Chinese contentAWS 团队招聘 2027 年夏季实习生，覆盖预训练、中期训练与后期训练
AWS 模型训练团队在招 2027 年暑期实习生，预训练、中期训练、对齐都能接触，岗位覆盖全流程比较少见。想做大模型训练方向的学生可以私信联系，提前两年开放申请，值得留意。
#large model training
14 stories · 1 in 7 days · 2 in 30 days
Stories, products and papers tagged "large model training", newest first.
Nov 11
Oct 18
Chinese content我国首个全国产十万卡 AI 超集群真机亮相，曙光 8000 走超算智算融合路线
中科曙光的曙光 8000 首次公开展出真机，十万卡互联，一套底座同时跑科学计算和大模型，跟单独建超算或智算中心的思路不一样。
IT之家Source
Oct 1
AllenAI launches Olmo-core 3
This open-source infrastructure democratizes access to MoE model training, potentially accelerating research beyond well-funded labs. Developers and researchers can now experiment with MoE architectures without massive resource requirements, though they'll still need significant GPU clusters for full-scale training.
Sep 22
Sep 18
Huawei Lingqu Ascend 950 Smart Computing Cluster Cloud Service
For users needing high-performance AI training, the Ascend 950 cluster offers a new, powerful option from a major hardware vendor, potentially competing with other leading solutions.
IT之家Source
Accelerating Sharded Data Parallelism with Federated Learning
This research provides a concrete method for improving the efficiency of training large AI models, which is relevant for anyone working with or studying large language models.
Sep 15
Chinese Internet Corpus 4.0 released with 120GB data
This is a significant release of a large-scale, high-quality Chinese language dataset, which is crucial for training and improving the performance of large language models in the Chinese language.
IT之家Source
Sep 7
Jun 24
Products20:00
Chinese content中科曙光ParaStor F9000包揽IO500双榜冠军，刷新全球存储性能纪录中科曙光的ParaStor F9000在IO500上拿了两个第一，存储性能全球领先，还搞定了414.7亿原子模拟，比之前的290亿原子纪录高出不少。
IT之家Source
Jun 18
Jun 2
Chinese content顾全全离开字节 Seed：AI 药物发现与 LLM Scaling 双线突破
顾全全的离开是字节 Seed 的重大人才变动，关注 AI 药物发现或大模型训练的从业者，值得了解他过去三年的核心成果——SeedFold 和 SeedProteo 直接对标 AlphaFold 3 等顶级模型，做生物计算或 LLM 预训练的团队建议点开看看。
May 21
May 19
Chinese contentCursor 与 SpaceXAI 合作训练更大模型，算力提升 10 倍
AI 编程工具正在从“辅助”走向“自主”，Cursor 这次联手 SpaceXAI 把算力堆到 10 倍，做 AI 编程或模型训练的开发者值得关注——这可能是下一代编程助手的雏形。
May 18
Chinese contentAsteria：运行时编排二阶优化，让大模型训练更高效
Asteria 让二阶优化从理论走向实用，做大模型训练的团队可以大幅降低显存压力，同时保持收敛速度优势，值得关注其开源进展。