Training in progress, epoch 1
Browse files- .gitignore +1 -0
- config.json +95 -0
- pytorch_model.bin +3 -0
- runs/Mar03_00-56-57_5a8ebeccb3d2/1677805030.42313/events.out.tfevents.1677805030.5a8ebeccb3d2.154.3 +3 -0
- runs/Mar03_00-56-57_5a8ebeccb3d2/events.out.tfevents.1677805030.5a8ebeccb3d2.154.2 +3 -0
- runs/Mar03_00-59-23_5a8ebeccb3d2/1677805171.9033334/events.out.tfevents.1677805171.5a8ebeccb3d2.154.5 +3 -0
- runs/Mar03_00-59-23_5a8ebeccb3d2/events.out.tfevents.1677805171.5a8ebeccb3d2.154.4 +3 -0
- runs/Mar03_01-00-05_5a8ebeccb3d2/1677805214.2310016/events.out.tfevents.1677805214.5a8ebeccb3d2.154.7 +3 -0
- runs/Mar03_01-00-05_5a8ebeccb3d2/events.out.tfevents.1677805214.5a8ebeccb3d2.154.6 +3 -0
- runs/Mar03_01-03-14_5a8ebeccb3d2/1677805407.0421698/events.out.tfevents.1677805407.5a8ebeccb3d2.14327.1 +3 -0
- runs/Mar03_01-03-14_5a8ebeccb3d2/events.out.tfevents.1677805407.5a8ebeccb3d2.14327.0 +3 -0
- runs/Mar03_01-09-20_5a8ebeccb3d2/1677805770.8406239/events.out.tfevents.1677805770.5a8ebeccb3d2.14327.3 +3 -0
- runs/Mar03_01-09-20_5a8ebeccb3d2/events.out.tfevents.1677805770.5a8ebeccb3d2.14327.2 +3 -0
- sentencepiece.bpe.model +3 -0
- special_tokens_map.json +20 -0
- tokenizer.json +0 -0
- tokenizer_config.json +26 -0
- training_args.bin +3 -0
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
checkpoint-*/
|
config.json
ADDED
@@ -0,0 +1,95 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "airesearch/wangchanberta-base-att-spm-uncased",
|
3 |
+
"architectures": [
|
4 |
+
"CamembertForTokenClassification"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"classifier_dropout": null,
|
9 |
+
"eos_token_id": 2,
|
10 |
+
"gradient_checkpointing": false,
|
11 |
+
"hidden_act": "gelu",
|
12 |
+
"hidden_dropout_prob": 0.1,
|
13 |
+
"hidden_size": 768,
|
14 |
+
"id2label": {
|
15 |
+
"0": "O",
|
16 |
+
"1": "B_BRN",
|
17 |
+
"2": "B_DES",
|
18 |
+
"3": "B_DTM",
|
19 |
+
"4": "B_LOC",
|
20 |
+
"5": "B_MEA",
|
21 |
+
"6": "B_NUM",
|
22 |
+
"7": "B_ORG",
|
23 |
+
"8": "B_PER",
|
24 |
+
"9": "B_TRM",
|
25 |
+
"10": "B_TTL",
|
26 |
+
"11": "I_BRN",
|
27 |
+
"12": "I_DES",
|
28 |
+
"13": "I_DTM",
|
29 |
+
"14": "I_LOC",
|
30 |
+
"15": "I_MEA",
|
31 |
+
"16": "I_NUM",
|
32 |
+
"17": "I_ORG",
|
33 |
+
"18": "I_PER",
|
34 |
+
"19": "I_TRM",
|
35 |
+
"20": "I_TTL",
|
36 |
+
"21": "E_BRN",
|
37 |
+
"22": "E_DES",
|
38 |
+
"23": "E_DTM",
|
39 |
+
"24": "E_LOC",
|
40 |
+
"25": "E_MEA",
|
41 |
+
"26": "E_NUM",
|
42 |
+
"27": "E_ORG",
|
43 |
+
"28": "E_PER",
|
44 |
+
"29": "E_TRM",
|
45 |
+
"30": "E_TTL"
|
46 |
+
},
|
47 |
+
"initializer_range": 0.02,
|
48 |
+
"intermediate_size": 3072,
|
49 |
+
"label2id": {
|
50 |
+
"B_BRN": 1,
|
51 |
+
"B_DES": 2,
|
52 |
+
"B_DTM": 3,
|
53 |
+
"B_LOC": 4,
|
54 |
+
"B_MEA": 5,
|
55 |
+
"B_NUM": 6,
|
56 |
+
"B_ORG": 7,
|
57 |
+
"B_PER": 8,
|
58 |
+
"B_TRM": 9,
|
59 |
+
"B_TTL": 10,
|
60 |
+
"E_BRN": 21,
|
61 |
+
"E_DES": 22,
|
62 |
+
"E_DTM": 23,
|
63 |
+
"E_LOC": 24,
|
64 |
+
"E_MEA": 25,
|
65 |
+
"E_NUM": 26,
|
66 |
+
"E_ORG": 27,
|
67 |
+
"E_PER": 28,
|
68 |
+
"E_TRM": 29,
|
69 |
+
"E_TTL": 30,
|
70 |
+
"I_BRN": 11,
|
71 |
+
"I_DES": 12,
|
72 |
+
"I_DTM": 13,
|
73 |
+
"I_LOC": 14,
|
74 |
+
"I_MEA": 15,
|
75 |
+
"I_NUM": 16,
|
76 |
+
"I_ORG": 17,
|
77 |
+
"I_PER": 18,
|
78 |
+
"I_TRM": 19,
|
79 |
+
"I_TTL": 20,
|
80 |
+
"O": 0
|
81 |
+
},
|
82 |
+
"layer_norm_eps": 1e-12,
|
83 |
+
"max_position_embeddings": 512,
|
84 |
+
"model_type": "camembert",
|
85 |
+
"num_attention_head": 12,
|
86 |
+
"num_attention_heads": 12,
|
87 |
+
"num_hidden_layers": 12,
|
88 |
+
"pad_token_id": 1,
|
89 |
+
"position_embedding_type": "absolute",
|
90 |
+
"torch_dtype": "float32",
|
91 |
+
"transformers_version": "4.26.1",
|
92 |
+
"type_vocab_size": 1,
|
93 |
+
"use_cache": true,
|
94 |
+
"vocab_size": 25005
|
95 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:25f10775fce5990fcc87eee2bff8ee2b665d7114b5d8e91361a0a0745dee7e9f
|
3 |
+
size 418783405
|
runs/Mar03_00-56-57_5a8ebeccb3d2/1677805030.42313/events.out.tfevents.1677805030.5a8ebeccb3d2.154.3
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9b57c0cc91ab5f88e6a004e7b12a9fe144bfada9f6f7588e3e24c6526513529f
|
3 |
+
size 5670
|
runs/Mar03_00-56-57_5a8ebeccb3d2/events.out.tfevents.1677805030.5a8ebeccb3d2.154.2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:834e9ef0149de4da2214af5227f3059deb72b2e5a524e2bbe6d2a2c129676e0b
|
3 |
+
size 5043
|
runs/Mar03_00-59-23_5a8ebeccb3d2/1677805171.9033334/events.out.tfevents.1677805171.5a8ebeccb3d2.154.5
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bc26271ff26a464e774072b7ebcf14430625561dc49820c8cd16f5453bc16b03
|
3 |
+
size 5670
|
runs/Mar03_00-59-23_5a8ebeccb3d2/events.out.tfevents.1677805171.5a8ebeccb3d2.154.4
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:05aab8f81c34ace6ce822ef5a853dbd20c3a4b509b1483b1263e4e279f113819
|
3 |
+
size 4136
|
runs/Mar03_01-00-05_5a8ebeccb3d2/1677805214.2310016/events.out.tfevents.1677805214.5a8ebeccb3d2.154.7
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fd85391e4b0adddc118dbbe6710091b4e1987af8f7601cd4005803046e2f9ea7
|
3 |
+
size 5670
|
runs/Mar03_01-00-05_5a8ebeccb3d2/events.out.tfevents.1677805214.5a8ebeccb3d2.154.6
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8f7a5642638aa04127946310b997a63f1b8a27f182444d41a9cc46d18e24bade
|
3 |
+
size 4136
|
runs/Mar03_01-03-14_5a8ebeccb3d2/1677805407.0421698/events.out.tfevents.1677805407.5a8ebeccb3d2.14327.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e56f5d20b525324e9e553c8e8b90aa0aeec04afa1fa5f995dfd72f0b84f4e476
|
3 |
+
size 5670
|
runs/Mar03_01-03-14_5a8ebeccb3d2/events.out.tfevents.1677805407.5a8ebeccb3d2.14327.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c0e0cfd87ffe01d03e17350f722dcbae723f754960fd27ae4f8e8a2c132c918d
|
3 |
+
size 5055
|
runs/Mar03_01-09-20_5a8ebeccb3d2/1677805770.8406239/events.out.tfevents.1677805770.5a8ebeccb3d2.14327.3
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1cb46f127f9eb5b6118e531ef7be832c8799af32c5cb13acbb29b5c0758919c3
|
3 |
+
size 5670
|
runs/Mar03_01-09-20_5a8ebeccb3d2/events.out.tfevents.1677805770.5a8ebeccb3d2.14327.2
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:805a9342ff85849c34e2297a2e16ac3bfdf51d23da7e49a24541fa60f7008029
|
3 |
+
size 7882
|
sentencepiece.bpe.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:49c4ba4e495ddf31eb2fdba7fc6aef3c233091d25d35bc9d24694ccf48ae114c
|
3 |
+
size 904693
|
special_tokens_map.json
ADDED
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"additional_special_tokens": [
|
3 |
+
"<s>NOTUSED",
|
4 |
+
"</s>NOTUSED",
|
5 |
+
"<_>"
|
6 |
+
],
|
7 |
+
"bos_token": "<s>",
|
8 |
+
"cls_token": "<s>",
|
9 |
+
"eos_token": "</s>",
|
10 |
+
"mask_token": {
|
11 |
+
"content": "<mask>",
|
12 |
+
"lstrip": true,
|
13 |
+
"normalized": true,
|
14 |
+
"rstrip": false,
|
15 |
+
"single_word": false
|
16 |
+
},
|
17 |
+
"pad_token": "<pad>",
|
18 |
+
"sep_token": "</s>",
|
19 |
+
"unk_token": "<unk>"
|
20 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"additional_special_tokens": [
|
3 |
+
"<s>NOTUSED",
|
4 |
+
"</s>NOTUSED",
|
5 |
+
"<_>"
|
6 |
+
],
|
7 |
+
"bos_token": "<s>",
|
8 |
+
"cls_token": "<s>",
|
9 |
+
"eos_token": "</s>",
|
10 |
+
"mask_token": {
|
11 |
+
"__type": "AddedToken",
|
12 |
+
"content": "<mask>",
|
13 |
+
"lstrip": true,
|
14 |
+
"normalized": true,
|
15 |
+
"rstrip": false,
|
16 |
+
"single_word": false
|
17 |
+
},
|
18 |
+
"model_max_length": 1000000000000000019884624838656,
|
19 |
+
"name_or_path": "airesearch/wangchanberta-base-att-spm-uncased",
|
20 |
+
"pad_token": "<pad>",
|
21 |
+
"sep_token": "</s>",
|
22 |
+
"sp_model_kwargs": {},
|
23 |
+
"special_tokens_map_file": null,
|
24 |
+
"tokenizer_class": "CamembertTokenizer",
|
25 |
+
"unk_token": "<unk>"
|
26 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b6c2aec947240122348be75297cb4ffc764fe0a4462e0c9d412da8433a53cec1
|
3 |
+
size 3515
|