Updated on 2024-12-26

Author: lxaw

papers/list.json

@@ -1,4 +1,13 @@
 [
+  {
+    "title": "Remix-DiT: Mixing Diffusion Transformers for Multi-Expert Denoising",
+    "author": "Gongfan Fang et al",
+    "year": "2024",
+    "topic": "dit, diffusion, moe",
+    "venue": "NeurIPS",
+    "description": "This paper introduces a method of mixing diffusion models for multi-expert denoising. Basically, they increase the width of the linear layers by a factor of K, and then modify the forward pass to support it. This allows for K experts that are initialized from the original weights. ",
+    "link": "https://arxiv.org/pdf/2412.05628"
+  },
   {
     "title": "Hymba: A Hybrid-head Architecture for Small Language Models",
     "author": "Xin Dong et al",

papers_read.html

@@ -16,10 +16,10 @@ <h1>Here's where I keep a list of papers I have read.</h1>
         I typically use this to organize papers I found interesting. Please feel free to do whatever you want with it. Note that this is not every single paper I have ever read, just a collection of ones that I remember to put down.
     </p>
     <p id="paperCount">
-        So far, we have read 194 papers. Let's keep it up!
+        So far, we have read 195 papers. Let's keep it up!
     </p> 
     <small id="searchCount">
-        Your search returned 194 papers. Nice! 
+        Your search returned 195 papers. Nice! 
     </small>
 
     <div class="search-inputs">
@@ -46,6 +46,16 @@ <h1>Here's where I keep a list of papers I have read.</h1>
         </thead>
         <tbody>
 
+            <tr>
+                <td>Remix-DiT: Mixing Diffusion Transformers for Multi-Expert Denoising</td>
+                <td>Gongfan Fang et al</td>
+                <td>2024</td>
+                <td>dit, diffusion, moe</td>
+                <td>NeurIPS</td>
+                <td>This paper introduces a method of mixing diffusion models for multi-expert denoising. Basically, they increase the width of the linear layers by a factor of K, and then modify the forward pass to support it. This allows for K experts that are initialized from the original weights. </td>
+                <td><a href="https://arxiv.org/pdf/2412.05628" target="_blank">Link</a></td>
+            </tr>
+        
             <tr>
                 <td>Hymba: A Hybrid-head Architecture for Small Language Models</td>
                 <td>Xin Dong et al</td>