{
  "type": "tweet",
  "id": "2024906241567260916",
  "url": "https://x.com/PyTorch/status/2024906241567260916",
  "twitterUrl": "https://twitter.com/PyTorch/status/2024906241567260916",
  "text": "Trying to tune your Expert Parallel (EP) communication for hyperscale mixture-of-experts (MoE) models? \n\nThis post, ‘Optimizing Communication for Mixture-of-Experts Training with Hybrid Expert Parallel’, details an efficient MoE EP communication solution, Hybrid-EP, and its use in the NVIDIA Megatron family of frameworks, on NVIDIA Quantum InfiniBand and NVIDIA Spectrum-X Ethernet platforms. It also dives into the effectiveness of Hybrid-EP in real-world model training. \n\nRead the full post: \nhttps://t.co/4NOFpaiFYz\n\n#PyTorch #OpenSourceAI #AI #Inference #Innovation",
  "source": "Twitter for iPhone",
  "retweetCount": 13,
  "replyCount": 0,
  "likeCount": 121,
  "quoteCount": 3,
  "viewCount": 12411,
  "createdAt": "Fri Feb 20 17:57:23 +0000 2026",
  "lang": "en",
  "bookmarkCount": 85,
  "isReply": false,
  "inReplyToId": null,
  "conversationId": "2024906241567260916",
  "displayTextRange": [
    0,
    278
  ],
  "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": 476861,
    "following": 81,
    "status": "",
    "canDm": false,
    "canMediaTag": true,
    "createdAt": "Fri Sep 16 00:56:26 +0000 2016",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {}
    },
    "fastFollowersCount": 0,
    "favouritesCount": 837,
    "hasCustomTimelines": true,
    "isTranslator": false,
    "mediaCount": 1328,
    "statusesCount": 3065,
    "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": {},
  "card": {
    "binding_values": [
      {
        "key": "photo_image_full_size_large",
        "value": {
          "image_value": {
            "height": 419,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=800x419",
            "width": 800
          }
        }
      },
      {
        "key": "thumbnail_image",
        "value": {
          "image_value": {
            "height": 150,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=280x150",
            "width": 267
          }
        }
      },
      {
        "key": "description",
        "value": {
          "string_value": "In LLM training, Expert Parallel (EP) communication for hyperscale mixture-of-experts (MoE) models is challenging. EP communication is essentially all-to-all, but due to its dynamics and sparseness…"
        }
      },
      {
        "key": "domain",
        "value": {
          "string_value": "developer.nvidia.com"
        }
      },
      {
        "key": "thumbnail_image_large",
        "value": {
          "image_value": {
            "height": 320,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=800x320_1",
            "width": 569
          }
        }
      },
      {
        "key": "summary_photo_image_small",
        "value": {
          "image_value": {
            "height": 202,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=386x202",
            "width": 386
          }
        }
      },
      {
        "key": "thumbnail_image_original",
        "value": {
          "image_value": {
            "height": 675,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=orig",
            "width": 1200
          }
        }
      },
      {
        "key": "photo_image_full_size_small",
        "value": {
          "image_value": {
            "height": 202,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=386x202",
            "width": 386
          }
        }
      },
      {
        "key": "summary_photo_image_large",
        "value": {
          "image_value": {
            "height": 419,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=800x419",
            "width": 800
          }
        }
      },
      {
        "key": "thumbnail_image_small",
        "value": {
          "image_value": {
            "height": 81,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=144x144",
            "width": 144
          }
        }
      },
      {
        "key": "thumbnail_image_x_large",
        "value": {
          "image_value": {
            "height": 675,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=png&name=2048x2048_2_exp",
            "width": 1200
          }
        }
      },
      {
        "key": "photo_image_full_size_original",
        "value": {
          "image_value": {
            "height": 675,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=orig",
            "width": 1200
          }
        }
      },
      {
        "key": "vanity_url",
        "value": {
          "scribe_key": "vanity_url",
          "string_value": "developer.nvidia.com"
        }
      },
      {
        "key": "photo_image_full_size",
        "value": {
          "image_value": {
            "height": 314,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=600x314",
            "width": 600
          }
        }
      },
      {
        "key": "thumbnail_image_color",
        "value": {
          "image_color_value": {
            "palette": [
              {
                "percentage": 96.9,
                "rgb": {
                  "blue": 1,
                  "green": 1,
                  "red": 1
                }
              },
              {
                "percentage": 2.21,
                "rgb": {
                  "blue": 82,
                  "green": 105,
                  "red": 115
                }
              },
              {
                "percentage": 0.41,
                "rgb": {
                  "blue": 200,
                  "green": 208,
                  "red": 208
                }
              },
              {
                "percentage": 0.25,
                "rgb": {
                  "blue": 145,
                  "green": 146,
                  "red": 147
                }
              },
              {
                "percentage": 0.23,
                "rgb": {
                  "blue": 54,
                  "green": 82,
                  "red": 93
                }
              }
            ]
          }
        }
      },
      {
        "key": "title",
        "value": {
          "string_value": "Optimizing Communication for Mixture-of-Experts Training with Hybrid Expert Parallel | NVIDIA..."
        }
      },
      {
        "key": "summary_photo_image_color",
        "value": {
          "image_color_value": {
            "palette": [
              {
                "percentage": 96.9,
                "rgb": {
                  "blue": 1,
                  "green": 1,
                  "red": 1
                }
              },
              {
                "percentage": 2.21,
                "rgb": {
                  "blue": 82,
                  "green": 105,
                  "red": 115
                }
              },
              {
                "percentage": 0.41,
                "rgb": {
                  "blue": 200,
                  "green": 208,
                  "red": 208
                }
              },
              {
                "percentage": 0.25,
                "rgb": {
                  "blue": 145,
                  "green": 146,
                  "red": 147
                }
              },
              {
                "percentage": 0.23,
                "rgb": {
                  "blue": 54,
                  "green": 82,
                  "red": 93
                }
              }
            ]
          }
        }
      },
      {
        "key": "summary_photo_image_x_large",
        "value": {
          "image_value": {
            "height": 675,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=png&name=2048x2048_2_exp",
            "width": 1200
          }
        }
      },
      {
        "key": "summary_photo_image",
        "value": {
          "image_value": {
            "height": 314,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=600x314",
            "width": 600
          }
        }
      },
      {
        "key": "photo_image_full_size_color",
        "value": {
          "image_color_value": {
            "palette": [
              {
                "percentage": 96.9,
                "rgb": {
                  "blue": 1,
                  "green": 1,
                  "red": 1
                }
              },
              {
                "percentage": 2.21,
                "rgb": {
                  "blue": 82,
                  "green": 105,
                  "red": 115
                }
              },
              {
                "percentage": 0.41,
                "rgb": {
                  "blue": 200,
                  "green": 208,
                  "red": 208
                }
              },
              {
                "percentage": 0.25,
                "rgb": {
                  "blue": 145,
                  "green": 146,
                  "red": 147
                }
              },
              {
                "percentage": 0.23,
                "rgb": {
                  "blue": 54,
                  "green": 82,
                  "red": 93
                }
              }
            ]
          }
        }
      },
      {
        "key": "photo_image_full_size_x_large",
        "value": {
          "image_value": {
            "height": 675,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=png&name=2048x2048_2_exp",
            "width": 1200
          }
        }
      },
      {
        "key": "card_url",
        "value": {
          "scribe_key": "card_url",
          "string_value": "https://t.co/4NOFpaiFYz"
        }
      },
      {
        "key": "summary_photo_image_original",
        "value": {
          "image_value": {
            "height": 675,
            "url": "https://pbs.twimg.com/card_img/2027443826743554050/dD4PAM2d?format=jpg&name=orig",
            "width": 1200
          }
        }
      }
    ],
    "card_platform": {
      "platform": {
        "audience": {
          "name": "production"
        },
        "device": {
          "name": "iPhone",
          "version": "13"
        }
      }
    },
    "name": "summary_large_image",
    "url": "https://t.co/4NOFpaiFYz",
    "user_refs_results": []
  },
  "place": {},
  "entities": {
    "hashtags": [
      {
        "indices": [
          523,
          531
        ],
        "text": "PyTorch"
      },
      {
        "indices": [
          532,
          545
        ],
        "text": "OpenSourceAI"
      },
      {
        "indices": [
          546,
          549
        ],
        "text": "AI"
      },
      {
        "indices": [
          550,
          560
        ],
        "text": "Inference"
      },
      {
        "indices": [
          561,
          572
        ],
        "text": "Innovation"
      }
    ],
    "symbols": [],
    "urls": [
      {
        "display_url": "hubs.la/Q0440F0l0",
        "expanded_url": "https://hubs.la/Q0440F0l0",
        "indices": [
          498,
          521
        ],
        "url": "https://t.co/4NOFpaiFYz"
      }
    ],
    "user_mentions": []
  },
  "quoted_tweet": null,
  "retweeted_tweet": null,
  "isLimitedReply": false,
  "article": null
}