jscore2023 commited on
Commit
1e23382
1 Parent(s): 2982534

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +9 -7
config.json CHANGED
@@ -8,7 +8,11 @@
8
  "attention_dropout": 0.0,
9
  "auto_map": {
10
  "AutoConfig": "configuration_RW.RWConfig",
11
- "AutoModelForCausalLM": "modelling_RW.RWForCausalLM"
 
 
 
 
12
  },
13
  "bias": false,
14
  "bos_token_id": 11,
@@ -19,13 +23,11 @@
19
  "layer_norm_epsilon": 1e-05,
20
  "model_type": "RefinedWebModel",
21
  "multi_query": true,
22
- "new_decoder_architecture": false,
23
- "num_attention_heads": 71,
24
- "num_hidden_layers": 32,
25
- "num_kv_heads": 71,
26
  "parallel_attn": true,
27
  "torch_dtype": "bfloat16",
28
- "transformers_version": "4.30.0.dev0",
29
- "use_cache": true,
30
  "vocab_size": 65024
31
  }
 
8
  "attention_dropout": 0.0,
9
  "auto_map": {
10
  "AutoConfig": "configuration_RW.RWConfig",
11
+ "AutoModel": "modelling_RW.RWModel",
12
+ "AutoModelForCausalLM": "modelling_RW.RWForCausalLM",
13
+ "AutoModelForQuestionAnswering": "modelling_RW.RWForQuestionAnswering",
14
+ "AutoModelForSequenceClassification": "modelling_RW.RWForSequenceClassification",
15
+ "AutoModelForTokenClassification": "modelling_RW.RWForTokenClassification"
16
  },
17
  "bias": false,
18
  "bos_token_id": 11,
 
23
  "layer_norm_epsilon": 1e-05,
24
  "model_type": "RefinedWebModel",
25
  "multi_query": true,
26
+ "n_head": 71,
27
+ "n_layer": 32,
 
 
28
  "parallel_attn": true,
29
  "torch_dtype": "bfloat16",
30
+ "transformers_version": "4.28.1",
31
+ "use_cache": false,
32
  "vocab_size": 65024
33
  }