Yishi Xu
commited on
Commit
•
700743b
1
Parent(s):
b3d4aee
added CerebrasGPT-6.7B_SP
Browse files
CerebrasGPT-6.7B_SP/checkpoint_24000_to_hf.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bbaf1ae47230cebe2512ae471c28f05ee385e223bce4cd33593d229805ecc9f6
|
3 |
+
size 26767985043
|
CerebrasGPT-6.7B_SP/checkpoint_42000_to_hf.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c4d8f897f3b503f52fc561ab82092233683e3aa43aa467592ea725e856841e81
|
3 |
+
size 26767985043
|
CerebrasGPT-6.7B_SP/params_to_hf.json
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"n_embd": 4096,
|
3 |
+
"tie_word_embeddings": true,
|
4 |
+
"n_positions": 2048,
|
5 |
+
"vocab_size": 50257,
|
6 |
+
"n_layer": 32,
|
7 |
+
"resid_pdrop": 0.0,
|
8 |
+
"layer_norm_epsilon": 1e-05,
|
9 |
+
"n_head": 32,
|
10 |
+
"scale_attn_weights": true,
|
11 |
+
"attn_pdrop": 0.0,
|
12 |
+
"n_inner": 16384,
|
13 |
+
"activation_function": "gelu",
|
14 |
+
"embd_pdrop": 0.0
|
15 |
+
}
|