Hugging Face's logo

Surpem
/
Sarden1

Token Classification
Transformers
Safetensors
PyTorch
sarden1
pii
ner
privacy
Model card Files
xet
 Community
vincent-benton commited on
Commit
55b8a6a
·
verified ·
1 Parent(s): 0bee8f7

Add files using upload-large-folder tool

Browse files
Files changed (5) hide show
  1. config.json +72 -0
  2. label_map.json +56 -0
  3. model.safetensors +3 -0
  4. tokenizer.json +0 -0
  5. tokenizer_config.json +13 -0
config.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "sarden1",
3
+ "architectures": [
4
+ "Sarden1"
5
+ ],
6
+ "vocab_size": 50257,
7
+ "hidden_size": 1024,
8
+ "num_hidden_layers": 18,
9
+ "num_q_heads": 16,
10
+ "num_kv_heads": 4,
11
+ "ffn_multiplier": 2.667,
12
+ "max_position_embeddings": 128,
13
+ "rope_theta": 500000.0,
14
+ "dropout": 0.05,
15
+ "num_labels": 25,
16
+ "label2id": {
17
+ "O": 0,
18
+ "B-PERSON": 1,
19
+ "I-PERSON": 2,
20
+ "B-EMAIL": 3,
21
+ "I-EMAIL": 4,
22
+ "B-PHONE": 5,
23
+ "I-PHONE": 6,
24
+ "B-ADDRESS": 7,
25
+ "I-ADDRESS": 8,
26
+ "B-SSN": 9,
27
+ "I-SSN": 10,
28
+ "B-CREDITCARD": 11,
29
+ "I-CREDITCARD": 12,
30
+ "B-IP": 13,
31
+ "I-IP": 14,
32
+ "B-DATE": 15,
33
+ "I-DATE": 16,
34
+ "B-ORG": 17,
35
+ "I-ORG": 18,
36
+ "B-USERNAME": 19,
37
+ "I-USERNAME": 20,
38
+ "B-PASSPORT": 21,
39
+ "I-PASSPORT": 22,
40
+ "B-DRIVERSLICENSE": 23,
41
+ "I-DRIVERSLICENSE": 24
42
+ },
43
+ "id2label": {
44
+ "0": "O",
45
+ "1": "B-PERSON",
46
+ "2": "I-PERSON",
47
+ "3": "B-EMAIL",
48
+ "4": "I-EMAIL",
49
+ "5": "B-PHONE",
50
+ "6": "I-PHONE",
51
+ "7": "B-ADDRESS",
52
+ "8": "I-ADDRESS",
53
+ "9": "B-SSN",
54
+ "10": "I-SSN",
55
+ "11": "B-CREDITCARD",
56
+ "12": "I-CREDITCARD",
57
+ "13": "B-IP",
58
+ "14": "I-IP",
59
+ "15": "B-DATE",
60
+ "16": "I-DATE",
61
+ "17": "B-ORG",
62
+ "18": "I-ORG",
63
+ "19": "B-USERNAME",
64
+ "20": "I-USERNAME",
65
+ "21": "B-PASSPORT",
66
+ "22": "I-PASSPORT",
67
+ "23": "B-DRIVERSLICENSE",
68
+ "24": "I-DRIVERSLICENSE"
69
+ },
70
+ "val_loss": 0.001700466678647155,
71
+ "torch_dtype": "bfloat16"
72
+ }
label_map.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "label2id": {
3
+ "O": 0,
4
+ "B-PERSON": 1,
5
+ "I-PERSON": 2,
6
+ "B-EMAIL": 3,
7
+ "I-EMAIL": 4,
8
+ "B-PHONE": 5,
9
+ "I-PHONE": 6,
10
+ "B-ADDRESS": 7,
11
+ "I-ADDRESS": 8,
12
+ "B-SSN": 9,
13
+ "I-SSN": 10,
14
+ "B-CREDITCARD": 11,
15
+ "I-CREDITCARD": 12,
16
+ "B-IP": 13,
17
+ "I-IP": 14,
18
+ "B-DATE": 15,
19
+ "I-DATE": 16,
20
+ "B-ORG": 17,
21
+ "I-ORG": 18,
22
+ "B-USERNAME": 19,
23
+ "I-USERNAME": 20,
24
+ "B-PASSPORT": 21,
25
+ "I-PASSPORT": 22,
26
+ "B-DRIVERSLICENSE": 23,
27
+ "I-DRIVERSLICENSE": 24
28
+ },
29
+ "id2label": {
30
+ "0": "O",
31
+ "1": "B-PERSON",
32
+ "2": "I-PERSON",
33
+ "3": "B-EMAIL",
34
+ "4": "I-EMAIL",
35
+ "5": "B-PHONE",
36
+ "6": "I-PHONE",
37
+ "7": "B-ADDRESS",
38
+ "8": "I-ADDRESS",
39
+ "9": "B-SSN",
40
+ "10": "I-SSN",
41
+ "11": "B-CREDITCARD",
42
+ "12": "I-CREDITCARD",
43
+ "13": "B-IP",
44
+ "14": "I-IP",
45
+ "15": "B-DATE",
46
+ "16": "I-DATE",
47
+ "17": "B-ORG",
48
+ "18": "I-ORG",
49
+ "19": "B-USERNAME",
50
+ "20": "I-USERNAME",
51
+ "21": "B-PASSPORT",
52
+ "22": "I-PASSPORT",
53
+ "23": "B-DRIVERSLICENSE",
54
+ "24": "I-DRIVERSLICENSE"
55
+ }
56
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d46db240796db464b20bc4dbc93096fab2844ef01255e4e7f2312d833cd2d2a
3
+ size 1003569304
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<|endoftext|>",
5
+ "eos_token": "<|endoftext|>",
6
+ "errors": "replace",
7
+ "is_local": false,
8
+ "local_files_only": false,
9
+ "model_max_length": 1024,
10
+ "pad_token": "<|endoftext|>",
11
+ "tokenizer_class": "GPT2Tokenizer",
12
+ "unk_token": "<|endoftext|>"
13
+ }