add qwen3-32b grpo script on npu

Author: tardis-key

docs/ascend_tutorial/ascend_quick_start.rst

@@ -179,6 +179,8 @@ vllm & vllm-ascend
 +-----------+-------------------------+-------------+-------------------+-------------------+-------------------+--------------------------+
 |   GRPO    | Qwen2.5-VL-32B-instruct |    0.79%    |        0.568      |        FSDP       |    vllm-ascend    |    Atlas 200T A2 Box16   |
 +-----------+-------------------------+-------------+-------------------+-------------------+-------------------+--------------------------+
+|   GRPO    | Qwen3-32B               |    0.64%    |        0.690      |        FSDP       |    vllm-ascend    |    Atlas 200T A2 Box16   |
++-----------+-------------------------+-------------+-------------------+-------------------+-------------------+--------------------------+
 |   DAPO    | Qwen2.5-7B-instruct     |    3.83%    |        pending    |        FSDP       |    vllm-ascend    |    Atlas 200T A2 Box16   |
 +-----------+-------------------------+-------------+-------------------+-------------------+-------------------+--------------------------+
 |   DAPO    | Qwen2.5-32B             |    3.42%    |        pending    |        FSDP       |    vllm-ascend    |    Atlas 200T A2 Box16   |

examples/grpo_trainer/run_qwen3_32b_npu.sh

@@ -0,0 +1,51 @@
+set -x
+
+python3 -m verl.trainer.main_ppo \
+    algorithm.adv_estimator=grpo \
+    data.train_files=$HOME/data/gsm8k/train.parquet \
+    data.val_files=$HOME/data/gsm8k/test.parquet \
+    data.train_batch_size=1024 \
+    data.max_prompt_length=2048 \
+    data.max_response_length=2048 \
+    data.filter_overlong_prompts=True \
+    data.truncation='error' \
+    data.shuffle=False \
+    actor_rollout_ref.model.path=Qwen/Qwen3-32B \
+    actor_rollout_ref.actor.optim.lr=1e-6 \
+    actor_rollout_ref.model.use_remove_padding=True \
+    actor_rollout_ref.actor.ulysses_sequence_parallel_size=4 \
+    +actor_rollout_ref.actor.fsdp_config.mixed_precision.param_dtype=bf16 \
+    +actor_rollout_ref.actor.fsdp_config.mixed_precision.reduce_dtype=bf16 \
+    +actor_rollout_ref.actor.fsdp_config.mixed_precision.buffer_dtype=fp32 \
+    actor_rollout_ref.actor.ppo_mini_batch_size=64 \
+    actor_rollout_ref.actor.ppo_micro_batch_size_per_gpu=8 \
+    actor_rollout_ref.actor.use_kl_loss=True \
+    actor_rollout_ref.actor.entropy_coeff=0 \
+    actor_rollout_ref.actor.kl_loss_coef=0.001 \
+    actor_rollout_ref.actor.kl_loss_type=low_var_kl \
+    actor_rollout_ref.model.enable_gradient_checkpointing=True \
+    actor_rollout_ref.actor.fsdp_config.param_offload=True \
+    actor_rollout_ref.actor.fsdp_config.optimizer_offload=False \
+    actor_rollout_ref.rollout.log_prob_micro_batch_size_per_gpu=8 \
+    actor_rollout_ref.rollout.tensor_model_parallel_size=4 \
+    actor_rollout_ref.rollout.name=vllm \
+    actor_rollout_ref.rollout.gpu_memory_utilization=0.7 \
+    actor_rollout_ref.rollout.n=4 \
+    actor_rollout_ref.ref.log_prob_micro_batch_size_per_gpu=8 \
+    actor_rollout_ref.ref.fsdp_config.param_offload=True \
+    actor_rollout_ref.actor.use_torch_compile=False \
+    actor_rollout_ref.ref.use_torch_compile=False \
+    actor_rollout_ref.rollout.enable_chunked_prefill=True \
+    actor_rollout_ref.rollout.max_num_batched_tokens=32768 \
+    algorithm.use_kl_in_reward=False \
+    trainer.critic_warmup=0 \
+    trainer.logger=['console','tensorboard'] \
+    trainer.project_name='verl_grpo_example_gsm8k_fsdp' \
+    trainer.experiment_name='qwen3_32b_fsdp' \
+    trainer.n_gpus_per_node=16 \
+    trainer.nnodes=2 \
+    trainer.resume_from_path=checkpoints/ \
+    trainer.save_freq=500 \
+    trainer.test_freq=50 \
+    trainer.total_epochs=50 \
+    trainer.device=npu $@