File size: 2,964 Bytes
80b7188 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 | #!/bin/bash
set -x
# Warning: Export VLLM_ATTENTION_BACKEND on every machine before starting Ray cluster.
# vLLM without XFORMERS will results in CUDA errors.
export VLLM_ATTENTION_BACKEND=FLASH_ATTN
# Parse command line arguments
while [[ $# -gt 0 ]]; do
case $1 in
--model)
MODEL_PATH="$2"
shift 2
;;
*)
break
;;
esac
done
# Set default model path if not provided
if [ -z "$MODEL_PATH" ]; then
MODEL_PATH="Qwen/Qwen2.5-1.5B-Instruct"
fi
python3 -m rllm.trainer.verl.train_agent_ppo \
algorithm.adv_estimator=grpo \
data.train_files=$HOME/rllm/data/train.parquet \
data.val_files=$HOME/rllm/data/math.parquet \
env.name=math \
agent.name=math_agent \
agent.max_steps=5 \
agent.async_engine=True \
data.train_batch_size=8 \
data.val_batch_size=512 \
data.max_prompt_length=2048 \
data.max_response_length=8192 \
actor_rollout_ref.model.path=$MODEL_PATH \
actor_rollout_ref.actor.optim.lr=1e-6 \
actor_rollout_ref.model.use_remove_padding=True \
actor_rollout_ref.hybrid_engine=False \
actor_rollout_ref.actor.ppo_mini_batch_size=2 \
actor_rollout_ref.actor.use_dynamic_bsz=True \
actor_rollout_ref.actor.ppo_max_token_len_per_gpu=3072 \
actor_rollout_ref.actor.use_kl_loss=False \
actor_rollout_ref.actor.kl_loss_coef=0.001 \
actor_rollout_ref.actor.kl_loss_type=low_var_kl \
actor_rollout_ref.actor.ulysses_sequence_parallel_size=1 \
actor_rollout_ref.model.enable_gradient_checkpointing=True \
actor_rollout_ref.actor.fsdp_config.param_offload=False \
+actor_rollout_ref.actor.fsdp_config.grad_offload=False \
actor_rollout_ref.actor.fsdp_config.optimizer_offload=False \
actor_rollout_ref.rollout.tensor_model_parallel_size=1 \
actor_rollout_ref.rollout.name=vllm \
actor_rollout_ref.rollout.mode="async" \
actor_rollout_ref.rollout.chat_scheduler=verl.schedulers.naive_chat_scheduler.NaiveChatCompletionScheduler \
actor_rollout_ref.rollout.temperature=0.6 \
actor_rollout_ref.rollout.val_kwargs.temperature=0.6 \
actor_rollout_ref.rollout.gpu_memory_utilization=0.5 \
actor_rollout_ref.rollout.n=1 \
+actor_rollout_ref.rollout.n_val=1 \
actor_rollout_ref.rollout.enforce_eager=False \
actor_rollout_ref.ref.fsdp_config.param_offload=True \
algorithm.kl_ctrl.kl_coef=0.001 \
trainer.critic_warmup=0 \
trainer.logger=['console','wandb'] \
trainer.project_name='debug' \
trainer.experiment_name='deepscaler-debug' \
+trainer.val_before_train=False \
trainer.n_gpus_per_node=2 \
+trainer.n_training_gpus_per_node=1 \
trainer.nnodes=1 \
trainer.save_freq=5 \
trainer.test_freq=5 \
trainer.remove_previous_ckpt_in_save=True \
trainer.default_hdfs_dir=null \
trainer.total_epochs=30 "${@:1}" |