Update model.py
Browse files
model.py
CHANGED
|
@@ -36,19 +36,19 @@ except ImportError:
|
|
| 36 |
@dataclass
|
| 37 |
class SLMConfig:
|
| 38 |
# Vocab / embedding
|
| 39 |
-
vocab_size: int =
|
| 40 |
-
hidden_size: int =
|
| 41 |
tie_embeddings: bool = True
|
| 42 |
|
| 43 |
# Depth / width
|
| 44 |
-
num_layers: int =
|
| 45 |
-
intermediate_size: int =
|
| 46 |
|
| 47 |
# Attention
|
| 48 |
-
num_heads: int =
|
| 49 |
-
num_kv_heads: int =
|
| 50 |
-
max_seq_len: int =
|
| 51 |
-
rope_theta: float =
|
| 52 |
qk_norm: bool = True
|
| 53 |
use_bias: bool = False
|
| 54 |
|
|
|
|
| 36 |
@dataclass
|
| 37 |
class SLMConfig:
|
| 38 |
# Vocab / embedding
|
| 39 |
+
vocab_size: int = 4096
|
| 40 |
+
hidden_size: int = 256
|
| 41 |
tie_embeddings: bool = True
|
| 42 |
|
| 43 |
# Depth / width
|
| 44 |
+
num_layers: int = 11
|
| 45 |
+
intermediate_size: int = 768
|
| 46 |
|
| 47 |
# Attention
|
| 48 |
+
num_heads: int = 4
|
| 49 |
+
num_kv_heads: int = 2
|
| 50 |
+
max_seq_len: int = 1024
|
| 51 |
+
rope_theta: float = 10000.0
|
| 52 |
qk_norm: bool = True
|
| 53 |
use_bias: bool = False
|
| 54 |
|