Update README.md
Browse files
README.md
CHANGED
|
@@ -1,15 +1,33 @@
|
|
| 1 |
---
|
| 2 |
-
license:
|
| 3 |
datasets:
|
| 4 |
- facebook/belebele
|
| 5 |
---
|
| 6 |
-
Pretrained toy
|
| 7 |
|
| 8 |
-
|
|
|
|
| 9 |
* batch_size = 64
|
| 10 |
* block_size = 256
|
| 11 |
* n_layer = 8
|
| 12 |
* n_head = 8
|
| 13 |
* n_embd = 768
|
|
|
|
| 14 |
|
| 15 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
license: other
|
| 3 |
datasets:
|
| 4 |
- facebook/belebele
|
| 5 |
---
|
| 6 |
+
Pretrained toy models. Made with Andrej Karpathy's NanoGPT.
|
| 7 |
|
| 8 |
+
# nano_35m
|
| 9 |
+
* Trained late 2023 on part of Tagalog portion of Belebele.
|
| 10 |
* batch_size = 64
|
| 11 |
* block_size = 256
|
| 12 |
* n_layer = 8
|
| 13 |
* n_head = 8
|
| 14 |
* n_embd = 768
|
| 15 |
+
* Everything else is left as is.
|
| 16 |
|
| 17 |
+
# nano_76m
|
| 18 |
+
* Trained January 2024 on part of Tagalog portion of Belebele.
|
| 19 |
+
* batch_size = 64
|
| 20 |
+
* block_size = 256
|
| 21 |
+
* n_layer = 11
|
| 22 |
+
* n_head = 16
|
| 23 |
+
* n_embd = 768
|
| 24 |
+
* Everything else is left as is.
|
| 25 |
+
|
| 26 |
+
# nano-ito_35m
|
| 27 |
+
* Trained March 2024 on part of PALITO Tagalog dataset.
|
| 28 |
+
* batch_size = 64
|
| 29 |
+
* block_size = 256
|
| 30 |
+
* n_layer = 11
|
| 31 |
+
* n_head = 16
|
| 32 |
+
* n_embd = 512
|
| 33 |
+
* Everything else is left as is.
|