executing-plans
executing-plans
Use when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available
Skills 둘러보기
생산성 및 협업 · SKILL.md
sequence-and-rnn
sequence-and-rnnUse Sonnet RNN cores, sequence unroll helpers, trainable recurrent state, and convolutional LSTMs.
Independent GitHub snapshot · Automated classification · Original documentation governs
Skill 문서
GitHubname: sequence-and-rnn description: "Use Sonnet RNN cores, sequence unroll helpers, trainable recurrent state, and convolutional LSTMs." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0
Sequence and RNN
Use this sub-skill for snt.RNNCore, VanillaRNN, LSTM, GRU, DeepRNN, skip/residual helpers, static_unroll, dynamic_unroll, TrainableState, UnrolledLSTM, recurrent dropout, and ConvLSTM modules.
Start here
- references/api-reference.md: signatures, state structures, and shape conventions.
- references/workflows.md: LSTM/GRU unrolls,
DeepRNN, trainable state, recurrent dropout, and ConvLSTM recipes. - references/troubleshooting.md: time-major, state, skip/residual, ConvLSTM, and backend-specialization errors.
- scripts/rnn_unroll_smoke.py: safe LSTM/GRU unroll shape smoke.
Boundaries
- Generic
snt.Moduleauthoring: ../module-authoring/SKILL.md. - Feed-forward layers/nets: ../layers-and-nets/SKILL.md.
- Optimizer loops: ../training-and-optimization/SKILL.md.
- XLA/distribution/backend policy: ../serialization-and-distribution/SKILL.md.