ybelkada commited on
Commit
5da43cd
1 Parent(s): 3da9fc1

add model weights

Browse files
Files changed (2) hide show
  1. config.json +42 -0
  2. pytorch_model.bin +3 -0
config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_router_probs": false,
3
+ "architectures": [
4
+ "SwitchTransformersForConditionalGeneration"
5
+ ],
6
+ "d_ff": 3072,
7
+ "d_kv": 64,
8
+ "d_model": 768,
9
+ "decoder_sparse_step": 2,
10
+ "dense_act_fn": "relu",
11
+ "dropout_rate": 0.1,
12
+ "encoder_sparse_step": 2,
13
+ "eos_token_id": 1,
14
+ "expert_capacity": 64,
15
+ "feed_forward_proj": "relu",
16
+ "initializer_factor": 1.0,
17
+ "is_encoder_decoder": true,
18
+ "is_gated_act": false,
19
+ "layer_norm_epsilon": 1e-06,
20
+ "model_type": "switch_transformers",
21
+ "num_decoder_layers": 12,
22
+ "num_experts": 32,
23
+ "num_heads": 12,
24
+ "num_layers": 12,
25
+ "num_selected_experts": 1,
26
+ "num_sparse_decoder_layers": 6,
27
+ "num_sparse_encoder_layers": 6,
28
+ "pad_token_id": 0,
29
+ "relative_attention_max_distance": 128,
30
+ "relative_attention_num_buckets": 32,
31
+ "router_aux_loss_coef": 0.001,
32
+ "router_bias": false,
33
+ "router_dtype": "float32",
34
+ "router_ignore_padding_tokens": false,
35
+ "router_jitter_noise": 0.01,
36
+ "router_type": "tokens_masked",
37
+ "router_z_loss_coef": 0.001,
38
+ "torch_dtype": "float32",
39
+ "transformers_version": "4.25.0.dev0",
40
+ "use_cache": true,
41
+ "vocab_size": 32128
42
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57bbda2b9fb0663dc89b2ff3e34cdf38fd749c27610a7a7aee9348e279df151a
3
+ size 7914402428