config)
|
|
|
73 |
self.embeddings = BertEmbeddings(config)
|
74 |
self.encoder = BertEncoder(config)
|
75 |
|
76 |
+
self.pooler = BertPooler(config) if add_pooling_layer else None
|
77 |
|
78 |
# Initialize weights and apply final processing
|
79 |
self.post_init()
|
|
|
197 |
|
198 |
class BertForSequenceClassification(BertPreTrainedModel):
|
199 |
|
200 |
+
config_class = BertConfig
|
201 |
|
202 |
def __init__(self, config):
|
203 |
super().__init__(config)
|
|
|
286 |
hidden_states=outputs.hidden_states,
|
287 |
attentions=outputs.attentions,
|
288 |
)
|
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 438000689
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:601a8071a8e164093f8cbf0ed22b304427f5feff9f93aee5963fc4081a735fe5
|
3 |
size 438000689
|