ASEpochs commited on
Commit
6aa7d57
·
verified ·
1 Parent(s): dad4dc2

Upload config.json

Browse files
Files changed (1) hide show
  1. config.json +38 -0
config.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "endexformer",
3
+ "apply_aux_loss": true,
4
+ "architectures": [
5
+ "TimeMoeForPrediction"
6
+ ],
7
+ "auto_map": {
8
+ "AutoConfig": "configuration_time_moe.TimeMoeConfig",
9
+ "AutoModelForCausalLM": "modeling_time_moe.TimeMoeForPrediction"
10
+ },
11
+ "attention_dropout": 0.0,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 384,
14
+ "horizon_lengths": [
15
+ 1,
16
+ 8,
17
+ 32,
18
+ 64
19
+ ],
20
+ "initializer_range": 0.02,
21
+ "input_size": 1,
22
+ "intermediate_size": 1536,
23
+ "max_position_embeddings": 4096,
24
+ "model_type": "time_moe",
25
+ "num_attention_heads": 12,
26
+ "num_experts": 8,
27
+ "num_experts_per_tok": 2,
28
+ "num_hidden_layers": 12,
29
+ "num_key_value_heads": 12,
30
+ "rms_norm_eps": 1e-06,
31
+ "rope_theta": 10000,
32
+ "router_aux_loss_factor": 0.02,
33
+ "tie_word_embeddings": false,
34
+ "torch_dtype": "bfloat16",
35
+ "transformers_version": "4.40.1",
36
+ "use_cache": true,
37
+ "use_dense": false
38
+ }