Update README.md
Browse files
README.md
CHANGED
|
@@ -69,6 +69,53 @@ Explore LWM concepts and applications in this compact video series:
|
|
| 69 |
</table>
|
| 70 |
-->
|
| 71 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 72 |
### **How is LWM 1.1 built?**
|
| 73 |
|
| 74 |
LWM 1.1 is a **transformer-based architecture** designed to model **spatial and frequency dependencies** in wireless channel data. It utilizes an enhanced **Masked Channel Modeling (MCM)** pretraining approach, with an increased masking ratio to improve feature learning and generalization. The introduction of **2D patch segmentation** allows the model to jointly process spatial (antenna) and frequency (subcarrier) relationships, providing a more structured representation of the channel. Additionally, **bucket-based batching** is employed to efficiently handle variable-sized inputs without excessive padding, ensuring memory-efficient training and inference. These modifications enable LWM 1.1 to extract meaningful embeddings from a wide range of wireless scenarios, improving its applicability across different system configurations.
|
|
|
|
| 69 |
</table>
|
| 70 |
-->
|
| 71 |
|
| 72 |
+
### 🎥 LWM Tutorial Series
|
| 73 |
+
|
| 74 |
+
Explore LWM concepts and applications in this compact video series:
|
| 75 |
+
|
| 76 |
+
<table>
|
| 77 |
+
<tr>
|
| 78 |
+
<td align="center">
|
| 79 |
+
<a href="https://www.youtube.com/watch?v=3sxJR86EFOo" target="_blank">
|
| 80 |
+
<img src="https://img.youtube.com/vi/3sxJR86EFOo/0.jpg" width="180"/>
|
| 81 |
+
<div style="margin-top:4px;padding:4px 12px;background:#f97316;color:white;border-radius:6px;font-weight:600;">▶ Watch</div>
|
| 82 |
+
</a>
|
| 83 |
+
</td>
|
| 84 |
+
<td align="center">
|
| 85 |
+
<a href="https://www.youtube.com/watch?v=Coqcya9NzFs" target="_blank">
|
| 86 |
+
<img src="https://img.youtube.com/vi/Coqcya9NzFs/0.jpg" width="180"/>
|
| 87 |
+
<div style="margin-top:4px;padding:4px 12px;background:#f97316;color:white;border-radius:6px;font-weight:600;">▶ Watch</div>
|
| 88 |
+
</a>
|
| 89 |
+
</td>
|
| 90 |
+
<td align="center">
|
| 91 |
+
<a href="https://www.youtube.com/watch?v=e9KvAXMUuQg" target="_blank">
|
| 92 |
+
<img src="https://img.youtube.com/vi/e9KvAXMUuQg/0.jpg" width="180"/>
|
| 93 |
+
<div style="margin-top:4px;padding:4px 12px;background:#f97316;color:white;border-radius:6px;font-weight:600;">▶ Watch</div>
|
| 94 |
+
</a>
|
| 95 |
+
</td>
|
| 96 |
+
</tr>
|
| 97 |
+
<tr>
|
| 98 |
+
<td align="center">
|
| 99 |
+
<a href="https://www.youtube.com/watch?v=ZB5WVvo6q6U" target="_blank">
|
| 100 |
+
<img src="https://img.youtube.com/vi/ZB5WVvo6q6U/0.jpg" width="180"/>
|
| 101 |
+
<div style="margin-top:4px;padding:4px 12px;background:#f97316;color:white;border-radius:6px;font-weight:600;">▶ Watch</div>
|
| 102 |
+
</a>
|
| 103 |
+
</td>
|
| 104 |
+
<td align="center">
|
| 105 |
+
<a href="https://www.youtube.com/watch?v=5oNnJjos0mo" target="_blank">
|
| 106 |
+
<img src="https://img.youtube.com/vi/5oNnJjos0mo/0.jpg" width="180"/>
|
| 107 |
+
<div style="margin-top:4px;padding:4px 12px;background:#f97316;color:white;border-radius:6px;font-weight:600;">▶ Watch</div>
|
| 108 |
+
</a>
|
| 109 |
+
</td>
|
| 110 |
+
<td align="center">
|
| 111 |
+
<a href="https://www.youtube.com/watch?v=_RObWck3MMw" target="_blank">
|
| 112 |
+
<img src="https://img.youtube.com/vi/_RObWck3MMw/0.jpg" width="180"/>
|
| 113 |
+
<div style="margin-top:4px;padding:4px 12px;background:#f97316;color:white;border-radius:6px;font-weight:600;">▶ Watch</div>
|
| 114 |
+
</a>
|
| 115 |
+
</td>
|
| 116 |
+
</tr>
|
| 117 |
+
</table>
|
| 118 |
+
|
| 119 |
### **How is LWM 1.1 built?**
|
| 120 |
|
| 121 |
LWM 1.1 is a **transformer-based architecture** designed to model **spatial and frequency dependencies** in wireless channel data. It utilizes an enhanced **Masked Channel Modeling (MCM)** pretraining approach, with an increased masking ratio to improve feature learning and generalization. The introduction of **2D patch segmentation** allows the model to jointly process spatial (antenna) and frequency (subcarrier) relationships, providing a more structured representation of the channel. Additionally, **bucket-based batching** is employed to efficiently handle variable-sized inputs without excessive padding, ensuring memory-efficient training and inference. These modifications enable LWM 1.1 to extract meaningful embeddings from a wide range of wireless scenarios, improving its applicability across different system configurations.
|