docs: update README.md
#34
by
eltociear
- opened
README.md
CHANGED
@@ -25,7 +25,7 @@ library_name: transformers
|
|
25 |
- Training Stage: Pretraining & Post-training
|
26 |
- Architecture: transformers with RoPE, SwiGLU, RMSNorm, and Attention QKV bias
|
27 |
- Number of Parameters: 32.5B
|
28 |
-
- Number of
|
29 |
- Number of Layers: 64
|
30 |
- Number of Attention Heads (GQA): 40 for Q and 8 for KV
|
31 |
- Context Length: Full 32,768 tokens
|
|
|
25 |
- Training Stage: Pretraining & Post-training
|
26 |
- Architecture: transformers with RoPE, SwiGLU, RMSNorm, and Attention QKV bias
|
27 |
- Number of Parameters: 32.5B
|
28 |
+
- Number of Parameters (Non-Embedding): 31.0B
|
29 |
- Number of Layers: 64
|
30 |
- Number of Attention Heads (GQA): 40 for Q and 8 for KV
|
31 |
- Context Length: Full 32,768 tokens
|