km
commited on
Commit
•
5e0f47f
1
Parent(s):
d6cb10f
Model release
Browse files- config.json +31 -0
- pytorch_model.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"d_model": 2560,
|
| 3 |
+
"d_intermediate": 0,
|
| 4 |
+
"n_layer": 64,
|
| 5 |
+
"vocab_size": 50277,
|
| 6 |
+
"ssm_cfg": {
|
| 7 |
+
"layer": "Mamba2"
|
| 8 |
+
},
|
| 9 |
+
"attn_layer_idx": [
|
| 10 |
+
9,
|
| 11 |
+
18,
|
| 12 |
+
27,
|
| 13 |
+
36,
|
| 14 |
+
45,
|
| 15 |
+
56
|
| 16 |
+
],
|
| 17 |
+
"attn_cfg": {
|
| 18 |
+
"causal": true,
|
| 19 |
+
"d_conv": 4,
|
| 20 |
+
"head_dim": 128,
|
| 21 |
+
"num_heads": 30,
|
| 22 |
+
"out_proj_bias": false,
|
| 23 |
+
"qkv_proj_bias": false,
|
| 24 |
+
"rotary_emb_dim": 64
|
| 25 |
+
},
|
| 26 |
+
"rms_norm": true,
|
| 27 |
+
"residual_in_fp32": true,
|
| 28 |
+
"fused_add_norm": true,
|
| 29 |
+
"pad_vocab_size_multiple": 16,
|
| 30 |
+
"tie_embeddings": true
|
| 31 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:730142f91e21e6225291625c3f7cbe8fbd2ccc824bf567b13aadead4d2909354
|
| 3 |
+
size 5396089002
|