ybelkada HF staff commited on
Commit
cec1694
1 Parent(s): f04ae36

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +71 -0
config.json CHANGED
@@ -0,0 +1,71 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Arthur",
3
+ "architectures": [
4
+ "SwitchTransformersForConditionalGeneration"
5
+ ],
6
+ "batch_prioritized_routing": false,
7
+ "d_ff": 3072,
8
+ "d_kv": 64,
9
+ "d_model": 768,
10
+ "decoder_sparse_step": 2,
11
+ "dense_act_fn": "relu",
12
+ "dropout_rate": 0.1,
13
+ "encoder_sparse_step": 2,
14
+ "eos_token_id": 1,
15
+ "expert_capacity": 64,
16
+ "feed_forward_proj": "relu",
17
+ "initializer_factor": 1.0,
18
+ "decoder_start_token_id": 0,
19
+ "is_encoder_decoder": true,
20
+ "is_gated_act": false,
21
+ "layer_norm_epsilon": 1e-06,
22
+ "model_type": "switch_transformers",
23
+ "num_decoder_layers": 12,
24
+ "num_experts": 16,
25
+ "num_heads": 12,
26
+ "num_layers": 12,
27
+ "num_selected_experts": 1,
28
+ "num_sparse_decoder_layers": 6,
29
+ "num_sparse_encoder_layers": 6,
30
+ "pad_token_id": 0,
31
+ "relative_attention_max_distance": 128,
32
+ "relative_attention_num_buckets": 32,
33
+ "router_bias": false,
34
+ "router_dtype": "float32",
35
+ "router_ignore_padding_tokens": false,
36
+ "router_jitter_noise": 0,
37
+ "router_type": "tokens_masked",
38
+ "task_specific_params": {
39
+ "summarization": {
40
+ "early_stopping": true,
41
+ "length_penalty": 2.0,
42
+ "max_length": 200,
43
+ "min_length": 30,
44
+ "no_repeat_ngram_size": 3,
45
+ "num_beams": 4,
46
+ "prefix": "summarize: "
47
+ },
48
+ "translation_en_to_de": {
49
+ "early_stopping": true,
50
+ "max_length": 300,
51
+ "num_beams": 4,
52
+ "prefix": "translate English to German: "
53
+ },
54
+ "translation_en_to_fr": {
55
+ "early_stopping": true,
56
+ "max_length": 300,
57
+ "num_beams": 4,
58
+ "prefix": "translate English to French: "
59
+ },
60
+ "translation_en_to_ro": {
61
+ "early_stopping": true,
62
+ "max_length": 300,
63
+ "num_beams": 4,
64
+ "prefix": "translate English to Romanian: "
65
+ }
66
+ },
67
+ "torch_dtype": "bfloat16",
68
+ "transformers_version": "4.24.0.dev0",
69
+ "use_cache": true,
70
+ "vocab_size": 32128
71
+ }