[ { "id": "b933a0edb8cb4d1a85fdd6ca50fc4376", "createdAt": "2026-01-07T16:40:38.094131+08:00", "customerName": "予芯", "intendedProduct": "2025/12/22", "version": "1.9.4", "description": "训练失败看不到错误日志", "solution": "已解决", "type": "功能问题", "module": "模型工坊", "statusProgress": "已修复", "reporter": "" }, { "id": "179dd08fe93d4fbf9c078c362486aa65", "createdAt": "2026-01-07T16:40:38.095578+08:00", "customerName": "诺因智能", "intendedProduct": "2025/12/23", "version": "1.9.4", "description": "客户端上传大文件失败", "solution": "技术优化，修复中", "type": "功能问题", "module": "数据空间", "statusProgress": "已修复", "reporter": "" }, { "id": "2700f88f74c94903a48369c86e05221a", "createdAt": "2026-01-07T16:40:38.096324+08:00", "customerName": "良业", "intendedProduct": "2025/12/23", "version": "1.9.4", "description": "数据集发布感觉操作繁琐，如果在模型工坊里直接选择需要训练的图片会更方便", "solution": "", "type": "反馈", "module": "数据集,模型工坊", "statusProgress": "", "reporter": "" }, { "id": "7ae437bfeaa84e579b012a4ec5fae794", "createdAt": "2026-01-07T16:40:38.097176+08:00", "customerName": "斯蒂尔", "intendedProduct": "2025/12/25", "version": "1.9.4", "description": "1.试用账号图片生成速度比较慢(2分钟1张图)\n2.生图效果不理想", "solution": "试用资源有限", "type": "反馈", "module": "数据生成", "statusProgress": "", "reporter": "" }, { "id": "e45d0a13be724c129a3c4e446517a940", "createdAt": "2026-01-07T16:40:38.098053+08:00", "customerName": "求之", "intendedProduct": "2025/12/26", "version": "1.9.4", "description": "1.训练过程不太透明，以及无法提前终止模型训练\n2.训练过程无法断点继续训练；\n3.试用账号资源有限，训练时长过长", "solution": "1、训练过程的透明化（日志化）和目前的易用化，有一些产品设计矛盾，之前主要还是在做易用化。\n接下来我们会在更透明，更强的客户对过程的操控粒度上提升。\n2、训练时长的事情实在抱歉，最近客户比较多，我们资源有限", "type": "反馈", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "157ee3180f9c47efb38ab52e3a3b2b4f", "createdAt": "2026-01-07T16:40:38.099239+08:00", "customerName": "诺因智能", "intendedProduct": "2025/12/26", "version": "1.9.4", "description": "训练进程显示什么时候上线", "solution": "训练日志可视化，我们大约需要一个月的时间可以满足上线使用", "type": "需求", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "b7371743ec5549af8ce643e136334ccf", "createdAt": "2026-01-07T16:40:38.100073+08:00", "customerName": "诺因智能", "intendedProduct": "2025/12/26", "version": "1.9.4", "description": "(igev模型)全量训练是指从0开始？ 增量是在现在的ckpt上做微调是吗？ ", "solution": "全量训练：在加载开源基础的权重上进行增量训练\n增量训练：在您的数据上训练的checkpoint上进行的增量训练", "type": "咨询", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "c87211e0dc6b42f1bb2127259cf64e81", "createdAt": "2026-01-07T16:40:38.101288+08:00", "customerName": "斯蒂尔", "intendedProduct": "2025/12/26", "version": "1.9.4", "description": "数据生成效果改进优化方式？试用账号配置卡数", "solution": "4090，24g，然后生图是单卡单线程\n目前系统上挂载了2张GPU", "type": "咨询", "module": "数据生成", "statusProgress": "", "reporter": "" }, { "id": "d14245e86d3d4ee399996ba6d26b23de", "createdAt": "2026-01-07T16:40:38.101877+08:00", "customerName": "良业", "intendedProduct": "2025/12/26", "version": "1.9.4", "description": "1、训练时需要添加测试图，无法理解，模型分析结果可以用训练集\n2、训练失败时需提示原因，方便下一步操作。", "solution": "1、训练过程中，根据训练集的数据来更新模型参数；在通过验证集来验证；\n2、训练失败有日志可以查看，您滚动下屏幕至右侧可以看到哈", "type": "咨询", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "e6fe3693b67d406e9edd056fc48c7119", "createdAt": "2026-01-07T16:40:38.102556+08:00", "customerName": "良业", "intendedProduct": "2025/12/29", "version": "1.9.4", "description": "1.试了几个模型，没训练出来，感觉不出效果\n", "solution": "补充数据集", "type": "咨询", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "ec24619e11014cb4a3e2f6d9e4b2d618", "createdAt": "2026-01-07T16:40:38.103257+08:00", "customerName": "诺因智能", "intendedProduct": "2025/12/29", "version": "1.9.4", "description": "1.总轮数与拓展参数(num_steps)不一致的时候，以哪个为准结束训练", "solution": "目前StereoNet-V2.5支持使用steps参数，其他的还是以epoch为准", "type": "咨询", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "a9fc0da9bee44acba2d43d15e06079ed", "createdAt": "2026-01-07T16:40:38.10396+08:00", "customerName": "良业", "intendedProduct": "2025/12/30", "version": "1.9.4", "description": "户外同一个草坪测试，用yolo检测640*480图片，检测时间\u003c100ms，分割精度\u003c10像素，出错率1/10万，使用x3可以吗？", "solution": "100ms可以", "type": "咨询", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "0bd2e8d859d7466293182181490d56b4", "createdAt": "2026-01-07T16:40:38.104695+08:00", "customerName": "良业", "intendedProduct": "2025/12/30", "version": "1.9.4", "description": "训练任务无效果", "solution": "数据集少了，补数后需要保存快照", "type": "咨询", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "3576f30f7b7d489e91f2c638b9aae7af", "createdAt": "2026-01-07T16:40:38.105387+08:00", "customerName": "诺因智能", "intendedProduct": "2025/12/30", "version": "1.9.4", "description": "现在平台上的V2.5就可以，需要完成的ckpt，这个预计什么时候可以更新啊", "solution": "", "type": "咨询", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "1cdb5f53ee734caf859fc0597a60f13a", "createdAt": "2026-01-07T16:40:38.106029+08:00", "customerName": "诺因智能", "intendedProduct": "2025/12/31", "version": "1.9.4", "description": "境外网站的加速，下载部分数据集时速度会很慢", "solution": "", "type": "反馈", "module": "数据集", "statusProgress": "", "reporter": "" }, { "id": "e01001a4667d45e0872d506df39be397", "createdAt": "2026-01-07T16:40:38.106833+08:00", "customerName": "诺因智能", "intendedProduct": "2025/12/31", "version": "1.9.4", "description": "诺因智能需要完整的模型 checkpoint，而不仅仅是 backbone 部分", "solution": "", "type": "需求", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "e52a8badaab840a2a6dd27a614d1e929", "createdAt": "2026-01-07T16:40:38.107656+08:00", "customerName": "诺因智能", "intendedProduct": "2025/12/31", "version": "1.9.4", "description": "模型推理的指标对比按钮无法选择", "solution": "", "type": "需求", "module": "模型工坊", "statusProgress": "", "reporter": "" }, { "id": "bbf8dd91213b4d96b96bbeb43ed9930e", "createdAt": "2026-01-07T16:40:38.108654+08:00", "customerName": "诺因智能", "intendedProduct": "2025/12/31", "version": "1.9.4", "description": "推理评测可以增加定量参数，推理的时候能否增加一下epe衡量模型精度的指标", "solution": "", "type": "需求", "module": "模型工坊", "statusProgress": "", "reporter": "" } ]