关于GCU、沐曦GPGPU、MLU、0卡V100资源4月7日恢复上架的公告>>> 关于共建具身智能开源数据集的倡议>>> 关于云脑任务中统一路径访问方式的公告>>> 关于将启智集群GPU资源迁移至智算集群的公告>>>
deng
Public Activity Repositories Datasets Starred Repositories
83
 Following
18
 Followers
4
 Achievement Badge
1
Loading Heatmap…

deng synced commits to r1.0 at deng/mindformers from mirror

  • d1ad5dd7e8 !2764 【qwen】bugfix: mslite下Qwen-14B运行batch_size=10时报告内存不足 Merge pull request !2764 from Yang Guilong/r1.0
  • 9055b9538a !2681 【r1.0】release note 1.0.2 Merge pull request !2681 from Xinrui Chen/r1.0-release
  • 3e16edb000 release note 1.0.2
  • 2688c88795 [qwen] lite.ini: don't enable ge.externalWeight by default
  • Compare 4 commits »

5 hours ago

deng synced commits to kbk-infer at deng/mindformers from mirror

5 hours ago

deng synced commits to dynamic_parallel at deng/mindformers from mirror

  • 941a5848b0 !2810 【BUGFIX】修复AdamWeightDecayZeRO2优化器register_hook失效问题 Merge pull request !2810 from kairui_kou/dynamic_parallel_opt_bugfix
  • 94b4618993 !2765 梯度累加 Merge pull request !2765 from 张立夫-杭州电子科技大学云技术研究中心(曾艳)&中央研究院(杨宇)/dynamic_parallel
  • b30333b352 梯度累加
  • 37218e048b !2773 add ring attention Merge pull request !2773 from xiaosh/add-ring-attention
  • 5e661508c6 bugfix for adamwzero2
  • Compare 8 commits »

5 hours ago

deng synced commits to dev at deng/mindformers from mirror

  • d2f0f78d8f !2793 【dev】run_qwen.py 支持--predict_data传入多个值，支持batch推理 Merge pull request !2793 from Yang Guilong/dev
  • 7495ce0a2a !2811 【bugfix】【dev】Qwen7B/14B 单卡和多卡推理seq_length设置成1k Merge pull request !2811 from liyang/dev
  • 781e3d4bb2 !2796 llama3 readme更新,使用msrun启动 Merge pull request !2796 from niyuxin94520/dev
  • 05edd0eb4d 【bugfix】【dev】Qwen7B/14B 单卡和多卡推理seq_length设置成1k
  • 345f8e7d01 !2805 [qwen] 修复kbk推理 Merge pull request !2805 from Yang Guilong/kbk-infer
  • Compare 38 commits »

5 hours ago

deng synced commits to master at deng/mindscience from mirror

5 hours ago

deng synced commits to r1.1.rc1 at deng/mindformers from mirror

2 days ago

deng synced commits to r1.0 at deng/mindformers from mirror

  • 45d1fa53db !2762 【r1.0】修复glm2_6b_ptuning2在增量推理时kvcache的序列维度没有扩充prefix的问题 Merge pull request !2762 from Xinrui Chen/r1.0-glm2config
  • 68728434c7 修复glm2_6b_ptuning2在增量推理时kvcache和attention mask的序列维度没有扩充prefix的问题
  • aca5be9e2f !2763 修复在910B3环境单机微调失败问题 Merge pull request !2763 from 冯浩/r1.0-bugfix
  • d1d8908f00 bugfix
  • 1f547d51e3 !2749 【r1.0】修复权重自动转换时src_strategy未合并问题 Merge pull request !2749 from 森镇/src_strategy_unmerge
  • Compare 8 commits »

2 days ago

deng synced commits to dev at deng/mindformers from mirror

  • dadb2c071e !2745 yi-34b适配kbk-infer训推一体 Merge pull request !2745 from 吴昊天/yi-34b-dev
  • fe2d63b35c !2759 glm32k适配kbk-infer训推一体 Merge pull request !2759 from 吴昊天/glm32k-dev
  • e46f212628 yi-34b适配kbk-infer训推一体
  • 7518c0691e glm32k适配kbk-infer训推一体
  • 33bce4e248 !2767 llama3 修复special tokenizer Merge pull request !2767 from xwkgch/dev
  • Compare 20 commits »

2 days ago

deng synced commits to 1.0.a at deng/mindformers from mirror

2 days ago

deng synced commits to r0.6 at deng/mindscience from mirror

  • e58626d936 !1901 [MindChemistry] matformer添加推理脚本 Merge pull request !1901 from jian981105/matformer_4_16
  • 4345013b7b !1900 [MindChemistry] deephe3nn和规范整改，添加推理脚本 Merge pull request !1900 from jian981105/deephe3nn_4_16
  • 34c3bc8bf4 reformat. reformat deepe3nn set device id from arguement
  • 6170d70ce6 add predict file
  • Compare 4 commits »

2 days ago

deng synced commits to r1.1.rc1 at deng/mindformers from mirror

  • 8c53301456 !2751 【r1.1】修复权重自动转换时src_strategy未合并问题 Merge pull request !2751 from 森镇/src_strategy_unmerge_r1.1
  • d1883b14cf !2730 【r1.1.rc1】wikitext-2下载链接修复 Merge pull request !2730 from Xinrui Chen/r1.1.rc1-wikitext
  • 1f2ec48e7d !2675 [r1.1][qwen] 适配 bf16训推 Merge pull request !2675 from Yang Guilong/r1.1.rc1
  • 365658e6d8 [qwen] 适配bf16训推
  • 2035ca861f 修复src_strategy不合并问题
  • Compare 6 commits »

5 days ago

deng synced commits to r1.0 at deng/mindformers from mirror

  • 70d88b8c5d !2739 【r1.0】baichuan2-13B双机训练配置修改 Merge pull request !2739 from 森镇/ascend_config
  • 000a5cff94 !2734 【r1.0】glm32k网络全参微调失败修复 Merge pull request !2734 from wuzhiyuan1996/r1.0
  • d73a472701 !2735 【r1.0】glm32k评测部分文档完善 Merge pull request !2735 from Xinrui Chen/r1.0-glm32k
  • 0ae704e63b 修改baichuan2双机配置
  • d502c199a8 glm32k评测部分文档完善
  • Compare 12 commits »

5 days ago

deng synced commits to feature-dev-qwenvl at deng/mindformers from mirror

  • 47f74073f0 !2738 【bugfix】【feature-dev-qwenvl】修复qwenvl模型的若干问题 Merge pull request !2738 from hsshuai/bugfix/feature-dev-qwenvl
  • f61a2bb2fe fix qwenvl issues, including image padding, infer batch size, weights shape
  • Compare 2 commits »

5 days ago

deng synced commits to dynamic_parallel at deng/mindformers from mirror

5 days ago

deng synced commits to dev at deng/mindformers from mirror

  • b85be44d09 !2747 glm2 bf16适配修改 Merge pull request !2747 from nashturing/glm2
  • 85f47cebc2 !2732 [dev][qwen] 适配 bf16训推 Merge pull request !2732 from Yang Guilong/dev
  • e7598000c6 [qwen] 适配bf16训推
  • 0cc7e4f099 update mindformers/models/glm2/glm2.py. Signed-off-by: nashturing <jinrencao@huawei.com>
  • 5403adb473 !2741 【bugfix】【dev】Qwen-7B/14B 启动脚本引入optim Merge pull request !2741 from liyang/dev
  • Compare 16 commits »

5 days ago

deng synced commits to r0.6 at deng/mindscience from mirror

5 days ago

deng synced commits to master at deng/mindscience from mirror

5 days ago

deng synced commits to r1.1.rc1 at deng/mindformers from mirror

  • eb23abcd7e !2717 【r1.1.rc1】使用choice_fun替换specify_prefix Merge pull request !2717 from 森镇/choice_fun
  • b64cfbda8b 使用choice_fun替换specify_prefix
  • 3e75ce9000 !2715 【r1.1.rc1】增加check_path_include_total_ckpt对load_checkpoint参数为None校验 Merge pull request !2715 from huanglei/cherry-pick-1713174866
  • 2cf8ff027b fixed 4ce34b3 from https://gitee.com/huanglei_Sorry/mindformers/pulls/2704 fix only_save_strategy bug
  • Compare 4 commits »

1 week ago

deng synced commits to r1.0 at deng/mindformers from mirror

  • 937d3e5109 !2641 [llama]修复RMSNorm使用大算子时，layernorm_compute_type参数不生效的问题 Merge pull request !2641 from xwkgch/cherry-pick-1712484278
  • ffaff8cf4c !2721 glm2教程文档 Merge pull request !2721 from nashturing/r1.0
  • f530efcf23 !2707 镜像链接修改 Merge pull request !2707 from nashturing/r1.0
  • 5e7a9dcacf !2719 【r1.0】修复glm2config中model_name字段冲突 Merge pull request !2719 from Xinrui Chen/r1.0-glm2config
  • cd613f135c glm2教程文档
  • Compare 16 commits »

1 week ago

deng synced commits to kbk-infer at deng/mindformers from mirror

  • 0f8fc32fda !2724 bugfix batch_size changed Merge pull request !2724 from wangpingan/kbk3
  • 91762da687 bugfix batch_size changed
  • 3d27f6ed75 !2709 Adapt MindIE Merge pull request !2709 from moran/kbk-infer
  • cc09037bd1 Adapt ModelRunner api for MindIE
  • ece6c2244f !2703 add kbk context and fix multi batch bug Merge pull request !2703 from tan-wei-cheng/develop-twc-dev
  • Compare 6 commits »

1 week ago

Please read the following content carefully:

Dear OpenI User

Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.

For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》

Agree and continue
Disagree, exit