Nidum.AI
commited on
Commit
·
2391047
1
Parent(s):
e853b1e
Add model card and license
Browse files
LICENSE
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
Apache-2.0
|
README.md
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags: [moe, pruning, minimax, bfloat16, sglang]
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
---
|
| 5 |
+
|
| 6 |
+
# MiniMax-M2-THRIFT (Pruned to 192 experts)
|
| 7 |
+
Base: ModelCloud/MiniMax-M2-BF16
|
| 8 |
+
Compression: 25% expert pruning (256 -> 192), top_k = 8
|