# returncode: 0

### 0
"pkill"
"-9"
"sglang"

### 1
"sleep"
"3"

### 2
"ray"
"stop"
"--force"

### 3
"pkill"
"-9"
"ray"

### 4
"pkill"
"-9"
"python"

### 5
"sleep"
"3"

### 6
"pkill"
"-9"
"ray"

### 7
"pkill"
"-9"
"python"

### 8
"nvidia-smi"

### 9
"date"
"+%Y%m%d_%H%M%S"

### 10
"ray"
"start"
"--head"
"--node-ip-address"
"127.0.0.1"
"--num-gpus"
"8"
"--disable-usage-stats"

### 11
"ray"
"job"
"submit"
"--address=http://127.0.0.1:8265"
"--runtime-env-json={\n  \"env_vars\": {\n    \"PYTHONPATH\": \"/root/Megatron-LM/:<REPO_ROOT>/examples/infra_features/train_infer_mismatch_helper\",\n    \"CUDA_DEVICE_MAX_CONNECTIONS\": \"1\"\n  }\n}"
"--"
"python3"
"train.py"
"--hf-checkpoint"
"/root/Qwen3-4B"
"--load"
"/root/Qwen3-4B"
"--ref-load"
"/root/Qwen3-4B"
"--prompt-data"
"/root/dapo-math-17k/dapo-math-17k.jsonl"
"--input-key"
"prompt"
"--label-key"
"label"
"--apply-chat-template"
"--rollout-shuffle"
"--balance-data"
"--rm-type"
"deepscaler"
"--num-rollout"
"100"
"--rollout-batch-size"
"8"
"--n-samples-per-prompt"
"8"
"--rollout-max-response-len"
"4096"
"--rollout-temperature"
"0.8"
"--global-batch-size"
"64"
"--optimizer"
"adam"
"--lr"
"1e-6"
"--lr-decay-style"
"constant"
"--weight-decay"
"0.1"
"--adam-beta1"
"0.9"
"--adam-beta2"
"0.98"
"--use-kl-loss"
"--advantage-estimator"
"grpo"
"--kl-loss-coef"
"0.00"
"--kl-loss-type"
"low_var_kl"
"--kl-coef"
"0.00"
"--entropy-coef"
"0.00"
"--eps-clip"
"0.2"
"--eps-clip-high"
"0.28"
"--use-tis"
"--use-wandb"
"--wandb-project"
"miles-dev-mcore-fsdp"
"--wandb-group"
"qwen3-4B-fsdp-1130-ref"
"--wandb-key"
"frozen-wandb-api-key"
"--rollout-num-gpus-per-engine"
"1"
"--sglang-mem-fraction-static"
"0.75"
"--sglang-decode-log-interval"
"1000"
"--sglang-chunked-prefill-size"
"4096"
"--sglang-attention-backend"
"fa3"
"--train-backend"
"fsdp"
"--update-weight-buffer-size"
"536870912"
"--gradient-checkpointing"
"--attn-implementation"
"flash_attention_3"
"--train-env-vars"
"{\"PYTORCH_CUDA_ALLOC_CONF\":\"expandable_segments:True\"}"
"--use-dynamic-batch-size"
"--max-tokens-per-gpu"
"9216"
"--actor-num-nodes"
"1"
"--actor-num-gpus-per-node"
"8"
"--colocate"
"--dump-details"
"/root/shared_data/qwen3-4B-fsdp-1116-noref/dump_details"
"--custom-config-path"
"examples/infra_features/train_infer_mismatch_helper/mis.yaml"
"--custom-tis-function-path"
"examples.infra_features.train_infer_mismatch_helper.mis.compute_mis_weights_fsdp"
