GRPO LoRA adapters for JMH microbenchmark generation. One repo per run: final adapter at root, every-200-step checkpoints, training metrics.
Jonas Drechsel
bookxd
AI & ML interests
None yet
Recent Activity
updated a dataset 10 days ago
bookxd/jmh-grpo-replication-data published a model 10 days ago
bookxd/gemma-4-E2B-it-jmh-grpo-mutation70-run8 updated a model 10 days ago
bookxd/gemma-4-E2B-it-jmh-grpo-mutation70-run9Organizations
None yet