Feature request: produce tf.train.SequenceExample s from compiler logs
Feature request: produce tf.train.SequenceExample s from compiler logs: a task in MiMo-V2.6-RL-oss: Agentic RL Environments (MiMo RL release). Feature request: produce tf.train.SequenceExamples from compiler logs compiler opt/rl/log reader.py exposes read log, which yields Records (features +…
The task
Feature request: produce tf.train.SequenceExamples from compiler logs compiler_opt/rl/log_reader.py exposes read_log, which yields Records (features + score per step) from the simple log format the compiler emits. That's great for inspection, but for actually training a policy…
Part of XiaomiMiMo/MiMo-V2.6-RL-oss.