Upload 6 files
Browse files- config.json +2 -5
- pytorch_model.bin +2 -2
config.json
CHANGED
|
@@ -17,8 +17,7 @@
|
|
| 17 |
"hidden_act": "gelu",
|
| 18 |
"hidden_size": 16,
|
| 19 |
"hybrid_layer_ids": [
|
| 20 |
-
2
|
| 21 |
-
4
|
| 22 |
],
|
| 23 |
"initializer_range": 0.02,
|
| 24 |
"intermediate_size": 64,
|
|
@@ -27,8 +26,6 @@
|
|
| 27 |
"mamba",
|
| 28 |
"mamba",
|
| 29 |
"hybrid",
|
| 30 |
-
"mamba",
|
| 31 |
-
"hybrid",
|
| 32 |
"mamba"
|
| 33 |
],
|
| 34 |
"mamba_d_conv": 4,
|
|
@@ -40,7 +37,7 @@
|
|
| 40 |
"model_type": "zamba2",
|
| 41 |
"n_mamba_heads": 8,
|
| 42 |
"num_attention_heads": 32,
|
| 43 |
-
"num_hidden_layers":
|
| 44 |
"num_key_value_heads": 32,
|
| 45 |
"num_logits_to_keep": 1,
|
| 46 |
"num_mem_blocks": 1,
|
|
|
|
| 17 |
"hidden_act": "gelu",
|
| 18 |
"hidden_size": 16,
|
| 19 |
"hybrid_layer_ids": [
|
| 20 |
+
2
|
|
|
|
| 21 |
],
|
| 22 |
"initializer_range": 0.02,
|
| 23 |
"intermediate_size": 64,
|
|
|
|
| 26 |
"mamba",
|
| 27 |
"mamba",
|
| 28 |
"hybrid",
|
|
|
|
|
|
|
| 29 |
"mamba"
|
| 30 |
],
|
| 31 |
"mamba_d_conv": 4,
|
|
|
|
| 37 |
"model_type": "zamba2",
|
| 38 |
"n_mamba_heads": 8,
|
| 39 |
"num_attention_heads": 32,
|
| 40 |
+
"num_hidden_layers": 4,
|
| 41 |
"num_key_value_heads": 32,
|
| 42 |
"num_logits_to_keep": 1,
|
| 43 |
"num_mem_blocks": 1,
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fa28ee93e29e40c5fc35e5727122220d5dc937b65f27f71995c8edbb056e551c
|
| 3 |
+
size 3408858
|