🐦 Twitter Post Details

Viewing enriched Twitter post

@PyTorch

Normalization methods (LayerNorm/RMSNorm) are foundational in modern deep learning models. We evaluate and improve torch.compile performance for LayerNorm/RMSNorm on NVIDIA H100 and B200 to reach near SOTA performance on a kernel-by-kernel basis, whilst providing automatic fusion capabilities with torch.compile for peak e2e performance. 🔗 Read our latest blog from Shunting Zhang, Paul Zhang, Markus Höhnerbach, Elias Ellison, Jason Ansel, and Natalia Gimelshein: https://t.co/ie5UZag3qx Today at PyTorch Conference EU: Lightning Talk: Faster Than SOTA Kernels in Torch.compile With Subgraph Fusions and Custom Op Autotuning - Elias Ellison & Paul Zhang, Meta, 15:40 - 15:50 #PyTorch #torchcompile #OpenSourceAI #PyTorchCon

Media 1

📊 Media Metadata

{
  "media": [
    {
      "type": "photo",
      "url": "https://crmoxkoizveukayfjuyo.supabase.co/storage/v1/object/public/media/posts/2041788204471251007/media_0.jpg",
      "filename": "media_0.jpg"
    }
  ],
  "processed_at": "2026-04-08T08:07:38.000480",
  "pipeline_version": "2.0"
}

🔧 Raw API Response

{
  "type": "tweet",
  "id": "2041788204471251007",
  "url": "https://x.com/PyTorch/status/2041788204471251007",
  "twitterUrl": "https://twitter.com/PyTorch/status/2041788204471251007",
  "text": "Normalization methods (LayerNorm/RMSNorm) are foundational in modern deep learning models. We evaluate and improve torch.compile performance for LayerNorm/RMSNorm on NVIDIA H100 and B200 to reach near SOTA performance on a kernel-by-kernel basis, whilst providing automatic fusion capabilities with torch.compile for peak e2e performance. \n\n🔗 Read our latest blog from Shunting Zhang, Paul Zhang, Markus Höhnerbach, Elias Ellison, Jason Ansel, and Natalia Gimelshein: https://t.co/ie5UZag3qx\n\nToday at PyTorch Conference EU: Lightning Talk: Faster Than SOTA Kernels in Torch.compile With Subgraph Fusions and Custom Op Autotuning - Elias Ellison & Paul Zhang, Meta, 15:40 - 15:50\n\n#PyTorch #torchcompile #OpenSourceAI #PyTorchCon",
  "source": "Twitter for iPhone",
  "retweetCount": 0,
  "replyCount": 0,
  "likeCount": 3,
  "quoteCount": 0,
  "viewCount": 295,
  "createdAt": "Wed Apr 08 08:00:17 +0000 2026",
  "lang": "en",
  "bookmarkCount": 3,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "2041788204471251007",
  "displayTextRange": [
    0,
    280
  ],
  "inReplyToUserId": null,
  "inReplyToUsername": null,
  "author": {
    "type": "user",
    "userName": "PyTorch",
    "url": "https://x.com/PyTorch",
    "twitterUrl": "https://twitter.com/PyTorch",
    "id": "776585502606721024",
    "name": "PyTorch",
    "isVerified": false,
    "isBlueVerified": true,
    "verifiedType": null,
    "profilePicture": "https://pbs.twimg.com/profile_images/1813965160702451712/yXV1vRhr_normal.jpg",
    "coverPicture": "https://pbs.twimg.com/profile_banners/776585502606721024/1761575044",
    "description": "",
    "location": "",
    "followers": 483957,
    "following": 83,
    "status": "",
    "canDm": false,
    "canMediaTag": true,
    "createdAt": "Fri Sep 16 00:56:26 +0000 2016",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 870,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 1418,
    "statusesCount": 3180,
    "withheldInCountries": [],
    "affiliatesHighlightedLabel": {},
    "possiblySensitive": false,
    "pinnedTweetIds": [],
    "profile_bio": {
      "description": "Tensors and neural networks in Python with strong hardware acceleration. PyTorch is an open source project at the Linux Foundation. #PyTorchFoundation",
      "entities": {
        "description": {
          "hashtags": [
            {
              "indices": [
                132,
                150
              ],
              "text": "PyTorchFoundation"
            }
          ],
          "symbols": [],
          "urls": [],
          "user_mentions": []
        },
        "url": {
          "urls": [
            {
              "display_url": "pytorch.org",
              "expanded_url": "http://pytorch.org",
              "indices": [
                0,
                23
              ],
              "url": "https://t.co/6SwTBhUwTJ"
            }
          ]
        }
      }
    },
    "isAutomated": false,
    "automatedBy": null
  },
  "extendedEntities": {
    "media": [
      {
        "display_url": "pic.twitter.com/HQy4mXjIb5",
        "expanded_url": "https://twitter.com/PyTorch/status/2041788204471251007/photo/1",
        "ext_media_availability": {
          "status": "Available"
        },
        "features": {
          "large": {
            "faces": [
              {
                "h": 65,
                "w": 65,
                "x": 1222,
                "y": 420
              }
            ]
          },
          "orig": {
            "faces": [
              {
                "h": 65,
                "w": 65,
                "x": 1222,
                "y": 420
              }
            ]
          }
        },
        "id_str": "2041788183461994496",
        "indices": [
          281,
          304
        ],
        "media_key": "3_2041788183461994496",
        "media_results": {
          "id": "QXBpTWVkaWFSZXN1bHRzOgwAAQoAARxV44j116AACgACHFXjjdoXcD8AAA==",
          "result": {
            "__typename": "ApiMedia",
            "id": "QXBpTWVkaWE6DAABCgABHFXjiPXXoAAKAAIcVeON2hdwPwAA",
            "media_key": "3_2041788183461994496"
          }
        },
        "media_url_https": "https://pbs.twimg.com/media/HFXjiPXXoAAFxXT.jpg",
        "original_info": {
          "focus_rects": [
            {
              "h": 1075,
              "w": 1920,
              "x": 0,
              "y": 0
            },
            {
              "h": 1080,
              "w": 1080,
              "x": 0,
              "y": 0
            },
            {
              "h": 1080,
              "w": 947,
              "x": 55,
              "y": 0
            },
            {
              "h": 1080,
              "w": 540,
              "x": 258,
              "y": 0
            },
            {
              "h": 1080,
              "w": 1920,
              "x": 0,
              "y": 0
            }
          ],
          "height": 1080,
          "width": 1920
        },
        "sizes": {
          "large": {
            "h": 1080,
            "w": 1920
          }
        },
        "type": "photo",
        "url": "https://t.co/HQy4mXjIb5"
      }
    ]
  },
  "card": null,
  "place": {},
  "entities": {
    "hashtags": [
      {
        "indices": [
          681,
          689
        ],
        "text": "PyTorch"
      },
      {
        "indices": [
          690,
          703
        ],
        "text": "torchcompile"
      },
      {
        "indices": [
          704,
          717
        ],
        "text": "OpenSourceAI"
      },
      {
        "indices": [
          718,
          729
        ],
        "text": "PyTorchCon"
      }
    ],
    "symbols": [],
    "urls": [
      {
        "display_url": "pytorch.org/blog/sota-norm…",
        "expanded_url": "https://pytorch.org/blog/sota-normalization-performance-with-torch-compile/",
        "indices": [
          468,
          491
        ],
        "url": "https://t.co/ie5UZag3qx"
      }
    ],
    "user_mentions": []
  },
  "quoted_tweet": null,
  "retweeted_tweet": null,
  "isLimitedReply": false,
  "communityInfo": null,
  "article": null
}