keras-transformer：变压器的Keras实现-源码
变压器
该项目实施“注意就是您所需要的”论文。
我们提供了详细的指南，以了解变压器的概念和所需的Keras功能/技巧（请参阅Guide.md和Walkthrough.md）。
我们希望该软件和文档将使社区能够以更健壮和更快的方式创建和自动化自定义Keras图层/模型/解决方案。
我们还分享了可以通过DSL设置的机器翻译演示。
我们将不胜感激：）
如果可以训练任何大型模型并共享结果，那就太好了。 由于计算资源有限，目前无法测试。
尽管我们在开发过程中进行了正式和非正式的测试； 我们将不胜感激。 还有任何建议和要求。
如果您还有其他Keras技巧或替代方法，也请为文档做出贡献。
关键贡献
可以通过DSL定义模型训练和解码。 因此执行是自动化的。
从定制培训师那里抽象培训基础知识（检查点，工件管理等）。
为生成模型的生成器（所谓的inner-generator ）保留明确的接口
文件下载
资源详情
[{"title":"（ 77 个子文件 80KB ） keras-transformer：变压器的Keras实现-源码","children":[{"title":"keras-transformer-master","children":[{"title":"Walkthrough.md <span style='color:#111;'> 22.47KB </span>","children":null,"spread":false},{"title":"keras_transformer","children":[{"title":"keras_transformer","children":[{"title":"utils","children":[{"title":"context_utils.py <span style='color:#111;'> 1.11KB </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"common_utils.py <span style='color:#111;'> 589B </span>","children":null,"spread":false},{"title":"io_utils.py <span style='color:#111;'> 800B </span>","children":null,"spread":false}],"spread":true},{"title":"generators","children":[{"title":"InnerGenerator.py <span style='color:#111;'> 1.18KB </span>","children":null,"spread":false},{"title":"outer","children":[{"title":"OuterGeneratorABC.py <span style='color:#111;'> 335B </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"PandasGenerator.py <span style='color:#111;'> 1.90KB </span>","children":null,"spread":false}],"spread":true},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false}],"spread":true},{"title":"core","children":[{"title":"TiedProjectionLayer.py <span style='color:#111;'> 1.39KB </span>","children":null,"spread":false},{"title":"Transformer.py <span style='color:#111;'> 8.35KB </span>","children":null,"spread":false},{"title":"attention","children":[{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"DotProductAttention.py <span style='color:#111;'> 5.79KB </span>","children":null,"spread":false},{"title":"MultiHeadAttention.py <span style='color:#111;'> 12.29KB </span>","children":null,"spread":false}],"spread":true},{"title":"PositionalEncodingLayer.py <span style='color:#111;'> 3.56KB </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"encoder_decoder","children":[{"title":"EncoderDecoder.py <span style='color:#111;'> 2.82KB </span>","children":null,"spread":false},{"title":"sub_layers","children":[{"title":"PositionWiseFeedForwardSublayer.py <span style='color:#111;'> 2.18KB </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"SelfAttentionSublayer.py <span style='color:#111;'> 2.43KB </span>","children":null,"spread":false}],"spread":false},{"title":"DecoderBlockStack.py <span style='color:#111;'> 2.54KB </span>","children":null,"spread":false},{"title":"EncoderBlockStack.py <span style='color:#111;'> 2.78KB </span>","children":null,"spread":false},{"title":"EncoderBlock.py <span style='color:#111;'> 2.05KB </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"normalization","children":[{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"LayerNormalization.py <span style='color:#111;'> 1.06KB </span>","children":null,"spread":false}],"spread":false},{"title":"DecoderBlock.py <span style='color:#111;'> 2.88KB </span>","children":null,"spread":false}],"spread":true},{"title":"ProjectionLayer.py <span style='color:#111;'> 1.71KB </span>","children":null,"spread":false},{"title":"TiedEmbedderProjector.py <span style='color:#111;'> 2.23KB </span>","children":null,"spread":false}],"spread":true},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"processors","children":[{"title":"ProcessorABC.py <span style='color:#111;'> 173B </span>","children":null,"spread":false},{"title":"processor_factory.py <span style='color:#111;'> 447B </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"SubWordProcessor.py <span style='color:#111;'> 4.35KB </span>","children":null,"spread":false},{"title":"DecodingCompatibleProcessorABC.py <span style='color:#111;'> 658B </span>","children":null,"spread":false}],"spread":true},{"title":"decoding","children":[{"title":"GreedyDecoder.py <span style='color:#111;'> 1.62KB </span>","children":null,"spread":false},{"title":"DecoderABC.py <span style='color:#111;'> 1008B </span>","children":null,"spread":false},{"title":"BeamSearchDecoder.py <span style='color:#111;'> 3.21KB </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false}],"spread":true},{"title":"training","children":[{"title":"custom_serialization","children":[{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"custom_layers.py <span style='color:#111;'> 2.21KB </span>","children":null,"spread":false},{"title":"custom_object_handler.py <span style='color:#111;'> 732B </span>","children":null,"spread":false}],"spread":true},{"title":"Trainer.py <span style='color:#111;'> 1.76KB </span>","children":null,"spread":false},{"title":"TrainerABC.py <span style='color:#111;'> 3.13KB </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"factories","children":[{"title":"loss_factory.py <span style='color:#111;'> 831B </span>","children":null,"spread":false},{"title":"optimizer_factory.py <span style='color:#111;'> 394B </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"callback_factory.py <span style='color:#111;'> 1.74KB </span>","children":null,"spread":false},{"title":"custom_weight_factory.py <span style='color:#111;'> 1.17KB </span>","children":null,"spread":false}],"spread":false},{"title":"custom_callbacks","children":[{"title":"CustomLearningRateScheduler.py <span style='color:#111;'> 1.15KB </span>","children":null,"spread":false},{"title":"CustomCallbackABC.py <span style='color:#111;'> 309B </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"CustomCheckpointer.py <span style='color:#111;'> 1.71KB </span>","children":null,"spread":false}],"spread":false}],"spread":true}],"spread":true},{"title":"demo","children":[{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"translation","children":[{"title":"TranslationDataGenerator.py <span style='color:#111;'> 3.00KB </span>","children":null,"spread":false},{"title":"training_conf1.json <span style='color:#111;'> 2.43KB </span>","children":null,"spread":false},{"title":"decoding_conf.json <span style='color:#111;'> 302B </span>","children":null,"spread":false},{"title":"training_conf2.json <span style='color:#111;'> 3.20KB </span>","children":null,"spread":false},{"title":"demo_execution.py <span style='color:#111;'> 1.13KB </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"samples.json <span style='color:#111;'> 312B </span>","children":null,"spread":false}],"spread":true}],"spread":true},{"title":"test","children":[{"title":"TestTiedEmbeddings.py <span style='color:#111;'> 13.87KB </span>","children":null,"spread":false},{"title":"TestCustomLoss.py <span style='color:#111;'> 1.62KB </span>","children":null,"spread":false},{"title":"TestPositionalEncoding.py <span style='color:#111;'> 2.58KB </span>","children":null,"spread":false},{"title":"TestMasking.py <span style='color:#111;'> 1.21KB </span>","children":null,"spread":false},{"title":"TestGenerators.py <span style='color:#111;'> 1.48KB </span>","children":null,"spread":false},{"title":"TestQKMask.py <span style='color:#111;'> 807B </span>","children":null,"spread":false},{"title":"__init__.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"TestBPE.py <span style='color:#111;'> 3.18KB </span>","children":null,"spread":false},{"title":"TestContextMask.py <span style='color:#111;'> 809B </span>","children":null,"spread":false}],"spread":true}],"spread":true},{"title":"LICENSE.txt <span style='color:#111;'> 578B </span>","children":null,"spread":false},{"title":"requirements.txt <span style='color:#111;'> 777B </span>","children":null,"spread":false},{"title":"README.md <span style='color:#111;'> 2.09KB </span>","children":null,"spread":false},{"title":"Guide.md <span style='color:#111;'> 26.90KB </span>","children":null,"spread":false},{"title":".gitignore <span style='color:#111;'> 50B </span>","children":null,"spread":false}],"spread":true}],"spread":true}]
评论信息
其他资源
- 传输矩阵法计算薄膜的反射率和透射率
- VC++实现悬浮窗 迅雷360悬浮球
- STM32F4简易示波器（基于原子探索者STM32F407开发板）
- android网上预约挂号服务端和客户端毕设源码
- SPAD SiPM 激光雷达sensor 技术简介
- 高通平台msm8096上用bq27541电量计驱动
- FBX格式人物模型，资源齐全，带动画，超值！！！
- 华为OJ题目部分合集
- OPENmp 快速排序的实现带部分注释
- aes对称加密示例，CBC模式，192密钥
- DBeaver Enterprise v7.3.0.zip
- 数学建模算法与应用(第2版)图书、课件及程序
- ieee 33节点
- 大数据基础-Linux基础详解课程46.sudo实例.mp4
- 2020年中国网络安全产业白皮书.pdf
- 自动问答系统(后台+小程序).zip
- sonar最大平均圈复杂度
- 桌面助手，不下载会后悔的
免责申明
【只为小站】的资源来自网友分享，仅供学习研究，请务必在下载后24小时内给予删除，不得用于其他任何用途，否则后果自负。基于互联网的特殊性，【只为小站】 无法对用户传输的作品、信息、内容的权属或合法性、合规性、真实性、科学性、完整权、有效性等进行实质审查；无论 【只为小站】 经营者是否已进行审查，用户均应自行承担因其传输的作品、信息、内容而可能或已经产生的侵权或权属纠纷等法律责任。
本站所有资源不代表本站的观点或立场，基于网友分享，根据中国法律《信息网络传播权保护条例》第二十二条之规定，若资源存在侵权或相关问题请联系本站客服人员，zhiweidada#qq.com，请把#换成@，本站将给予最大的支持与配合，做到及时反馈和处理。关于更多版权及免责申明参见 版权及免责申明
本站所有资源不代表本站的观点或立场，基于网友分享，根据中国法律《信息网络传播权保护条例》第二十二条之规定，若资源存在侵权或相关问题请联系本站客服人员，zhiweidada#qq.com，请把#换成@，本站将给予最大的支持与配合，做到及时反馈和处理。关于更多版权及免责申明参见 版权及免责申明