Eunhui commited on
Commit
206b0ce
1 Parent(s): 2b8e80e

Upload 3 files

Browse files

SciDeBERTa-full is trained with S2ORC dataset which includes all categories and not only abstract but also body texts, the total size of the dataset is about 260GB.
SciDeBERTa-full is trained from scratch and the total number of training steps is 500K.
The experimental result with SciDeBERTa-full model is as below picture.


![scideberta-experimental-result.png](https://s3.amazonaws.com/moonup/production/uploads/1678414581482-601ccd46cfb8edae5efa7556.png)

Files changed (3) hide show
  1. config.json +25 -0
  2. pytorch_model.bin +3 -0
  3. spm.model +3 -0
config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "deberta-v2",
3
+ "attention_head_size": 64,
4
+ "attention_probs_dropout_prob": 0.1,
5
+ "hidden_act": "gelu",
6
+ "hidden_dropout_prob": 0.1,
7
+ "hidden_size": 768,
8
+ "initializer_range": 0.02,
9
+ "intermediate_size": 3072,
10
+ "layer_norm_eps": 1e-07,
11
+ "max_position_embeddings": 512,
12
+ "max_relative_positions": -1,
13
+ "norm_rel_ebd": "layer_norm",
14
+ "num_attention_heads": 12,
15
+ "num_hidden_layers": 12,
16
+ "padding_idx": 0,
17
+ "pos_att_type": "p2c|c2p",
18
+ "position_biased_input": false,
19
+ "position_buckets": 256,
20
+ "relative_attention": true,
21
+ "share_att_key": true,
22
+ "type_vocab_size": 0,
23
+ "vocab_size": 128100
24
+ }
25
+
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:614b652d8e1518ca71f2b5792b169357f231e0b6818e756542a4b9634ebd7295
3
+ size 369920519
spm.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c679fbf93643d19aab7ee10c0b99e460bdbc02fedf34b92b05af343b4af586fd
3
+ size 2464616