[rollout, sglang] feat: Add sync mode for bash (#3186)

### What does this PR do?
- Use `sync` mode for `dapo`, `gsm8k` and `geo`

### Checklist Before Submitting

> [!IMPORTANT]
> Please check all the following items before requesting a review,
otherwise the reviewer might deprioritize this PR for review.

- [x] Read the [Contribute
Guide](https://github.com/volcengine/verl/blob/main/CONTRIBUTING.md).
- [x] Apply [pre-commit
checks](https://github.com/volcengine/verl/blob/main/CONTRIBUTING.md#code-linting-and-formatting):
`pre-commit install && pre-commit run --all-files --show-diff-on-failure
--color=always`

examples/grpo_trainer/run_qwen2_5_vl-7b-sglang.sh

@@ -40,6 +40,7 @@ python3 -m verl.trainer.main_ppo \
     actor_rollout_ref.rollout.n=5 \
     actor_rollout_ref.ref.log_prob_micro_batch_size_per_gpu=20 \
     actor_rollout_ref.ref.fsdp_config.param_offload=True \
+    actor_rollout_ref.rollout.mode=sync \
     algorithm.use_kl_in_reward=False \
     trainer.critic_warmup=0 \
     trainer.logger='["console","wandb"]' \

examples/sglang_multiturn/run_qwen2.5-3b_gsm8k_multiturn.sh

@@ -48,7 +48,8 @@ python3 -m verl.trainer.main_ppo \
     actor_rollout_ref.rollout.n=16 \
     actor_rollout_ref.ref.log_prob_micro_batch_size_per_gpu=32 \
     actor_rollout_ref.ref.fsdp_config.param_offload=True \
-    actor_rollout_ref.rollout.over_sample_rate=0 \
+    actor_rollout_ref.rollout.over_sample_rate=0.1 \
+    actor_rollout_ref.rollout.mode=sync \
     algorithm.use_kl_in_reward=False \
     trainer.critic_warmup=0 \
     trainer.logger='["console","wandb"]' \

examples/sglang_multiturn/run_qwen3-4b_gsm8k_multiturn.sh

@@ -35,6 +35,8 @@ python3 -m verl.trainer.main_ppo \
     actor_rollout_ref.rollout.name=sglang \
     actor_rollout_ref.rollout.gpu_memory_utilization=0.5 \
     actor_rollout_ref.rollout.n=16 \
+    actor_rollout_ref.rollout.over_sample_rate=0.1 \
+    actor_rollout_ref.rollout.mode=sync \
     actor_rollout_ref.ref.log_prob_micro_batch_size_per_gpu=32 \
     actor_rollout_ref.ref.fsdp_config.param_offload=True \
     algorithm.use_kl_in_reward=False \