abcsk123 commited on
Commit
7955339
·
verified ·
1 Parent(s): 4c6edb8

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -25,8 +25,8 @@ tags:
25
  This is a custom-architected model based on `Qwen2.5-Coder-1.5B`. We introduced a novel **Asymmetric Hybrid Architecture (GQA + MLA)** with **Cross-Layer Shared Latent Gates** and **Attention Sinks**, enabling efficient feature communication and reduced KV-Cache memory footprint.
26
 
27
  ## 🏗️ Architecture Innovations
28
- ![Uploading qwen2_hybrid_arch.png…]()
29
 
 
30
 
31
  Unlike standard Qwen2 models, this `Hybrid-v9` backbone features:
32
  1. **Asymmetric Layers:**
 
25
  This is a custom-architected model based on `Qwen2.5-Coder-1.5B`. We introduced a novel **Asymmetric Hybrid Architecture (GQA + MLA)** with **Cross-Layer Shared Latent Gates** and **Attention Sinks**, enabling efficient feature communication and reduced KV-Cache memory footprint.
26
 
27
  ## 🏗️ Architecture Innovations
 
28
 
29
+ ![qwen2_hybrid_arch](https://cdn-uploads.huggingface.co/production/uploads/67cd51087c6e6ea1cc18d236/2pYTkGmwcgRMxQr8qBVB3.png)
30
 
31
  Unlike standard Qwen2 models, this `Hybrid-v9` backbone features:
32
  1. **Asymmetric Layers:**