roberta
aljagne commited on
Commit
7f9f3e1
·
verified ·
1 Parent(s): 35459cb

Create config.json

Browse files

Initial model configuration for AfriLION-Base:

- RoBERTa-base architecture (12 layers, 768 hidden size)
- 50K vocabulary size for multilingual African language support
- 12 attention heads with GELU activation
- Max sequence length: 514 tokens
- Dropout: 0.1 for regularization
- Compatible with Hugging Face Transformers 4.36.0+

Files changed (1) hide show
  1. config.json +25 -0
config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "RobertaForMaskedLM"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 2,
7
+ "eos_token_id": 3,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.1,
10
+ "hidden_size": 768,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 3072,
13
+ "layer_norm_eps": 1e-05,
14
+ "max_position_embeddings": 514,
15
+ "model_type": "roberta",
16
+ "num_attention_heads": 12,
17
+ "num_hidden_layers": 12,
18
+ "pad_token_id": 0,
19
+ "position_embedding_type": "absolute",
20
+ "transformers_version": "4.36.0",
21
+ "type_vocab_size": 1,
22
+ "use_cache": true,
23
+ "vocab_size": 50000,
24
+ "tokenizer_class": "PreTrainedTokenizerFast"
25
+ }