{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":336274588,"defaultBranch":"develop","name":"PaddleNLP","ownerLogin":"PaddlePaddle","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2021-02-05T13:07:42.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/23534030?v=4","public":true,"private":false,"isOrgOwned":true},"refInfo":{"name":"","listCacheKey":"v0:1717039900.0","currentOid":""},"activityList":{"items":[{"before":"ac7145dae47899d096795f9a0cfda9e88726f80c","after":"e71540bfc6ceb4883c99cd89156b23b3b4922dae","ref":"refs/heads/develop","pushedAt":"2024-05-31T08:10:16.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"Add Sharding V1 broadcast and allgather overlap optimize (#8499)\n\n* update\r\n\r\n* update is_casual_mask to use_casual_mask\r\n\r\n* update by environment\r\n\r\n* add constraint\r\n\r\n* add pretrain and finetune enviroment\r\n\r\n* update\r\n\r\n* update\r\n\r\n* Update finetune_generation.py\r\n\r\nupdate use_casual_mask env\r\n\r\n* update\r\n\r\n* lint code\r\n\r\n---------\r\n\r\nCo-authored-by: zhengzhonghui \r\nCo-authored-by: lizhiyu <1528794076@qq.com>","shortMessageHtmlLink":"Add Sharding V1 broadcast and allgather overlap optimize (#8499)"}},{"before":"60c661546b985c318afd0e85139a4b39ba3a3a0e","after":"ac7145dae47899d096795f9a0cfda9e88726f80c","ref":"refs/heads/develop","pushedAt":"2024-05-31T06:43:36.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"[fea] moe support (#8498)\n\nCo-authored-by: kebo01 ","shortMessageHtmlLink":"[fea] moe support (#8498)"}},{"before":"92aa9f1767a88c8f92dfd9a1159babd7d24e0ee3","after":"60c661546b985c318afd0e85139a4b39ba3a3a0e","ref":"refs/heads/develop","pushedAt":"2024-05-31T03:05:26.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"Update merge_lora_params.py (#8514)\n\n* Update merge_lora_params.py\r\n\r\n* Update merge_lora_params.py\r\n\r\n* Update merge_lora_params.py","shortMessageHtmlLink":"Update merge_lora_params.py (#8514)"}},{"before":"a0b516d073df79310e53c1c76a8c37662d43d579","after":null,"ref":"refs/heads/ZHUI-patch-1","pushedAt":"2024-05-30T03:31:40.000Z","pushType":"branch_deletion","commitsCount":0,"pusher":{"login":"ZHUI","name":"Zhong Hui","path":"/ZHUI","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16911935?s=80&v=4"}},{"before":"d8adc2b9d12128463fcf31fce2721cae2f600037","after":"92aa9f1767a88c8f92dfd9a1159babd7d24e0ee3","ref":"refs/heads/develop","pushedAt":"2024-05-30T03:31:27.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"Update model_utils.py (#8509)","shortMessageHtmlLink":"Update model_utils.py (#8509)"}},{"before":null,"after":"a0b516d073df79310e53c1c76a8c37662d43d579","ref":"refs/heads/ZHUI-patch-1","pushedAt":"2024-05-29T08:52:23.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"ZHUI","name":"Zhong Hui","path":"/ZHUI","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16911935?s=80&v=4"},"commit":{"message":"Update model_utils.py","shortMessageHtmlLink":"Update model_utils.py"}},{"before":"bbf945b64ab611e491c429ef86887cf84f43d3a5","after":"82a71775424043042a8c672cd1e9fc09348fd594","ref":"refs/heads/release/2.8","pushedAt":"2024-05-29T08:50:59.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"ZHUI","name":"Zhong Hui","path":"/ZHUI","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16911935?s=80&v=4"},"commit":{"message":"quick fix os.path.split (#8508)","shortMessageHtmlLink":"quick fix os.path.split (#8508)"}},{"before":"c1cfe631df59d6273f8736843c4142b1805a92d0","after":"d8adc2b9d12128463fcf31fce2721cae2f600037","ref":"refs/heads/develop","pushedAt":"2024-05-29T05:18:34.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"[LLM] disable part of MC2 in lora (#8505)","shortMessageHtmlLink":"[LLM] disable part of MC2 in lora (#8505)"}},{"before":"85ba5735c38750fcb00847080a3eb2b23d04b825","after":"c1cfe631df59d6273f8736843c4142b1805a92d0","ref":"refs/heads/develop","pushedAt":"2024-05-28T10:52:56.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"modify block_multihead_attention api (#8456)\n\n* modify block_multihead_attention api\r\n\r\n* add param to blha\r\n\r\n* modify fused_transformer_layers\r\n\r\n* fix bug","shortMessageHtmlLink":"modify block_multihead_attention api (#8456)"}},{"before":"d4edd19194aef7ce8f9fee936395bc279b2fd606","after":"85ba5735c38750fcb00847080a3eb2b23d04b825","ref":"refs/heads/develop","pushedAt":"2024-05-28T02:12:40.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"[AutoParallel] Support qwen for auto_parallel (#8312)\n\n* run 4096\r\n\r\n* support 3d semi auto run of qwen model\r\n\r\n* [AutoParallel] Add qwen auto parallel.\r\n\r\n* Delete useless files.\r\n\r\n* Fix code style problem.\r\n\r\n* Fix problems.\r\n\r\n* Polish code.\r\n\r\n* Add CI-cases.\r\n\r\n* Add bf16 ci testcase.\r\n\r\n* Fix some problems.\r\n\r\n---------\r\n\r\nCo-authored-by: pangengzheng \r\nCo-authored-by: pangengzheng ","shortMessageHtmlLink":"[AutoParallel] Support qwen for auto_parallel (#8312)"}},{"before":"773497e2e63b7c434502bda582b4ba9479634e6e","after":"d4edd19194aef7ce8f9fee936395bc279b2fd606","ref":"refs/heads/develop","pushedAt":"2024-05-27T06:30:23.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"Fix rng_state in llm models (#8396)","shortMessageHtmlLink":"Fix rng_state in llm models (#8396)"}},{"before":"d8e1a6b2b6290a2f25b45ca830dcf0a642a69ad7","after":"773497e2e63b7c434502bda582b4ba9479634e6e","ref":"refs/heads/develop","pushedAt":"2024-05-27T04:51:20.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"quick fix from pretrained. (#8486)","shortMessageHtmlLink":"quick fix from pretrained. (#8486)"}},{"before":"a90f16353d32b9ed5f2a2418a449406ad7fc7612","after":"d8e1a6b2b6290a2f25b45ca830dcf0a642a69ad7","ref":"refs/heads/develop","pushedAt":"2024-05-27T02:54:28.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"【AutoParallel】Add split_backward for vpp (#8479)\n\n* add split_backward for vpp\r\n\r\n* polish\r\n\r\n* add test cast\r\n\r\n* polish\r\n\r\n* update test case\r\n\r\n* change the config\r\n\r\n* polish","shortMessageHtmlLink":"【AutoParallel】Add split_backward for vpp (#8479)"}},{"before":"0cd8fe7ee58180522168411606152dfe4ddda681","after":"a90f16353d32b9ed5f2a2418a449406ad7fc7612","ref":"refs/heads/develop","pushedAt":"2024-05-24T08:51:23.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"fix run_benchmark for llama2_70b in auto_parallel (#8484)\n\n* remove tsinghua pypi\r\n\r\n* modify gpt dateset addr for benchmark\r\n\r\n* fix run_benchmark for llama2_70b in auto_parallel","shortMessageHtmlLink":"fix run_benchmark for llama2_70b in auto_parallel (#8484)"}},{"before":"7aaa788069fc07ca109d5ac77e14603ed295f02d","after":"0cd8fe7ee58180522168411606152dfe4ddda681","ref":"refs/heads/develop","pushedAt":"2024-05-24T07:28:09.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"Revert \"Support Sharding Overlap (#8473)\" (#8491)\n\nThis reverts commit 7aaa788069fc07ca109d5ac77e14603ed295f02d.","shortMessageHtmlLink":"Revert \"Support Sharding Overlap (#8473)\" (#8491)"}},{"before":"8879f79f9857dc7831403064631ae32b0a0def23","after":"bbf945b64ab611e491c429ef86887cf84f43d3a5","ref":"refs/heads/release/2.8","pushedAt":"2024-05-24T05:43:46.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"Release/2.8 (#8437)\n\n* [XPU] llama add xpu support (#8282)\r\n\r\n* [XPU] llama add xpu support\r\n\r\n* fix\r\n\r\n* use try import\r\n\r\n* fix\r\n\r\n* refine\r\n\r\n* refine\r\n\r\n* refine\r\n\r\n* refine\r\n\r\n* update (#8399)\r\n\r\n* [LLM] Support fuse attention q, k, v weights (#8202)\r\n\r\n1. add use-interface & fuse action\r\n\r\n1.1. modify 1., code order\r\n\r\n2. switch to name_mapping\r\n\r\n3. solve tp branch\r\n\r\n3.2 follow hui, handel qkv separately\r\n\r\n3.3 handle pdparams\r\n\r\n3.4 from torch\r\n\r\n3.5 abandon low_cpu_mem_usage\r\n\r\n3.6 solve shard branch\r\n\r\n* 3.6.1 solve shard branch after rebase develop\r\n\r\n* code clean\r\n\r\n* remove debug comment\r\n\r\n* Redefine fuse and split functions\r\n\r\n* Redefine fuse and split functions\r\n\r\n* comment and fix\r\n\r\n* update method\r\n\r\n* update QKV fuse and split\r\n\r\n* support fuse weights in multi-files\r\n\r\n* add precision compare\r\n\r\n* simplify function call\r\n\r\n* support use_fast_ffn\r\n\r\n* clean modeling and configuration\r\n\r\n* add test for gpt and opt\r\n\r\n* fix tp_actions get\r\n\r\n* add fast_ffn test\r\n\r\n* add Qwen2Moe\r\n\r\n* Revert \"add Qwen2Moe\"\r\n\r\nThis reverts commit 113b8838a7c53f1d131928c30bf1071dfa583445.\r\n\r\n* add test for split\r\n\r\n* update doc\r\n\r\n* update filter_dict_keys\r\n\r\n---------\r\n\r\nCo-authored-by: Zii \r\n\r\n* [LLM] Fix fuse or split with same key (#8378)\r\n\r\n* fix fuse or split with same key\r\n\r\n* fix\r\n\r\n* fix eps\r\n\r\n* update format\r\n\r\n* [LLM] add decay steps option for finetuning (#8251)\r\n\r\n* [LLM] add memory stats to logger of trainer (#8269)\r\n\r\n* [Distributed] fix lora (#8325)\r\n\r\n* [LLM] fix lora target modules on llama (#8372)\r\n\r\n* [Distributed] metric calculation supports tp logits (#8370)\r\n\r\n* Update model_utils.py\r\n\r\n* Update model_utils.py\r\n\r\n* Update model_utils.py\r\n\r\n---------\r\n\r\nCo-authored-by: Jianbang Yang \r\nCo-authored-by: DrownFish19 \r\nCo-authored-by: Zii \r\nCo-authored-by: Tian <121000916+SylarTiaNII@users.noreply.github.com>","shortMessageHtmlLink":"Release/2.8 (#8437)"}},{"before":"c6f4159d9c972732e3798903a7f75b33eee7ac25","after":"7aaa788069fc07ca109d5ac77e14603ed295f02d","ref":"refs/heads/develop","pushedAt":"2024-05-24T02:17:21.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"Support Sharding Overlap (#8473)\n\n* update\r\n\r\n* update is_casual_mask to use_casual_mask\r\n\r\n* update by environment","shortMessageHtmlLink":"Support Sharding Overlap (#8473)"}},{"before":"7a24bccfd15348e818036c56335fccf984fd95d5","after":"8879f79f9857dc7831403064631ae32b0a0def23","ref":"refs/heads/release/2.8","pushedAt":"2024-05-23T13:09:14.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"ZHUI","name":"Zhong Hui","path":"/ZHUI","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16911935?s=80&v=4"},"commit":{"message":"quick fix from pretrained. (#8487)","shortMessageHtmlLink":"quick fix from pretrained. (#8487)"}},{"before":"a315046f4eecb0f5b8a776db1c73e0a9ff7cdc6a","after":"6b96825447ee5387d849812b13e8aa5ae23b5fa1","ref":"refs/heads/paddlenlp-dev-fleetv","pushedAt":"2024-05-23T09:09:36.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"sneaxiy","name":null,"path":"/sneaxiy","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/32832641?s=80&v=4"},"commit":{"message":"fix bug of sharding format (#8483)","shortMessageHtmlLink":"fix bug of sharding format (#8483)"}},{"before":"621118e4870bcce20b942329ba7a7ca7d3c94bc6","after":"c6f4159d9c972732e3798903a7f75b33eee7ac25","ref":"refs/heads/develop","pushedAt":"2024-05-23T03:26:21.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"[Ops] RoPE kernel support theta input (#8440)\n\n* [RoPE] feat: support theta\r\n\r\n* Update modeling.py\r\n\r\n* [RoPE] feat: support theta for qwen","shortMessageHtmlLink":"[Ops] RoPE kernel support theta input (#8440)"}},{"before":"70bffa8df3986d45bba0e4358adf4c4b92e04057","after":"621118e4870bcce20b942329ba7a7ca7d3c94bc6","ref":"refs/heads/develop","pushedAt":"2024-05-22T12:44:25.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"fix rotary_emb for llama (#8470)","shortMessageHtmlLink":"fix rotary_emb for llama (#8470)"}},{"before":"328d0a032714a02fa9bdd8709888f25be5e985f0","after":"a315046f4eecb0f5b8a776db1c73e0a9ff7cdc6a","ref":"refs/heads/paddlenlp-dev-fleetv","pushedAt":"2024-05-22T08:50:07.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"sneaxiy","name":null,"path":"/sneaxiy","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/32832641?s=80&v=4"},"commit":{"message":"Add release grad for SD (#8478)","shortMessageHtmlLink":"Add release grad for SD (#8478)"}},{"before":"87e4c4fa876a9eaa76d22860a6db3a7832d1322b","after":"70bffa8df3986d45bba0e4358adf4c4b92e04057","ref":"refs/heads/develop","pushedAt":"2024-05-22T05:58:25.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"[LLM] fix bug when loss is None in llama modeling.py (#8459)","shortMessageHtmlLink":"[LLM] fix bug when loss is None in llama modeling.py (#8459)"}},{"before":"08898bf1e0429db3da6d0b3e8a95e8b7d8c817d7","after":"7a24bccfd15348e818036c56335fccf984fd95d5","ref":"refs/heads/release/2.8","pushedAt":"2024-05-21T03:24:11.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"ZHUI","name":"Zhong Hui","path":"/ZHUI","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16911935?s=80&v=4"},"commit":{"message":"Cherry pick type promotion fix. (#8463)","shortMessageHtmlLink":"Cherry pick type promotion fix. (#8463)"}},{"before":"9064078e8d3c5fed6213a690a07e38c7a75d794f","after":"87e4c4fa876a9eaa76d22860a6db3a7832d1322b","ref":"refs/heads/develop","pushedAt":"2024-05-20T14:15:10.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"wawltor","name":null,"path":"/wawltor","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16698950?s=80&v=4"},"commit":{"message":"[NPU] support npu llama2-13B export & inference (#8442)\n\n* [NPU] support npu llama2-13B export & inference\r\n\r\n* move csrc_npu to csrc/npu","shortMessageHtmlLink":"[NPU] support npu llama2-13B export & inference (#8442)"}},{"before":"b36b6a01a6dd65ea45f4e8db0177a61c80ae410e","after":"9064078e8d3c5fed6213a690a07e38c7a75d794f","ref":"refs/heads/develop","pushedAt":"2024-05-20T11:29:07.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"ZHUI","name":"Zhong Hui","path":"/ZHUI","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16911935?s=80&v=4"},"commit":{"message":"Fix CI tests. (#8471)","shortMessageHtmlLink":"Fix CI tests. (#8471)"}},{"before":"fc860a3289804fbaf197d12c6d858d0d79e741af","after":"08898bf1e0429db3da6d0b3e8a95e8b7d8c817d7","ref":"refs/heads/release/2.8","pushedAt":"2024-05-20T05:05:30.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"ZHUI","name":"Zhong Hui","path":"/ZHUI","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/16911935?s=80&v=4"},"commit":{"message":"Cherry-Pick fast_safe_open (#8458)\n\n* [Performance] Optimize unified checkpoint save/load speed. (#8204)\r\n\r\n* opt unified checkpoint save/load speed.","shortMessageHtmlLink":"Cherry-Pick fast_safe_open (#8458)"}},{"before":"d84cfe4e1515e790b67d04a5e01132e4679eac95","after":"328d0a032714a02fa9bdd8709888f25be5e985f0","ref":"refs/heads/paddlenlp-dev-fleetv","pushedAt":"2024-05-18T16:24:32.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"sneaxiy","name":null,"path":"/sneaxiy","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/32832641?s=80&v=4"},"commit":{"message":"fix bug in sync_param_name (#8465)","shortMessageHtmlLink":"fix bug in sync_param_name (#8465)"}},{"before":"14f66ddf60ec6fa5d9c59f71a24c8fdc72622aa6","after":"d84cfe4e1515e790b67d04a5e01132e4679eac95","ref":"refs/heads/paddlenlp-dev-fleetv","pushedAt":"2024-05-18T10:35:02.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"sneaxiy","name":null,"path":"/sneaxiy","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/32832641?s=80&v=4"},"commit":{"message":"sync param/grad/moment (#8464)","shortMessageHtmlLink":"sync param/grad/moment (#8464)"}},{"before":"5d4ce560ef3442b26ee544182d6a149aad351ba3","after":"b36b6a01a6dd65ea45f4e8db0177a61c80ae410e","ref":"refs/heads/develop","pushedAt":"2024-05-17T12:56:30.000Z","pushType":"pr_merge","commitsCount":1,"pusher":{"login":"zhiqiu","name":"Leo Chen","path":"/zhiqiu","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/6888866?s=80&v=4"},"commit":{"message":"Add llama2-70b for test_tipc (#8455)","shortMessageHtmlLink":"Add llama2-70b for test_tipc (#8455)"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAAEWMgEEAA","startCursor":null,"endCursor":null}},"title":"Activity · PaddlePaddle/PaddleNLP"}