Datasets:
cross referencing extended work.
Browse files
README.md
CHANGED
|
@@ -27,6 +27,7 @@ datasets:
|
|
| 27 |
8. [Optimization Techniques](#8-optimization-techniques)
|
| 28 |
9. [Lessons Learned](#9-lessons-learned-from-implementing-muon)
|
| 29 |
10. [Concludsion](#10-conclusion)
|
|
|
|
| 30 |
|
| 31 |
## 🧪 Try It Yourself
|
| 32 |
|
|
@@ -253,7 +254,14 @@ Muon represents a significant advancement in neural network optimization by addr
|
|
| 253 |
|
| 254 |
As you build and train your own models, consider Muon for hidden layer optimization, especially during pre-training phases where building new capabilities is the priority.
|
| 255 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 256 |
---
|
|
|
|
| 257 |
👉 If you find this useful, please ⭐️ the repo or cite it in your work — it helps support the project.
|
| 258 |
|
| 259 |
## 📖 Citation
|
|
|
|
| 27 |
8. [Optimization Techniques](#8-optimization-techniques)
|
| 28 |
9. [Lessons Learned](#9-lessons-learned-from-implementing-muon)
|
| 29 |
10. [Concludsion](#10-conclusion)
|
| 30 |
+
10. [Extended Work](#11-extended-work)
|
| 31 |
|
| 32 |
## 🧪 Try It Yourself
|
| 33 |
|
|
|
|
| 254 |
|
| 255 |
As you build and train your own models, consider Muon for hidden layer optimization, especially during pre-training phases where building new capabilities is the priority.
|
| 256 |
|
| 257 |
+
|
| 258 |
+
## 11. Extended Work
|
| 259 |
+
For the distributed (DP × TP) implementation built for CPU/Gloo environments, see:
|
| 260 |
+
|
| 261 |
+
[🧩 The "Muon is Scalable" Blueprint: A Distributed Muon Engineering Breakdown (CPU-Friendly, Tutorial Style)](https://huggingface.co/datasets/bird-of-paradise/muon-distributed)
|
| 262 |
+
|
| 263 |
---
|
| 264 |
+
|
| 265 |
👉 If you find this useful, please ⭐️ the repo or cite it in your work — it helps support the project.
|
| 266 |
|
| 267 |
## 📖 Citation
|