Hugging Face's logo

jessietextstan
/
bart-large-mnli

Text Classification
Transformers
PyTorch
TensorBoard
bert
Generated from Trainer
text-embeddings-inference
Model card Files
xet
 Metrics Community
2
jessietextstan commited on
Commit
360439c
·
1 Parent(s): 4bb558a

Training in progress, epoch 1

Browse files
Files changed (19) hide show
  1. config.json +18 -32
  2. pytorch_model.bin +2 -2
  3. runs/Dec29_10-12-37_STAN-Dev/1672326760.7588515/events.out.tfevents.1672326760.STAN-Dev.20184.1 +3 -0
  4. runs/Dec29_10-12-37_STAN-Dev/events.out.tfevents.1672326760.STAN-Dev.20184.0 +3 -0
  5. runs/Dec29_10-15-06_STAN-Dev/1672326910.4012887/events.out.tfevents.1672326910.STAN-Dev.17732.1 +3 -0
  6. runs/Dec29_10-15-06_STAN-Dev/events.out.tfevents.1672326910.STAN-Dev.17732.0 +3 -0
  7. runs/Dec29_10-16-17_STAN-Dev/1672326981.0348406/events.out.tfevents.1672326981.STAN-Dev.4560.1 +3 -0
  8. runs/Dec29_10-16-17_STAN-Dev/events.out.tfevents.1672326981.STAN-Dev.4560.0 +3 -0
  9. runs/Dec29_10-17-32_STAN-Dev/1672327055.841558/events.out.tfevents.1672327055.STAN-Dev.3168.1 +3 -0
  10. runs/Dec29_10-17-32_STAN-Dev/events.out.tfevents.1672327055.STAN-Dev.3168.0 +3 -0
  11. runs/Dec29_11-29-16_STAN-Dev/1672331359.5547507/events.out.tfevents.1672331359.STAN-Dev.15260.1 +3 -0
  12. runs/Dec29_11-29-16_STAN-Dev/events.out.tfevents.1672331359.STAN-Dev.15260.0 +3 -0
  13. runs/Dec29_11-34-07_STAN-Dev/1672331651.1410754/events.out.tfevents.1672331651.STAN-Dev.17480.1 +3 -0
  14. runs/Dec29_11-34-07_STAN-Dev/events.out.tfevents.1672331651.STAN-Dev.17480.0 +3 -0
  15. runs/Dec29_11-43-21_STAN-Dev/1672332205.4430766/events.out.tfevents.1672332205.STAN-Dev.4960.1 +3 -0
  16. runs/Dec29_11-43-21_STAN-Dev/events.out.tfevents.1672332205.STAN-Dev.4960.0 +3 -0
  17. runs/Dec29_11-46-43_STAN-Dev/1672332407.3218734/events.out.tfevents.1672332407.STAN-Dev.14360.1 +3 -0
  18. runs/Dec29_11-46-43_STAN-Dev/events.out.tfevents.1672332407.STAN-Dev.14360.0 +3 -0
  19. training_args.bin +1 -1
config.json CHANGED
@@ -1,30 +1,14 @@
1
  {
2
- "_name_or_path": "facebook/bart-large-mnli",
3
- "_num_labels": 3,
4
- "activation_dropout": 0.0,
5
- "activation_function": "gelu",
6
- "add_final_layer_norm": false,
7
  "architectures": [
8
- "BartForSequenceClassification"
9
  ],
10
- "attention_dropout": 0.0,
11
- "bos_token_id": 0,
12
- "classif_dropout": 0.0,
13
- "classifier_dropout": 0.0,
14
- "d_model": 1024,
15
- "decoder_attention_heads": 16,
16
- "decoder_ffn_dim": 4096,
17
- "decoder_layerdrop": 0.0,
18
- "decoder_layers": 12,
19
- "decoder_start_token_id": 2,
20
- "dropout": 0.1,
21
- "encoder_attention_heads": 16,
22
- "encoder_ffn_dim": 4096,
23
- "encoder_layerdrop": 0.0,
24
- "encoder_layers": 12,
25
- "eos_token_id": 2,
26
- "forced_eos_token_id": 2,
27
  "gradient_checkpointing": false,
 
 
 
28
  "id2label": {
29
  "0": "fees",
30
  "1": "arc",
@@ -52,8 +36,8 @@
52
  "8": "parking",
53
  "9": "bicycle"
54
  },
55
- "init_std": 0.02,
56
- "is_encoder_decoder": true,
57
  "label2id": {
58
  "amenity": "12",
59
  "arc": "1",
@@ -81,15 +65,17 @@
81
  "service request": "3",
82
  "smoking": "18"
83
  },
84
- "max_position_embeddings": 1024,
85
- "model_type": "bart",
86
- "normalize_before": false,
 
87
  "num_hidden_layers": 12,
88
- "output_past": false,
89
- "pad_token_id": 1,
90
- "scale_embedding": false,
91
  "torch_dtype": "float32",
92
  "transformers_version": "4.25.1",
 
93
  "use_cache": true,
94
- "vocab_size": 50265
95
  }
 
1
  {
2
+ "_name_or_path": "bert-base-cased",
 
 
 
 
3
  "architectures": [
4
+ "BertForSequenceClassification"
5
  ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "classifier_dropout": null,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8
  "gradient_checkpointing": false,
9
+ "hidden_act": "gelu",
10
+ "hidden_dropout_prob": 0.1,
11
+ "hidden_size": 768,
12
  "id2label": {
13
  "0": "fees",
14
  "1": "arc",
 
36
  "8": "parking",
37
  "9": "bicycle"
38
  },
39
+ "initializer_range": 0.02,
40
+ "intermediate_size": 3072,
41
  "label2id": {
42
  "amenity": "12",
43
  "arc": "1",
 
65
  "service request": "3",
66
  "smoking": "18"
67
  },
68
+ "layer_norm_eps": 1e-12,
69
+ "max_position_embeddings": 512,
70
+ "model_type": "bert",
71
+ "num_attention_heads": 12,
72
  "num_hidden_layers": 12,
73
+ "pad_token_id": 0,
74
+ "position_embedding_type": "absolute",
75
+ "problem_type": "single_label_classification",
76
  "torch_dtype": "float32",
77
  "transformers_version": "4.25.1",
78
+ "type_vocab_size": 2,
79
  "use_cache": true,
80
+ "vocab_size": 28996
81
  }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0723349e9c249dbe8caa2a1c9611331e2498ed5bda412a9b88a8c38af50abc4a
3
- size 1629634905
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:930f42c19f197df4c1b00d3bcfd7939d36a6bdacc90d7cf9b5aae067b77659c7
3
+ size 433387957
runs/Dec29_10-12-37_STAN-Dev/1672326760.7588515/events.out.tfevents.1672326760.STAN-Dev.20184.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:91f43155c92e60dc57e0b503dcd1e18e44595de9daf52377dbbebea0b6c37da2
3
+ size 5532
runs/Dec29_10-12-37_STAN-Dev/events.out.tfevents.1672326760.STAN-Dev.20184.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4e2865924c4eb74b92f92e4b74a47923de5f49efe073208b2fd54a382e652c6
3
+ size 5315
runs/Dec29_10-15-06_STAN-Dev/1672326910.4012887/events.out.tfevents.1672326910.STAN-Dev.17732.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff38c40a5ed4fbf0b8297b4059541c333acced5accc015524dbbf2506077766d
3
+ size 5532
runs/Dec29_10-15-06_STAN-Dev/events.out.tfevents.1672326910.STAN-Dev.17732.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ece27ffd11cd63d4b4d610e63e1c99e63dfa7b3dbae2eb0595250a4e4f68ea5
3
+ size 5315
runs/Dec29_10-16-17_STAN-Dev/1672326981.0348406/events.out.tfevents.1672326981.STAN-Dev.4560.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2508ffe6f3e99604b314df03b8d95e9fc77066a2232de96d125af0555d1c39e2
3
+ size 5532
runs/Dec29_10-16-17_STAN-Dev/events.out.tfevents.1672326981.STAN-Dev.4560.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:279e19fb92d62a2fb08394c03248943feaa0b2d48ebba3c3bd7f07291981b75b
3
+ size 5315
runs/Dec29_10-17-32_STAN-Dev/1672327055.841558/events.out.tfevents.1672327055.STAN-Dev.3168.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dacffc3d64f135af1ec9130d96eb6d003963db0d6c0e7ccdb0914f00148513be
3
+ size 5532
runs/Dec29_10-17-32_STAN-Dev/events.out.tfevents.1672327055.STAN-Dev.3168.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7bd2356ab5e039a0476359123626753e30f60f01fd1abab2acd901d043ba1a07
3
+ size 5315
runs/Dec29_11-29-16_STAN-Dev/1672331359.5547507/events.out.tfevents.1672331359.STAN-Dev.15260.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e48711cb174464c0eb266154c0ac3d3f131fe18e74e0e938318d29e57ea37a87
3
+ size 5532
runs/Dec29_11-29-16_STAN-Dev/events.out.tfevents.1672331359.STAN-Dev.15260.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:058155e9ba21c9606a6eb8e0520929c78248041f0ea20134b8c871f6211aaf9a
3
+ size 5315
runs/Dec29_11-34-07_STAN-Dev/1672331651.1410754/events.out.tfevents.1672331651.STAN-Dev.17480.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6abc4d7bbf402427c5c806fd78b48c6f43f05888d85a58705657e002e9a3c23b
3
+ size 5532
runs/Dec29_11-34-07_STAN-Dev/events.out.tfevents.1672331651.STAN-Dev.17480.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8b2d59556556a67f5fba2332814d534496cee25b656fde4438b2bbb45ce61d2
3
+ size 5315
runs/Dec29_11-43-21_STAN-Dev/1672332205.4430766/events.out.tfevents.1672332205.STAN-Dev.4960.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:27cdba34252834c7e449b2f49d345d0cbccbc0ac6cfc5dff005fcc9afe94b8e4
3
+ size 5532
runs/Dec29_11-43-21_STAN-Dev/events.out.tfevents.1672332205.STAN-Dev.4960.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e16aee1fbedcb375cdd3b8ae7a34328e9020b722f69ce94f888405b7d92d360a
3
+ size 4924
runs/Dec29_11-46-43_STAN-Dev/1672332407.3218734/events.out.tfevents.1672332407.STAN-Dev.14360.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1282e2d324a2b5bdf343a6c1374d215e503398fa7ae372a247d8b828af83ce93
3
+ size 5532
runs/Dec29_11-46-43_STAN-Dev/events.out.tfevents.1672332407.STAN-Dev.14360.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:adde1c17d6127507236031ff92ded076f2d2a804c53622f859731648d2af8790
3
+ size 5241
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a29f03cb5a515558611c4a97adb2be721431657fde21b805f5be35c12a9daddf
3
  size 3451
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5bc6449b4c32b4be3cb8d18eeff38b808dbec6f58458ac470866fe7b1896ba2e
3
  size 3451