lukealonso commited on
Commit
fe9ce92
·
verified ·
1 Parent(s): 0b9dfab

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -4,6 +4,8 @@ base_model:
4
  license: mit
5
  ---
6
 
 
 
7
  ## Model Description
8
 
9
  **GLM-5.1-NVFP4** is an NVFP4-quantized version of [zai-org/GLM-5.1](https://huggingface.co/zai-org/GLM-5.1), a 744B-parameter Mixture-of-Experts language model with 40B active parameters, 256 experts per MoE layer (8 activated per token), and DeepSeek Sparse Attention (DSA).
@@ -28,8 +30,6 @@ Three calibration passes were run:
28
 
29
  Hardware: 8x RTX PRO 6000 Blackwell 96GB (b12x MoE runner recommended)
30
 
31
- Note: You must run with `--disable-shared-experts-fusion` in sglang, otherwise it will incorrectly attempt to fuse the BF16 shared expert.
32
-
33
  ### Community Testing
34
 
35
  ```
 
4
  license: mit
5
  ---
6
 
7
+ Note: You must run with --disable-shared-experts-fusion in sglang, otherwise it will incorrectly attempt to fuse the BF16 shared expert.
8
+
9
  ## Model Description
10
 
11
  **GLM-5.1-NVFP4** is an NVFP4-quantized version of [zai-org/GLM-5.1](https://huggingface.co/zai-org/GLM-5.1), a 744B-parameter Mixture-of-Experts language model with 40B active parameters, 256 experts per MoE layer (8 activated per token), and DeepSeek Sparse Attention (DSA).
 
30
 
31
  Hardware: 8x RTX PRO 6000 Blackwell 96GB (b12x MoE runner recommended)
32
 
 
 
33
  ### Community Testing
34
 
35
  ```