mohammadmahdinouri commited on
Commit
38f4b16
·
verified ·
1 Parent(s): 893c3ad

Upload ModernALBERTForMaskedLM

Browse files
Files changed (2) hide show
  1. config.json +4 -4
  2. model.safetensors +2 -2
config.json CHANGED
@@ -11,9 +11,9 @@
11
  "group_depth": 4,
12
  "hidden_act": "gelu",
13
  "hidden_dropout_prob": 0.1,
14
- "hidden_size": 1024,
15
  "initializer_range": 0.02,
16
- "intermediate_size": 2624,
17
  "layer_norm_eps": 1e-06,
18
  "load_balancing_loss_coef": 0.01,
19
  "lora_alpha": 32,
@@ -21,9 +21,9 @@
21
  "max_position_embeddings": 8192,
22
  "model_type": "ModernALBERT",
23
  "num_attention_heads": 16,
24
- "num_expert_modules": 3,
25
  "num_experts": 8,
26
- "num_hidden_layers": 16,
27
  "pad_token_id": 0,
28
  "router_jitter_noise": 0.01,
29
  "top_k": 2,
 
11
  "group_depth": 4,
12
  "hidden_act": "gelu",
13
  "hidden_dropout_prob": 0.1,
14
+ "hidden_size": 768,
15
  "initializer_range": 0.02,
16
+ "intermediate_size": 1152,
17
  "layer_norm_eps": 1e-06,
18
  "load_balancing_loss_coef": 0.01,
19
  "lora_alpha": 32,
 
21
  "max_position_embeddings": 8192,
22
  "model_type": "ModernALBERT",
23
  "num_attention_heads": 16,
24
+ "num_expert_modules": 5,
25
  "num_experts": 8,
26
+ "num_hidden_layers": 20,
27
  "pad_token_id": 0,
28
  "router_jitter_noise": 0.01,
29
  "top_k": 2,
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:486765c4dbe88800f6d998fe681a8dc0c8d8b8f335a21fed731f3aa6c6ae3520
3
- size 303132304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d68c71065b9e2005166b283da711cf94264ea1f834ba81616d192a8dee5b673
3
+ size 220012896