Add configuration class
Browse files- configuration_tinymind.py +19 -0
configuration_tinymind.py
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""TinyMind configuration."""
|
| 2 |
+
from transformers import PretrainedConfig
|
| 3 |
+
|
| 4 |
+
|
| 5 |
+
class TinyMindConfig(PretrainedConfig):
|
| 6 |
+
model_type = "tiny_smart_llm"
|
| 7 |
+
|
| 8 |
+
def __init__(self, vocab_size=50257, n_embd=256, n_heads=8, n_layers=6, max_seq_len=512, dropout=0.1, **kwargs):
|
| 9 |
+
self.vocab_size = vocab_size
|
| 10 |
+
self.n_embd = n_embd
|
| 11 |
+
self.n_heads = n_heads
|
| 12 |
+
self.n_layers = n_layers
|
| 13 |
+
self.num_hidden_layers = n_layers
|
| 14 |
+
self.hidden_size = n_embd
|
| 15 |
+
self.num_attention_heads = n_heads
|
| 16 |
+
self.max_seq_len = max_seq_len
|
| 17 |
+
self.max_position_embeddings = max_seq_len
|
| 18 |
+
self.dropout = dropout
|
| 19 |
+
super().__init__(**kwargs)
|