Add files using upload-large-folder tool

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb/wandb/debug-internal.log

@@ -0,0 +1,31 @@
+{"time":"2026-05-02T21:01:27.089287178+09:00","level":"INFO","msg":"wandb-core"}
+{"time":"2026-05-02T21:01:27.116487637+09:00","level":"INFO","msg":"stream: starting","core version":"0.25.1"}
+{"time":"2026-05-02T21:01:32.177703408+09:00","level":"ERROR","msg":"monitor: failed to initialize GPU resource: monitor: could not get GPU binary port: timeout reading portfile /tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/wandb-system-monitor-portfile-791433119"}
+{"time":"2026-05-02T21:01:32.463460082+09:00","level":"INFO","msg":"stream: created new stream","id":"m9xk3p9p"}
+{"time":"2026-05-02T21:01:32.464136548+09:00","level":"INFO","msg":"handler: started"}
+{"time":"2026-05-02T21:01:32.465451984+09:00","level":"INFO","msg":"stream: started"}
+{"time":"2026-05-02T21:01:32.466026932+09:00","level":"INFO","msg":"writer: started","stream_id":"m9xk3p9p"}
+{"time":"2026-05-02T21:01:32.467207947+09:00","level":"INFO","msg":"sender: started"}
+{"time":"2026-05-02T21:01:33.577986724+09:00","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":0,"console_lines":1}
+{"time":"2026-05-02T21:01:34.004592056+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:01:48.576628897+09:00","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":0,"events_lines":1,"console_offset":0,"console_lines":16,"uploaded_len":2}
+{"time":"2026-05-02T21:01:49.073487601+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:02:03.57739858+09:00","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":1,"events_lines":1,"console_offset":11,"console_lines":1}
+{"time":"2026-05-02T21:02:03.995233262+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:02:18.577863349+09:00","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":2,"events_lines":1,"console_offset":16,"console_lines":15}
+{"time":"2026-05-02T21:02:18.859874591+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:02:33.577360034+09:00","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":3,"events_lines":1,"console_offset":31,"console_lines":15}
+{"time":"2026-05-02T21:02:33.967842201+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:02:39.041087354+09:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":187}
+{"time":"2026-05-02T21:02:39.047778317+09:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":32}
+{"time":"2026-05-02T21:02:39.151241183+09:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":305}
+{"time":"2026-05-02T21:02:39.157195106+09:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":72}
+{"time":"2026-05-02T21:02:40.599169371+09:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":477}
+{"time":"2026-05-02T21:02:40.614228567+09:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":242}
+{"time":"2026-05-02T21:02:41.811243346+09:00","level":"INFO","msg":"stream: closing"}
+{"time":"2026-05-02T21:02:42.745957822+09:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-05-02T21:02:42.747911197+09:00","level":"INFO","msg":"filestream: sending request","total_files":3,"history_offset":0,"history_lines":1,"console_offset":46,"console_lines":516,"uploaded_len":3,"complete":true,"exit_code":1}
+{"time":"2026-05-02T21:02:43.119821689+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:02:43.120030377+09:00","level":"INFO","msg":"handler: closed"}
+{"time":"2026-05-02T21:02:43.120442815+09:00","level":"INFO","msg":"sender: closed"}
+{"time":"2026-05-02T21:02:43.120456852+09:00","level":"INFO","msg":"stream: closed"}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb/wandb/run-20260502_210121-m9xk3p9p/files/config.yaml

@@ -0,0 +1,376 @@
+_wandb:
+    value:
+        cli_version: 0.25.1
+        e:
+            4is3svrhq809lwa4dst4qbj79auqgiq9:
+                args:
+                    - --config
+                    - /home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/opsd_distillation_config.yaml
+                    - policy.model_name=/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+                    - policy.dtensor_cfg.tensor_parallel_size=4
+                    - policy.dtensor_cfg.context_parallel_size=1
+                    - policy.max_total_sequence_length=10000
+                    - policy.generation.max_new_tokens=8192
+                    - policy.generation.vllm_cfg.tensor_parallel_size=4
+                    - policy.generation.vllm_cfg.gpu_memory_utilization=0.7
+                    - policy.logprob_batch_size=1
+                    - policy.dtensor_cfg.cpu_offload=true
+                    - policy.dtensor_cfg.activation_checkpointing=true
+                    - paths.base_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd
+                    - paths.data_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/data/knowledge_d5_e10/atomic_sft
+                    - checkpointing.checkpoint_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927
+                    - logger.log_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs
+                    - logger.wandb.project=mix-teachers-opsd
+                    - logger.wandb.name=offpsd-knowledge-nll-gt-d5-e10-n1-t0-20260502-Llama-3.2-3B-Instruct
+                    - logger.tensorboard.log_dir=tb_logs-offpsd-knowledge-nll-gt-d5-e10-llama3-2-3b-n1-t0
+                    - cluster.gpus_per_node=4
+                    - cluster.num_nodes=1
+                    - loss_fn.kl_type=forward
+                    - loss_fn.zero_outside_topk=false
+                    - distillation.ema_decay=1.0
+                    - distillation.topk_logits_k=64
+                    - distillation.num_prompts_per_step=20
+                    - distillation.num_generations_per_prompt=1
+                    - distillation.rollout_policy=teacher
+                    - +distillation.use_nll_loss=true
+                    - distillation.val_period=20
+                    - distillation.max_num_epochs=40
+                    - checkpointing.save_per_epoch=false
+                    - checkpointing.save_period=20
+                    - checkpointing.save_optimizer=false
+                    - checkpointing.keep_top_k=null
+                    - checkpointing.metric_name=null
+                    - policy.train_global_batch_size=20
+                    - policy.generation_batch_size=20
+                    - policy.generation.temperature=0.0
+                    - policy.optimizer.kwargs.lr=3.0e-5
+                    - policy.optimizer.kwargs.weight_decay=0.01
+                    - 'policy.scheduler=[{name: "torch.optim.lr_scheduler.ConstantLR", kwargs: {factor: 1.0, total_iters: 10000000000}}, {milestones: []}]'
+                codePath: mix_teachers/train/knowledge_opsd/run_opsd.py
+                codePathLocal: run_opsd.py
+                cpu_count: 120
+                cpu_count_logical: 120
+                disk:
+                    /:
+                        total: "511749718016"
+                        used: "24012599296"
+                email: jiaruil5@andrew.cmu.edu
+                executable: /home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3
+                git:
+                    commit: 9938bfa2b66c98512427baa8c5adbea81dbf6dc8
+                    remote: https://github.com/jiarui-liu/TeacherMix.git
+                host: b0020
+                memory:
+                    total: "1081436749824"
+                os: Linux-4.18.0-477.27.1.el8_8.x86_64-x86_64-with-glibc2.28
+                program: /home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py
+                python: CPython 3.11.15
+                root: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb
+                startedAt: "2026-05-02T12:01:21.495560Z"
+                writerId: 4is3svrhq809lwa4dst4qbj79auqgiq9
+        m:
+            - "2": ray/*
+              "5": 2
+              "6":
+                - 1
+              "7": []
+            - "1": ray/ray_step
+              "6": []
+              "7": []
+        python_version: 3.11.15
+        t:
+            "1":
+                - 1
+                - 5
+                - 11
+                - 30
+                - 41
+                - 49
+                - 50
+                - 51
+                - 53
+                - 71
+            "2":
+                - 1
+                - 5
+                - 11
+                - 30
+                - 41
+                - 49
+                - 50
+                - 51
+                - 53
+                - 71
+            "3":
+                - 7
+                - 13
+                - 41
+            "4": 3.11.15
+            "5": 0.25.1
+            "6": 4.57.1
+            "12": 0.25.1
+            "13": linux-x86_64
+checkpointing:
+    value:
+        checkpoint_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927
+        checkpoint_must_save_by: null
+        enabled: true
+        higher_is_better: true
+        keep_top_k: null
+        metric_name: null
+        model_save_format: safetensors
+        save_consolidated: false
+        save_optimizer: false
+        save_per_epoch: false
+        save_period: 20
+cluster:
+    value:
+        gpus_per_node: 4
+        num_nodes: 1
+data:
+    value:
+        default:
+            env_name: knowledge
+            prompt_file: null
+            system_prompt_file: null
+        max_input_seq_length: 10000
+        shuffle: true
+        train:
+            - dataset_name: knowledge_opsd
+              env_name: knowledge
+              processor: knowledge_opsd
+              prompt_file: null
+              system_prompt_file: null
+        validation:
+            - dataset_name: knowledge_opsd_val
+              env_name: knowledge
+              processor: knowledge_opsd
+              prompt_file: null
+              system_prompt_file: null
+            - dataset_name: knowledge_opsd_comp_val
+              env_name: knowledge
+              processor: knowledge_opsd
+              prompt_file: null
+              system_prompt_file: null
+distillation:
+    value:
+        ema_decay: 1
+        max_num_epochs: 40
+        max_num_steps: 1000
+        max_rollout_turns: 1
+        max_val_samples: 200
+        num_generations_per_prompt: 1
+        num_prompts_per_step: 20
+        rollout_policy: teacher
+        seed: 42
+        topk_logits_k: 64
+        use_nll_loss: true
+        val_at_end: true
+        val_at_start: true
+        val_batch_size: 32
+        val_period: 20
+env:
+    value:
+        knowledge:
+            num_workers: 8
+logger:
+    value:
+        gpu_monitoring:
+            collection_interval: 10
+            flush_interval: 10
+        log_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001
+        mlflow_enabled: false
+        monitor_gpus: true
+        num_val_samples_to_print: 5
+        swanlab_enabled: false
+        tensorboard:
+            log_dir: tb_logs-offpsd-knowledge-nll-gt-d5-e10-llama3-2-3b-n1-t0
+        tensorboard_enabled: true
+        wandb:
+            entity: mathrl-jiarui
+            name: offpsd-knowledge-nll-gt-d5-e10-n1-t0-20260502-Llama-3.2-3B-Instruct
+            project: mix-teachers-opsd
+        wandb_enabled: true
+loss_fn:
+    value:
+        kl_type: forward
+        mixed_kl_weight: 0.5
+        zero_outside_topk: false
+paths:
+    value:
+        base_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd
+        checkpoint_dir: /home/fs01/yw2349/jiarui/mathrl_data/mix_teachers/checkpoints/knowledge/opsd_nemo_rl
+        data_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/data/knowledge_d5_e10/atomic_sft
+        log_dir: /home/fs01/yw2349/jiarui/mathrl_data/mix_teachers/checkpoints/knowledge/opsd_nemo_rl/logs
+        model_search_dirs:
+            - /home/fs01/yw2349/jiarui/mathrl_data/models
+policy:
+    value:
+        dtensor_cfg:
+            _v2: true
+            activation_checkpointing: true
+            context_parallel_size: 1
+            cpu_offload: true
+            custom_parallel_plan: null
+            enabled: true
+            sequence_parallel: false
+            tensor_parallel_size: 4
+        dynamic_batching:
+            enabled: true
+            logprob_mb_tokens: 10000
+            sequence_length_round: 64
+            train_mb_tokens: 10000
+        generation:
+            _pad_token_id: 128009
+            backend: vllm
+            colocated:
+                enabled: true
+                resources:
+                    gpus_per_node: null
+                    num_nodes: null
+            max_new_tokens: 8192
+            stop_strings: null
+            stop_token_ids:
+                - 128009
+            temperature: 0
+            top_k: null
+            top_p: 1
+            vllm_cfg:
+                async_engine: false
+                distributed_executor_backend: null
+                enforce_eager: false
+                expert_parallel_size: 1
+                gpu_memory_utilization: 0.7
+                kv_cache_dtype: auto
+                load_format: dummy
+                max_model_len: 10000
+                num_first_layers_in_bf16: 0
+                num_last_layers_in_bf16: 0
+                pipeline_parallel_size: 1
+                precision: bfloat16
+                skip_tokenizer_init: true
+                tensor_parallel_size: 4
+                use_deep_gemm: false
+        generation_batch_size: 20
+        logprob_batch_size: 1
+        logprob_chunk_size: null
+        make_sequence_length_divisible_by: 8
+        max_grad_norm: 1
+        max_total_sequence_length: 10000
+        megatron_cfg:
+            enabled: false
+        model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+        offload_optimizer_for_logprob: false
+        optimizer:
+            kwargs:
+                betas:
+                    - 0.9
+                    - 0.95
+                eps: 1e-08
+                foreach: false
+                fused: false
+                lr: 3e-05
+                weight_decay: 0.01
+            name: torch.optim.AdamW
+        precision: bfloat16
+        scheduler:
+            - kwargs:
+                factor: 1
+                total_iters: 10000000000
+              name: torch.optim.lr_scheduler.ConstantLR
+            - milestones: []
+        sequence_packing:
+            algorithm: modified_first_fit_decreasing
+            enabled: false
+            logprob_mb_tokens: 10000
+            sequence_length_round: 64
+            train_mb_tokens: 10000
+        tokenizer:
+            chat_template_kwargs:
+                enable_thinking: false
+            name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+        train_global_batch_size: 20
+        train_micro_batch_size: 1
+teacher:
+    value:
+        dtensor_cfg:
+            _v2: true
+            activation_checkpointing: true
+            context_parallel_size: 1
+            cpu_offload: true
+            custom_parallel_plan: null
+            enabled: true
+            sequence_parallel: false
+            tensor_parallel_size: 4
+        dynamic_batching:
+            enabled: true
+            logprob_mb_tokens: 10000
+            sequence_length_round: 64
+            train_mb_tokens: 10000
+        generation:
+            backend: vllm
+            colocated:
+                enabled: true
+                resources:
+                    gpus_per_node: null
+                    num_nodes: null
+            max_new_tokens: 8192
+            stop_strings: null
+            stop_token_ids: null
+            temperature: 1
+            top_k: null
+            top_p: 1
+            vllm_cfg:
+                async_engine: false
+                distributed_executor_backend: null
+                enforce_eager: false
+                expert_parallel_size: 1
+                gpu_memory_utilization: 0.7
+                kv_cache_dtype: auto
+                max_model_len: 10000
+                num_first_layers_in_bf16: 0
+                num_last_layers_in_bf16: 0
+                pipeline_parallel_size: 1
+                precision: bfloat16
+                tensor_parallel_size: 4
+                use_deep_gemm: false
+        generation_batch_size: 16
+        logprob_batch_size: 1
+        logprob_chunk_size: null
+        make_sequence_length_divisible_by: 8
+        max_grad_norm: 1
+        max_total_sequence_length: 10000
+        megatron_cfg:
+            enabled: false
+        model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+        offload_optimizer_for_logprob: false
+        optimizer:
+            kwargs:
+                betas:
+                    - 0.9
+                    - 0.95
+                eps: 1e-08
+                foreach: false
+                fused: false
+                lr: 1e-05
+                weight_decay: 0.01
+            name: torch.optim.AdamW
+        precision: bfloat16
+        scheduler:
+            - kwargs:
+                factor: 1
+                total_iters: 10000000000
+              name: torch.optim.lr_scheduler.ConstantLR
+            - milestones: []
+        sequence_packing:
+            algorithm: modified_first_fit_decreasing
+            enabled: false
+            logprob_mb_tokens: 10000
+            sequence_length_round: 64
+            train_mb_tokens: 10000
+        tokenizer:
+            chat_template_kwargs:
+                enable_thinking: false
+            name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+        train_global_batch_size: 16
+        train_micro_batch_size: 1

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb/wandb/run-20260502_210121-m9xk3p9p/files/output.log

@@ -0,0 +1,562 @@
+Unexpected error during git code logging: Artifact type 'code' is reserved for internal use. Please use a different type.
+No uncommitted changes found
+Error during git operations: Command '['git', 'diff', 'main']' returned non-zero exit status 128.
+Initialized WandbLogger for project mix-teachers-opsd, run offpsd-knowledge-nll-gt-d5-e10-n1-t0-20260502-Llama-3.2-3B-Instruct at /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb
+Initialized TensorboardLogger at /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/tensorboard
+GPU monitoring started with collection interval=10s, flush interval=10s
+  ✓ Training dataloader loaded with 100 samples
+  ✓ Validation dataloader loaded with 193 samples
+
+▶ Setting up compute cluster...
+  ✓ Ray cluster initialized with 1 nodes
+Initializing vllm_policy workers: 100%|██████████| 4/4 [00:13<00:00,  3.39s/worker]
+(IsolatedWorkerInitializer pid=9617) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(IsolatedWorkerInitializer pid=9616) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(IsolatedWorkerInitializer pid=9619) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(IsolatedWorkerInitializer pid=9618) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+  ✓ 4 workers initialized in 13.56s
+(pid=10124) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(pid=10122) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(pid=10123) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(pid=10125) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(VllmGenerationWorker pid=10124) VLLM_RAY_BUNDLE_INDICES environment variable set to: 0,1,2,3
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [utils.py:253] non-default args: {'served_model_name': '/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct', 'skip_tokenizer_init': True, 'trust_remote_code': True, 'load_format': 'dummy', 'dtype': 'bfloat16', 'seed': 0, 'max_model_len': 10000, 'distributed_executor_backend': 'ray', 'tensor_parallel_size': 4, 'enable_prefix_caching': True, 'gpu_memory_utilization': 0.7, 'logprobs_mode': 'processed_logprobs', 'worker_extension_cls': 'nemo_rl.models.generation.vllm.vllm_backend.VllmInternalWorkerExtension', 'enable_sleep_mode': True, 'model': '/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct'}
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [model.py:631] Resolved architecture: LlamaForCausalLM
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [model.py:1745] Using max model len 10000
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [arg_utils.py:1443] Using ray runtime env (env vars redacted): {'py_executable': '/home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3', 'env_vars': {'AI_AGENT': '***', 'AVAILABLE_ADDR_LIST': '***', 'AVAILABLE_PORT_LIST': '***', 'BASH_FUNC__module_raw%%': '***', 'BASH_FUNC_ml%%': '***', 'BASH_FUNC_module%%': '***', 'BASH_FUNC_scl%%': '***', 'BASH_FUNC_switchml%%': '***', 'BASH_FUNC_which%%': '***', 'CLAUDECODE': '***', 'CLAUDE_CODE_ENTRYPOINT': '***', 'CLAUDE_CODE_EXECPATH': '***', 'CONDA_DEFAULT_ENV': '***', 'CONDA_EXE': '***', 'CONDA_PREFIX': '***', 'CONDA_PREFIX_1': '***', 'CONDA_PROMPT_MODIFIER': '***', 'CONDA_PYTHON_EXE': '***', 'CONDA_SHLVL': '***', 'COREPACK_ENABLE_AUTO_PIN': '***', 'CPATH': '***', 'CPATH_modshare': '***', 'CUDA_DIR': '***', 'CUDA_HOME': '***', 'CUDA_INSTALL_PATH': '***', 'CUDA_MPS_LOG_DIRECTORY': '***', 'CUDA_MPS_PIPE_DIRECTORY': '***', 'CUDA_PATH': '***', 'CUDA_TOP': '***', 'CYCLECLOUD_BOOTSTRAP': '***', 'CYCLECLOUD_HOME': '***', 'C_INCLUDE_PATH': '***', 'C_INCLUDE_PATH_modshare': '***', 'DBUS_SESSION_BUS_ADDRESS': '***', 'DISPLAY': '***', 'FJPNCKPT_CONFIG': '***', 'GDK_BACKEND': '***', 'GENKAI_CACHE_DIR': '***', 'GENKAI_CUDA_NAME': '***', 'GENKAI_CUDA_VER': '***', 'GENKAI_FAST_DIR': '***', 'GIT_EDITOR': '***', 'GLOBUS_GSSAPI_NAME_COMPATIBILITY': '***', 'HF_TOKEN': '***', 'HISTCONTROL': '***', 'HISTSIZE': '***', 'HOME': '***', 'HOSTNAME': '***', 'HYDRA_FULL_ERROR': '***', 'JOBID': '***', 'KMP_DUPLICATE_LIB_OK': '***', 'KMP_INIT_AT_FORK': '***', 'LANG': '***', 'LD_LIBRARY_PATH': '***', 'LD_LIBRARY_PATH_modshare': '***', 'LESSOPEN': '***', 'LIBRARY_PATH': '***', 'LIBRARY_PATH_modshare': '***', 'LOADEDMODULES': '***', 'LOADEDMODULES_modshare': '***', 'LOCAL_RANK': '***', 'LOGNAME': '***', 'LS_COLORS': '***', 'LUSTRE_JOBSTAT': '***', 'MAIL': '***', 'MANPATH': '***', 'MASTER_ADDR': '***', 'MASTER_PORT': '***', 'MATHEMATICA_HOME': '***', 'MIG_PARTED_CHECKPOINT_FILE': '***', 'MIG_PARTED_CONFIG_FILE': '***', 'MIG_PARTED_HOOKS_FILE': '***', 'MODULEPATH': '***', 'MODULEPATH_modshare': '***', 'MODULESHOME': '***', 'MODULES_CMD': '***', 'MODULES_LMCONFLICT': '***', 'MODULES_LMCONFLICT_modshare': '***', 'MODULES_RUN_QUARANTINE': '***', 'MOD_GIT_ROOTDIR': '***', 'NCCL_ASYNC_ERROR_HANDLING': '***', 'NCCL_DEBUG': '***', 'NCCL_SHM_USE_CUDA_MEMCPY': '***', 'NCCL_TIMEOUT': '***', 'NEMO_RL_PY_EXECUTABLES_SYSTEM': '***', 'NEMO_RL_VENV_DIR': '***', 'NODE_RANK': '***', 'NoDefaultCurrentDirectoryInExePath': '***', 'OLDPWD': '***', 'OMPI_CKPTCONFIG': '***', 'OTEL_EXPORTER_OTLP_METRICS_TEMPORALITY_PREFERENCE': '***', 'PATH': '***', 'PATH_modshare': '***', 'PJM_ASSIGN_LOGICAL_CPU': '***', 'PJM_CORE_MEM_LIMIT': '***', 'PJM_CUSTOM_RESOURCES': '***', 'PJM_DPREFIX': '***', 'PJM_ELAPSED_TIME_MODE': '***', 'PJM_ELAPSE_LIMIT': '***', 'PJM_ENVIRONMENT': '***', 'PJM_JOBDIR': '***', 'PJM_JOBID': '***', 'PJM_JOBNAME': '***', 'PJM_MAILOPTION': '***', 'PJM_MPI_PROC': '***', 'PJM_NET_ROUTE': '***', 'PJM_NODE': '***', 'PJM_NODE_CPUTIME_LIMIT': '***', 'PJM_O_HOME': '***', 'PJM_O_HOST': '***', 'PJM_O_LANG': '***', 'PJM_O_LOGNAME': '***', 'PJM_O_MAIL': '***', 'PJM_O_NODEINF': '***', 'PJM_O_PATH': '***', 'PJM_O_SHELL': '***', 'PJM_O_WORKDIR': '***', 'PJM_PROC_BY_NODE': '***', 'PJM_RSCGRP': '***', 'PJM_RSCUNIT': '***', 'PJM_SHELL': '***', 'PJM_SSD_DIR': '***', 'PJM_STDERR_PATH': '***', 'PJM_STDOUT_PATH': '***', 'PJM_SUBJOBID': '***', 'PJM_VNODE_MEM_LIMIT': '***', 'PLE_SCRIPT_TYPE': '***', 'PWD': '***', 'PYTHONBREAKPOINT': '***', 'PYTHONPATH': '***', 'PYTHONUNBUFFERED': '***', 'RANK': '***', 'RAY_CLIENT_MODE': '***', 'RAY_DEDUP_LOGS': '***', 'RAY_ENABLE_UV_RUN_RUNTIME_ENV': '***', 'RAY_EXPERIMENTAL_NOSET_CUDA_VISIBLE_DEVICES': '***', 'RAY_TMPDIR': '***', 'RAY_USAGE_STATS_ENABLED': '***', 'RAY_memory_monitor_refresh_ms': '***', 'RAY_worker_heartbeat_timeout_milliseconds': '***', 'RAY_worker_register_timeout_seconds': '***', 'SHLVL': '***', 'SSH_ASKPASS': '***', 'SSH_CLIENT': '***', 'SSH_CONNECTION': '***', 'SSH_TTY': '***', 'SWANLAB_API_HOST': '***', 'SWANLAB_RUNTIME': '***'
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [scheduler.py:216] Chunked prefill is enabled with max_num_batched_tokens=16384.
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [core.py:93] Initializing a V1 LLM engine (v0.11.2) with config: model='/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct', speculative_config=None, tokenizer='/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct', skip_tokenizer_init=True, tokenizer_mode=auto, revision=None, tokenizer_revision=None, trust_remote_code=True, dtype=torch.bfloat16, max_seq_len=10000, download_dir=None, load_format=dummy, tensor_parallel_size=4, pipeline_parallel_size=1, data_parallel_size=1, disable_custom_all_reduce=False, quantization=None, enforce_eager=False, kv_cache_dtype=auto, device_config=cuda, structured_outputs_config=StructuredOutputsConfig(backend='auto', disable_fallback=False, disable_any_whitespace=False, disable_additional_properties=False, reasoning_parser='', reasoning_parser_plugin='', enable_in_reasoning=False), observability_config=ObservabilityConfig(show_hidden_metrics_for_version=None, otlp_traces_endpoint=None, collect_detailed_traces=None), seed=0, served_model_name=/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct, enable_prefix_caching=True, enable_chunked_prefill=True, pooler_config=None, compilation_config={'level': None, 'mode': <CompilationMode.VLLM_COMPILE: 3>, 'debug_dump_path': None, 'cache_dir': '', 'compile_cache_save_format': 'binary', 'backend': 'inductor', 'custom_ops': ['none'], 'splitting_ops': ['vllm::unified_attention', 'vllm::unified_attention_with_output', 'vllm::unified_mla_attention', 'vllm::unified_mla_attention_with_output', 'vllm::mamba_mixer2', 'vllm::mamba_mixer', 'vllm::short_conv', 'vllm::linear_attention', 'vllm::plamo2_mamba_mixer', 'vllm::gdn_attention_core', 'vllm::kda_attention', 'vllm::sparse_attn_indexer'], 'compile_mm_encoder': False, 'use_inductor': None, 'compile_sizes': [], 'inductor_compile_config': {'enable_auto_functionalized_v2': False, 'combo_kernels': True, 'benchmark_combo_kernel': True}, 'inductor_passes': {}, 'cudagraph_mode': <CUDAGraphMode.FULL_AND_PIECEWISE: (2, 1)>, 'cudagraph_num_of_warmups': 1, 'cudagraph_capture_sizes': [1, 2, 4, 8, 16, 24, 32, 40, 48, 56, 64, 72, 80, 88, 96, 104, 112, 120, 128, 136, 144, 152, 160, 168, 176, 184, 192, 200, 208, 216, 224, 232, 240, 248, 256, 272, 288, 304, 320, 336, 352, 368, 384, 400, 416, 432, 448, 464, 480, 496, 512], 'cudagraph_copy_inputs': False, 'cudagraph_specialize_lora': True, 'use_inductor_graph_partition': False, 'pass_config': {}, 'max_cudagraph_capture_size': 512, 'local_cache_dir': None}
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [ray_utils.py:343] Ray is already initialized. Skipping Ray initialization.
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [ray_utils.py:374] Using the existing placement group
+(VllmGenerationWorker pid=10124) The argument `trust_remote_code` is to be used with Auto classes. It has no effect here and is ignored.
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:22 [ray_env.py:66] RAY_NON_CARRY_OVER_ENV_VARS from config: set()
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:22 [ray_env.py:69] Copying the following environment variables to workers: ['VLLM_ALLOW_INSECURE_SERIALIZATION', 'CUDA_HOME', 'VLLM_SKIP_P2P_CHECK', 'VLLM_CACHE_ROOT', 'VLLM_RAY_BUNDLE_INDICES', 'VLLM_ENABLE_V1_MULTIPROCESSING', 'RAY_ENABLE_UV_RUN_RUNTIME_ENV', 'HF_TOKEN', 'LD_LIBRARY_PATH', 'VLLM_RAY_PER_WORKER_GPUS']
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:22 [ray_env.py:74] If certain env vars should NOT be copied, add them to /home/pj24003162/ku40003404/.config/vllm/ray_non_carry_over_env_vars.json file
+(RayWorkerWrapper pid=10547) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(RayWorkerWrapper pid=10551) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(RayWorkerWrapper pid=10554) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(RayWorkerWrapper pid=10548) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(RayWorkerWrapper pid=10547) INFO 05-02 21:02:32 [worker_base.py:283] Injected <class 'nemo_rl.models.generation.vllm.vllm_backend.VllmInternalWorkerExtension'> into <class 'vllm.v1.worker.gpu_worker.Worker'> for extended collective_rpc calls ['_maybe_process_fp8_kv_cache', 'cleanup', 'get_zmq_address', 'init_collective', 'maybe_init_zmq', 'prepare_refit_info', 'report_device_id', 'start_gpu_profiling', 'stop_gpu_profiling', 'update_weights_from_collective', 'update_weights_via_ipc_zmq']
+(RayWorkerWrapper pid=10547) WARNING 05-02 21:02:32 [worker_base.py:301] Missing `shared_worker_lock` argument from executor. This argument is needed for mm_processor_cache_type='shm'.
+(RayWorkerWrapper pid=10551) INFO 05-02 21:02:32 [worker_base.py:283] Injected <class 'nemo_rl.models.generation.vllm.vllm_backend.VllmInternalWorkerExtension'> into <class 'vllm.v1.worker.gpu_worker.Worker'> for extended collective_rpc calls ['_maybe_process_fp8_kv_cache', 'cleanup', 'get_zmq_address', 'init_collective', 'maybe_init_zmq', 'prepare_refit_info', 'report_device_id', 'start_gpu_profiling', 'stop_gpu_profiling', 'update_weights_from_collective', 'update_weights_via_ipc_zmq']
+(RayWorkerWrapper pid=10551) WARNING 05-02 21:02:32 [worker_base.py:301] Missing `shared_worker_lock` argument from executor. This argument is needed for mm_processor_cache_type='shm'.
+(RayWorkerWrapper pid=10554) INFO 05-02 21:02:32 [worker_base.py:283] Injected <class 'nemo_rl.models.generation.vllm.vllm_backend.VllmInternalWorkerExtension'> into <class 'vllm.v1.worker.gpu_worker.Worker'> for extended collective_rpc calls ['_maybe_process_fp8_kv_cache', 'cleanup', 'get_zmq_address', 'init_collective', 'maybe_init_zmq', 'prepare_refit_info', 'report_device_id', 'start_gpu_profiling', 'stop_gpu_profiling', 'update_weights_from_collective', 'update_weights_via_ipc_zmq']
+(RayWorkerWrapper pid=10554) WARNING 05-02 21:02:32 [worker_base.py:301] Missing `shared_worker_lock` argument from executor. This argument is needed for mm_processor_cache_type='shm'.
+(RayWorkerWrapper pid=10548) INFO 05-02 21:02:32 [worker_base.py:283] Injected <class 'nemo_rl.models.generation.vllm.vllm_backend.VllmInternalWorkerExtension'> into <class 'vllm.v1.worker.gpu_worker.Worker'> for extended collective_rpc calls ['_maybe_process_fp8_kv_cache', 'cleanup', 'get_zmq_address', 'init_collective', 'maybe_init_zmq', 'prepare_refit_info', 'report_device_id', 'start_gpu_profiling', 'stop_gpu_profiling', 'update_weights_from_collective', 'update_weights_via_ipc_zmq']
+(RayWorkerWrapper pid=10548) WARNING 05-02 21:02:32 [worker_base.py:301] Missing `shared_worker_lock` argument from executor. This argument is needed for mm_processor_cache_type='shm'.
+(RayWorkerWrapper pid=10547) INFO 05-02 21:02:34 [parallel_state.py:1208] world_size=4 rank=1 local_rank=1 distributed_init_method=tcp://127.0.0.1:48355 backend=nccl
+(RayWorkerWrapper pid=10551) INFO 05-02 21:02:34 [parallel_state.py:1208] world_size=4 rank=0 local_rank=0 distributed_init_method=tcp://127.0.0.1:48355 backend=nccl
+(RayWorkerWrapper pid=10554) INFO 05-02 21:02:34 [parallel_state.py:1208] world_size=4 rank=2 local_rank=2 distributed_init_method=tcp://127.0.0.1:48355 backend=nccl
+(RayWorkerWrapper pid=10548) INFO 05-02 21:02:34 [parallel_state.py:1208] world_size=4 rank=3 local_rank=3 distributed_init_method=tcp://127.0.0.1:48355 backend=nccl
+(RayWorkerWrapper pid=10547) [Gloo] Rank 1 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10551) [Gloo] Rank 0 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10554) [Gloo] Rank 2 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10548) [Gloo] Rank 3 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10547) [Gloo] Rank 1 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10551) [Gloo] Rank 0 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10551) INFO 05-02 21:02:34 [pynccl.py:111] vLLM is using nccl==2.27.5
+(RayWorkerWrapper pid=10551) NCCL version 2.27.5+cuda12.9
+(RayWorkerWrapper pid=10554) [Gloo] Rank 2 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10548) [Gloo] Rank 3 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10547) WARNING 05-02 21:02:38 [symm_mem.py:107] SymmMemCommunicator: symmetric memory multicast operations are not supported.
+(RayWorkerWrapper pid=10551) WARNING 05-02 21:02:38 [symm_mem.py:107] SymmMemCommunicator: symmetric memory multicast operations are not supported.
+(RayWorkerWrapper pid=10554) WARNING 05-02 21:02:38 [symm_mem.py:107] SymmMemCommunicator: symmetric memory multicast operations are not supported.
+(RayWorkerWrapper pid=10548) WARNING 05-02 21:02:38 [symm_mem.py:107] SymmMemCommunicator: symmetric memory multicast operations are not supported.
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342] Error executing method 'init_device'. This might cause deadlock in distributed execution.
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342] Traceback (most recent call last):
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     return run_method(self, method, args, kwargs)
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/util/tracing/tracing_helper.py", line 461, in _resume_span
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     return method(self, *_args, **_kwargs)
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     self.worker.init_device()  # type: ignore
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     init_worker_distributed_environment(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     ensure_model_parallel_initialized(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     initialize_model_parallel(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     _TP = init_model_parallel_group(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     return GroupCoordinator(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 707, in create_from_process_group
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     buffer_io = MessageQueue(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]                 ^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 281, in __init__
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     self.local_socket.bind(local_subscribe_addr)
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/zmq/sugar/socket.py", line 320, in bind
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     super().bind(addr)
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "zmq/backend/cython/_zmq.py", line 1002, in zmq.backend.cython._zmq.Socket.bind
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342] zmq.error.ZMQError: ipc path "/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355" is longer than 107 characters (sizeof(sockaddr_un.sun_path)). zmq.IPC_PATH_MAX_LEN constant can be used to check addr length (if it is defined). (addr='ipc:///tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355')
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:39 [ray_executor.py:121] Shutting down Ray distributed executor. If you see error log from logging.cc regarding SIGTERM received, please ignore because this is the expected termination process in Ray.
+(VllmGenerationWorker pid=10124) Exception raised in creation task: The actor died because of an error raised in its creation task, ray::vllm_policy-grp0-0:VllmGenerationWorker.__init__() (pid=10124, ip=172.16.17.20, actor_id=788c98f21211d983db82fd8401000000, repr=VllmGenerationWorker)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 462, in __init__
+(VllmGenerationWorker pid=10124)     self._create_engine(llm_kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 579, in _create_engine
+(VllmGenerationWorker pid=10124)     self.llm = vllm.LLM(**llm_kwargs)
+(VllmGenerationWorker pid=10124)                ^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/entrypoints/llm.py", line 343, in __init__
+(VllmGenerationWorker pid=10124)     self.llm_engine = LLMEngine.from_engine_args(
+(VllmGenerationWorker pid=10124)                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 174, in from_engine_args
+(VllmGenerationWorker pid=10124)     return cls(
+(VllmGenerationWorker pid=10124)            ^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 108, in __init__
+(VllmGenerationWorker pid=10124)     self.engine_core = EngineCoreClient.make_client(
+(VllmGenerationWorker pid=10124)                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 95, in make_client
+(VllmGenerationWorker pid=10124)     return InprocClient(vllm_config, executor_class, log_stats)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 264, in __init__
+(VllmGenerationWorker pid=10124)     self.engine_core = EngineCore(*args, **kwargs)
+(VllmGenerationWorker pid=10124)                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core.py", line 102, in __init__
+(VllmGenerationWorker pid=10124)     self.model_executor = executor_class(vllm_config)
+(VllmGenerationWorker pid=10124)                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/abstract.py", line 101, in __init__
+(VllmGenerationWorker pid=10124)     self._init_executor()
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 97, in _init_executor
+(VllmGenerationWorker pid=10124)     self._init_workers_ray(placement_group, runtime_env={"py_executable": "/home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3"})
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 370, in _init_workers_ray
+(VllmGenerationWorker pid=10124)     self.collective_rpc("init_device")
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 493, in collective_rpc
+(VllmGenerationWorker pid=10124)     return ray.get(ray_worker_outputs, timeout=timeout)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)                                   ^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124) ray.exceptions.RayTaskError(ZMQError): ray::RayWorkerWrapper.execute_method() (pid=10551, ip=172.16.17.20, actor_id=503a40dba590cd200475f83501000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x1518ed3f8490>)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+(VllmGenerationWorker pid=10124)     raise e
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(VllmGenerationWorker pid=10124)     return run_method(self, method, args, kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(VllmGenerationWorker pid=10124)     self.worker.init_device()  # type: ignore
+(VllmGenerationWorker pid=10124)     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(VllmGenerationWorker pid=10124)     init_worker_distributed_environment(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(VllmGenerationWorker pid=10124)     ensure_model_parallel_initialized(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(VllmGenerationWorker pid=10124)     initialize_model_parallel(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(VllmGenerationWorker pid=10124)     _TP = init_model_parallel_group(
+(VllmGenerationWorker pid=10124)           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(VllmGenerationWorker pid=10124)     return GroupCoordinator(
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(VllmGenerationWorker pid=10124)     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(VllmGenerationWorker pid=10124)                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 707, in create_from_process_group
+(VllmGenerationWorker pid=10124)     buffer_io = MessageQueue(
+(VllmGenerationWorker pid=10124)                 ^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 281, in __init__
+(VllmGenerationWorker pid=10124)     self.local_socket.bind(local_subscribe_addr)
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/zmq/sugar/socket.py", line 320, in bind
+(VllmGenerationWorker pid=10124)     super().bind(addr)
+(VllmGenerationWorker pid=10124)   File "zmq/backend/cython/_zmq.py", line 1002, in zmq.backend.cython._zmq.Socket.bind
+(VllmGenerationWorker pid=10124) zmq.error.ZMQError: ipc path "/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355" is longer than 107 characters (sizeof(sockaddr_un.sun_path)). zmq.IPC_PATH_MAX_LEN constant can be used to check addr length (if it is defined). (addr='ipc:///tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355')
+(RayWorkerWrapper pid=10551) [rank0]:[W502 21:02:40.774631326 ProcessGroupNCCL.cpp:1524] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+(VllmGenerationWorker pid=10124) Unhandled error (suppress with 'RAY_IGNORE_UNHANDLED_ERRORS=1'): ray::RayWorkerWrapper.execute_method() (pid=10554, ip=172.16.17.20, actor_id=3a062de3b6335cf114e4b89b01000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x14d9739f7c90>)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+(VllmGenerationWorker pid=10124)     raise e
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(VllmGenerationWorker pid=10124)     return run_method(self, method, args, kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(VllmGenerationWorker pid=10124)     self.worker.init_device()  # type: ignore
+(VllmGenerationWorker pid=10124)     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(VllmGenerationWorker pid=10124)     init_worker_distributed_environment(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(VllmGenerationWorker pid=10124)     ensure_model_parallel_initialized(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(VllmGenerationWorker pid=10124)     initialize_model_parallel(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(VllmGenerationWorker pid=10124)     _TP = init_model_parallel_group(
+(VllmGenerationWorker pid=10124)           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(VllmGenerationWorker pid=10124)     return GroupCoordinator(
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(VllmGenerationWorker pid=10124)     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(VllmGenerationWorker pid=10124)                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 724, in create_from_process_group
+(VllmGenerationWorker pid=10124)     dist.broadcast_object_list(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 3691, in broadcast_object_list
+(VllmGenerationWorker pid=10124)     broadcast(object_sizes_tensor, group_src=group_src, group=group)
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 2841, in broadcast
+(VllmGenerationWorker pid=10124)     work.wait()
+(VllmGenerationWorker pid=10124) RuntimeError: [/pytorch/third_party/gloo/gloo/transport/tcp/pair.cc:547] Connection closed by peer [172.16.17.20]:27801
+(VllmGenerationWorker pid=10124) Unhandled error (suppress with 'RAY_IGNORE_UNHANDLED_ERRORS=1'): ray::RayWorkerWrapper.execute_method() (pid=10547, ip=172.16.17.20, actor_id=afe18c46590e3690d9e9625e01000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x14338cc28f10>)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+(VllmGenerationWorker pid=10124)     raise e
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(VllmGenerationWorker pid=10124)     return run_method(self, method, args, kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(VllmGenerationWorker pid=10124)     self.worker.init_device()  # type: ignore
+(VllmGenerationWorker pid=10124)     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(VllmGenerationWorker pid=10124)     init_worker_distributed_environment(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(VllmGenerationWorker pid=10124)     ensure_model_parallel_initialized(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(VllmGenerationWorker pid=10124)     initialize_model_parallel(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(VllmGenerationWorker pid=10124)     _TP = init_model_parallel_group(
+(VllmGenerationWorker pid=10124)           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(VllmGenerationWorker pid=10124)     return GroupCoordinator(
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(VllmGenerationWorker pid=10124)     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(VllmGenerationWorker pid=10124)                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 724, in create_from_process_group
+(VllmGenerationWorker pid=10124)     dist.broadcast_object_list(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 3691, in broadcast_object_list
+(VllmGenerationWorker pid=10124)     broadcast(object_sizes_tensor, group_src=group_src, group=group)
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 2841, in broadcast
+(VllmGenerationWorker pid=10124)     work.wait()
+(VllmGenerationWorker pid=10124) RuntimeError: [/pytorch/third_party/gloo/gloo/transport/tcp/pair.cc:547] Connection closed by peer [172.16.17.20]:35480
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342] Error executing method 'init_device'. This might cause deadlock in distributed execution.
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342] Traceback (most recent call last):
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return run_method(self, method, args, kwargs)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/util/tracing/tracing_helper.py", line 461, in _resume_span
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return method(self, *_args, **_kwargs)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     self.worker.init_device()  # type: ignore
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     init_worker_distributed_environment(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     ensure_model_parallel_initialized(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     initialize_model_parallel(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     _TP = init_model_parallel_group(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return GroupCoordinator(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 724, in create_from_process_group
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     dist.broadcast_object_list(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 3691, in broadcast_object_list
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     broadcast(object_sizes_tensor, group_src=group_src, group=group)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 2841, in broadcast
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     work.wait()
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342] RuntimeError: [/pytorch/third_party/gloo/gloo/transport/tcp/pair.cc:547] Connection closed by peer [172.16.17.20]:35480
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342] Error executing method 'init_device'. This might cause deadlock in distributed execution.
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342] Traceback (most recent call last):
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return run_method(self, method, args, kwargs)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/util/tracing/tracing_helper.py", line 461, in _resume_span
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return method(self, *_args, **_kwargs)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     self.worker.init_device()  # type: ignore
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     init_worker_distributed_environment(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     ensure_model_parallel_initialized(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     initialize_model_parallel(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     _TP = init_model_parallel_group(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return GroupCoordinator(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 724, in create_from_process_group
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     dist.broadcast_object_list(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 3691, in broadcast_object_list
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     broadcast(object_sizes_tensor, group_src=group_src, group=group)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 2841, in broadcast
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     work.wait()
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342] RuntimeError: [/pytorch/third_party/gloo/gloo/transport/tcp/pair.cc:547] Connection closed by peer [172.16.17.20]:27801
+Traceback (most recent call last):
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py", line 205, in <module>
+    main()
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py", line 193, in main
+    run_opsd_training(
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/math_operations_opsd/opsd_common.py", line 394, in run_opsd_training
+    ) = setup(config, tokenizer, dataset, val_dataset)
+        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/algorithms/self_distillation.py", line 1038, in setup
+    student_generation = VllmGeneration(
+                         ^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_generation.py", line 235, in __init__
+    self._post_init()
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_generation.py", line 433, in _post_init
+    results = ray.get(futures)
+              ^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/auto_init_hook.py", line 22, in auto_init_wrapper
+    return fn(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/client_mode_hook.py", line 104, in wrapper
+    return func(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/worker.py", line 2981, in get
+    values, debugger_breakpoint = worker.get_objects(
+                                  ^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/worker.py", line 1014, in get_objects
+    raise value
+ray.exceptions.ActorDiedError: The actor died because of an error raised in its creation task, ray::vllm_policy-grp0-0:VllmGenerationWorker.__init__() (pid=10124, ip=172.16.17.20, actor_id=788c98f21211d983db82fd8401000000, repr=VllmGenerationWorker)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 462, in __init__
+    self._create_engine(llm_kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 579, in _create_engine
+    self.llm = vllm.LLM(**llm_kwargs)
+               ^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/entrypoints/llm.py", line 343, in __init__
+    self.llm_engine = LLMEngine.from_engine_args(
+                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 174, in from_engine_args
+    return cls(
+           ^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 108, in __init__
+    self.engine_core = EngineCoreClient.make_client(
+                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 95, in make_client
+    return InprocClient(vllm_config, executor_class, log_stats)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 264, in __init__
+    self.engine_core = EngineCore(*args, **kwargs)
+                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core.py", line 102, in __init__
+    self.model_executor = executor_class(vllm_config)
+                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/abstract.py", line 101, in __init__
+    self._init_executor()
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 97, in _init_executor
+    self._init_workers_ray(placement_group, runtime_env={"py_executable": "/home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3"})
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 370, in _init_workers_ray
+    self.collective_rpc("init_device")
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 493, in collective_rpc
+    return ray.get(ray_worker_outputs, timeout=timeout)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^
+                                  ^^^^^^^^^^^^^^^^^^^
+ray.exceptions.RayTaskError(ZMQError): ray::RayWorkerWrapper.execute_method() (pid=10551, ip=172.16.17.20, actor_id=503a40dba590cd200475f83501000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x1518ed3f8490>)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+    raise e
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+    return run_method(self, method, args, kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+    return func(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+    self.worker.init_device()  # type: ignore
+    ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+    init_worker_distributed_environment(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+    ensure_model_parallel_initialized(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+    initialize_model_parallel(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+    _TP = init_model_parallel_group(
+          ^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+    return GroupCoordinator(
+           ^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+    self.mq_broadcaster = MessageQueue.create_from_process_group(
+                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 707, in create_from_process_group
+    buffer_io = MessageQueue(
+                ^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 281, in __init__
+    self.local_socket.bind(local_subscribe_addr)
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/zmq/sugar/socket.py", line 320, in bind
+    super().bind(addr)
+  File "zmq/backend/cython/_zmq.py", line 1002, in zmq.backend.cython._zmq.Socket.bind
+zmq.error.ZMQError: ipc path "/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355" is longer than 107 characters (sizeof(sockaddr_un.sun_path)). zmq.IPC_PATH_MAX_LEN constant can be used to check addr length (if it is defined). (addr='ipc:///tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355')
+Traceback (most recent call last):
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py", line 205, in <module>
+    main()
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py", line 193, in main
+    run_opsd_training(
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/math_operations_opsd/opsd_common.py", line 394, in run_opsd_training
+    ) = setup(config, tokenizer, dataset, val_dataset)
+        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/algorithms/self_distillation.py", line 1038, in setup
+    student_generation = VllmGeneration(
+                         ^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_generation.py", line 235, in __init__
+    self._post_init()
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_generation.py", line 433, in _post_init
+    results = ray.get(futures)
+              ^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/auto_init_hook.py", line 22, in auto_init_wrapper
+    return fn(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/client_mode_hook.py", line 104, in wrapper
+    return func(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/worker.py", line 2981, in get
+    values, debugger_breakpoint = worker.get_objects(
+                                  ^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/worker.py", line 1014, in get_objects
+    raise value
+ray.exceptions.ActorDiedError: The actor died because of an error raised in its creation task, ray::vllm_policy-grp0-0:VllmGenerationWorker.__init__() (pid=10124, ip=172.16.17.20, actor_id=788c98f21211d983db82fd8401000000, repr=VllmGenerationWorker)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 462, in __init__
+    self._create_engine(llm_kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 579, in _create_engine
+    self.llm = vllm.LLM(**llm_kwargs)
+               ^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/entrypoints/llm.py", line 343, in __init__
+    self.llm_engine = LLMEngine.from_engine_args(
+                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 174, in from_engine_args
+    return cls(
+           ^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 108, in __init__
+    self.engine_core = EngineCoreClient.make_client(
+                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 95, in make_client
+    return InprocClient(vllm_config, executor_class, log_stats)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 264, in __init__
+    self.engine_core = EngineCore(*args, **kwargs)
+                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core.py", line 102, in __init__
+    self.model_executor = executor_class(vllm_config)
+                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/abstract.py", line 101, in __init__
+    self._init_executor()
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 97, in _init_executor
+    self._init_workers_ray(placement_group, runtime_env={"py_executable": "/home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3"})
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 370, in _init_workers_ray
+    self.collective_rpc("init_device")
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 493, in collective_rpc
+    return ray.get(ray_worker_outputs, timeout=timeout)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^
+                                  ^^^^^^^^^^^^^^^^^^^
+ray.exceptions.RayTaskError(ZMQError): ray::RayWorkerWrapper.execute_method() (pid=10551, ip=172.16.17.20, actor_id=503a40dba590cd200475f83501000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x1518ed3f8490>)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+    raise e
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+    return run_method(self, method, args, kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+    return func(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+    self.worker.init_device()  # type: ignore
+    ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+    init_worker_distributed_environment(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+    ensure_model_parallel_initialized(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+    initialize_model_parallel(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+    _TP = init_model_parallel_group(
+          ^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+    return GroupCoordinator(
+           ^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+    self.mq_broadcaster = MessageQueue.create_from_process_group(
+                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 707, in create_from_process_group
+    buffer_io = MessageQueue(
+                ^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 281, in __init__
+    self.local_socket.bind(local_subscribe_addr)
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/zmq/sugar/socket.py", line 320, in bind
+    super().bind(addr)
+  File "zmq/backend/cython/_zmq.py", line 1002, in zmq.backend.cython._zmq.Socket.bind
+zmq.error.ZMQError: ipc path "/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355" is longer than 107 characters (sizeof(sockaddr_un.sun_path)). zmq.IPC_PATH_MAX_LEN constant can be used to check addr length (if it is defined). (addr='ipc:///tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355')

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb/wandb/run-20260502_210121-m9xk3p9p/files/requirements.txt

@@ -0,0 +1,290 @@
+colorama==0.4.6
+psutil==7.2.2
+hydra-core==1.3.2
+tensorboard==2.20.0
+dnspython==2.8.0
+mdurl==0.1.2
+httpx==0.28.1
+ninja==1.13.0
+nvidia-ml-py==13.595.45
+google-api-core==2.30.3
+mistral_common==1.11.0
+mlflow-skinny==3.11.1
+python-multipart==0.0.27
+lark==1.2.2
+python-dotenv==1.2.2
+huggingface_hub==0.36.2
+cupy-cuda12x==14.0.1
+openai-harmony==0.0.8
+math-verify==0.9.0
+pyasn1_modules==0.4.2
+Mako==1.3.12
+email-validator==2.3.0
+pyparsing==3.3.2
+xformers==0.0.33.post1
+nvidia-nvjitlink-cu12==12.8.93
+dill==0.4.1
+graphql-core==3.2.8
+graphql-relay==3.2.0
+jiter==0.14.0
+sentencepiece==0.2.1
+msgpack==1.1.2
+multiprocess==0.70.19
+click==8.3.3
+nvidia-cufft-cu12==11.3.3.83
+protobuf==6.33.6
+interegular==0.3.3
+asttokens==3.0.1
+cuda-bindings==13.2.0
+blake3==1.0.8
+py-spy==0.4.2
+python-json-logger==4.1.0
+outlines_core==0.2.11
+openai==2.32.0
+contourpy==1.3.3
+mpmath==1.3.0
+annotated-doc==0.0.4
+nvidia-nccl-cu12==2.27.5
+distro==1.9.0
+wandb==0.25.1
+aiohttp==3.13.5
+opentelemetry-semantic-conventions==0.62b1
+pyecharts==2.1.0
+aiohttp-cors==0.8.1
+triton==3.5.0
+kiwisolver==1.5.0
+shellingham==1.5.4
+einops==0.8.2
+setproctitle==1.3.7
+wrapt==2.1.2
+aiohappyeyeballs==2.6.1
+annotated-types==0.7.0
+nvidia-cuda-nvrtc-cu12==12.8.93
+watchfiles==1.1.1
+nvidia-cuda-cupti-cu12==12.8.90
+scipy==1.17.1
+matplotlib==3.10.8
+stack-data==0.6.3
+absl-py==2.4.0
+antlr4-python3-runtime==4.9.3
+tzdata==2026.2
+distlib==0.4.0
+comm==0.2.3
+parso==0.8.6
+tensorboard-data-server==0.7.2
+ptyprocess==0.7.0
+opencv-python-headless==4.13.0.92
+msgspec==0.21.1
+uvicorn==0.46.0
+nvidia-cuda-runtime-cu12==12.8.90
+datasets==4.8.4
+requests==2.33.1
+Flask==3.1.3
+fastar==0.11.0
+executing==2.2.1
+Werkzeug==3.1.8
+nvidia-cutlass-dsl==4.4.2
+blobfile==3.2.0
+decord==0.6.0
+jsonschema==4.26.0
+llvmlite==0.44.0
+uvloop==0.22.1
+nvidia-cusparselt-cu12==0.7.1
+model-hosting-container-standards==0.1.14
+databricks-sdk==0.105.0
+nest-asyncio==1.6.0
+sentry-sdk==2.58.0
+cuda-python==13.2.0
+wcwidth==0.6.0
+Jinja2==3.1.6
+depyf==0.20.0
+nvidia-cufile-cu12==1.13.1.3
+colorful==0.5.8
+loguru==0.7.3
+blinker==1.9.0
+debugpy==1.8.20
+pycryptodomex==3.23.0
+regex==2026.4.4
+nvidia-cutlass-dsl-libs-base==4.4.2
+jedi==0.19.2
+python-discovery==1.2.2
+pydantic-extra-types==2.11.1
+mlflow-tracing==3.11.1
+opentelemetry-sdk==1.41.1
+anyio==4.13.0
+GitPython==3.1.48
+setuptools==82.0.1
+PyYAML==6.0.3
+starlette==0.52.1
+multidict==6.7.1
+attrs==26.1.0
+pip==26.0.1
+sympy==1.14.0
+pyasn1==0.6.3
+sqlparse==0.5.5
+opentelemetry-proto==1.41.1
+astor==0.8.1
+peft==0.18.1
+torchdata==0.11.0
+compressed-tensors==0.12.2
+Pygments==2.20.0
+transformers==4.57.1
+cryptography==46.0.7
+plotly==6.7.0
+typing-inspection==0.4.2
+flashinfer-python==0.5.2
+nvtx==0.2.15
+ipython_pygments_lexers==1.1.1
+rpds-py==0.30.0
+pillow==12.2.0
+accelerate==1.13.0
+yarl==1.23.0
+grpcio==1.80.0
+numba==0.61.2
+omegaconf==2.3.0
+googleapis-common-protos==1.74.0
+joblib==1.5.3
+graphene==3.4.3
+num2words==0.5.14
+docstring_parser==0.18.0
+referencing==0.37.0
+lxml==6.1.0
+smmap==5.0.3
+jsonschema-specifications==2025.9.1
+MarkupSafe==3.0.3
+opencensus-context==0.1.3
+charset-normalizer==3.4.7
+torchaudio==2.9.0
+nvidia-cusolver-cu12==11.7.3.90
+safetensors==0.7.0
+sniffio==1.3.1
+gguf==0.18.0
+threadpoolctl==3.6.0
+filelock==3.29.0
+pexpect==4.9.0
+networkx==3.6.1
+idna==3.13
+aiosignal==1.4.0
+fastapi-cloud-cli==0.17.1
+py-cpuinfo==9.0.0
+skops==0.14.0
+nvidia-cusparse-cu12==12.5.8.93
+flask-cors==6.0.2
+decorator==5.2.1
+nvidia-nvtx-cu12==12.8.90
+importlib_metadata==8.7.1
+rich==13.9.4
+opentelemetry-exporter-prometheus==0.62b1
+fastapi-cli==0.0.24
+pybase64==1.4.3
+pydantic-settings==2.14.0
+nvidia-cublas-cu12==12.8.4.1
+wheel==0.47.0
+smart_open==7.6.0
+xgrammar==0.1.25
+llguidance==1.3.0
+jupyter_client==8.8.0
+mlflow==3.11.1
+fastapi==0.136.1
+jmespath==1.1.0
+ipython==9.13.0
+colored==2.2.3
+pycparser==3.0
+anthropic==0.71.0
+pyzmq==27.1.0
+proto-plus==1.27.2
+prompt_toolkit==3.0.52
+python-dateutil==2.9.0.post0
+SQLAlchemy==2.0.49
+psutil==7.2.2
+urllib3==2.6.3
+httptools==0.7.1
+google-auth==2.49.2
+tornado==6.5.5
+gitdb==4.0.12
+cachetools==7.0.6
+narwhals==2.20.0
+rignore==0.7.6
+latex2sympy2_extended==1.11.0
+tabulate==0.10.0
+pytz==2026.1.post1
+prometheus_client==0.25.0
+h11==0.16.0
+partial-json-parser==0.2.1.1.post7
+gunicorn==25.3.0
+torch==2.9.0
+cloudpickle==3.1.2
+fonttools==4.62.1
+tiktoken==0.12.0
+scikit-learn==1.8.0
+cuda-pathfinder==1.5.4
+greenlet==3.5.0
+pycountry==26.2.16
+tokenizers==0.22.2
+propcache==0.4.1
+nvidia-cudnn-frontend==1.22.1
+virtualenv==21.3.0
+zipp==3.23.1
+ray==2.54.1
+alembic==1.18.4
+docker==7.1.0
+opentelemetry-api==1.41.1
+swanlab==0.7.16
+packaging==26.2
+diskcache==5.6.3
+numpy==2.2.6
+vllm==0.11.2
+typer==0.25.0
+pandas==2.3.3
+fsspec==2026.2.0
+tqdm==4.67.3
+nvidia-cudnn-cu12==9.10.2.21
+torchvision==0.24.0
+cycler==0.12.1
+nvidia-nvshmem-cu12==3.3.20
+typing_extensions==4.15.0
+hf-xet==1.4.3
+pyarrow==23.0.1
+prometheus-fastapi-instrumentator==7.1.0
+traitlets==5.14.3
+itsdangerous==2.2.0
+platformdirs==4.9.6
+cffi==2.0.0
+opencensus==0.11.4
+ipykernel==7.2.0
+lm-format-enforcer==0.11.3
+certifi==2026.4.22
+supervisor==4.3.0
+pydantic_core==2.46.3
+markdown-it-py==4.0.0
+httpcore==1.0.9
+docopt==0.6.2
+matplotlib-inline==0.2.1
+Markdown==3.10.2
+six==1.17.0
+cbor2==5.9.0
+jupyter_core==5.9.1
+pydantic==2.13.3
+frozenlist==1.8.0
+websockets==16.0
+rich-toolkit==0.19.7
+xxhash==3.7.0
+simplejson==4.1.1
+hf_transfer==0.1.9
+huey==2.6.0
+pure_eval==0.2.3
+apache-tvm-ffi==0.1.10
+nvidia-curand-cu12==10.3.9.90
+prettytable==3.17.0
+zipp==3.23.0
+tomli==2.4.0
+jaraco.text==4.0.0
+backports.tarfile==1.2.0
+more-itertools==10.8.0
+packaging==26.0
+autocommand==2.2.2
+platformdirs==4.4.0
+importlib_metadata==8.7.1
+wheel==0.46.3
+jaraco.context==6.1.0
+jaraco.functools==4.4.0

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb/wandb/run-20260502_210121-m9xk3p9p/files/wandb-metadata.json

@@ -0,0 +1,72 @@
+{
+  "os": "Linux-4.18.0-477.27.1.el8_8.x86_64-x86_64-with-glibc2.28",
+  "python": "CPython 3.11.15",
+  "startedAt": "2026-05-02T12:01:21.495560Z",
+  "args": [
+    "--config",
+    "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/opsd_distillation_config.yaml",
+    "policy.model_name=/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct",
+    "policy.dtensor_cfg.tensor_parallel_size=4",
+    "policy.dtensor_cfg.context_parallel_size=1",
+    "policy.max_total_sequence_length=10000",
+    "policy.generation.max_new_tokens=8192",
+    "policy.generation.vllm_cfg.tensor_parallel_size=4",
+    "policy.generation.vllm_cfg.gpu_memory_utilization=0.7",
+    "policy.logprob_batch_size=1",
+    "policy.dtensor_cfg.cpu_offload=true",
+    "policy.dtensor_cfg.activation_checkpointing=true",
+    "paths.base_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd",
+    "paths.data_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/data/knowledge_d5_e10/atomic_sft",
+    "checkpointing.checkpoint_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927",
+    "logger.log_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs",
+    "logger.wandb.project=mix-teachers-opsd",
+    "logger.wandb.name=offpsd-knowledge-nll-gt-d5-e10-n1-t0-20260502-Llama-3.2-3B-Instruct",
+    "logger.tensorboard.log_dir=tb_logs-offpsd-knowledge-nll-gt-d5-e10-llama3-2-3b-n1-t0",
+    "cluster.gpus_per_node=4",
+    "cluster.num_nodes=1",
+    "loss_fn.kl_type=forward",
+    "loss_fn.zero_outside_topk=false",
+    "distillation.ema_decay=1.0",
+    "distillation.topk_logits_k=64",
+    "distillation.num_prompts_per_step=20",
+    "distillation.num_generations_per_prompt=1",
+    "distillation.rollout_policy=teacher",
+    "+distillation.use_nll_loss=true",
+    "distillation.val_period=20",
+    "distillation.max_num_epochs=40",
+    "checkpointing.save_per_epoch=false",
+    "checkpointing.save_period=20",
+    "checkpointing.save_optimizer=false",
+    "checkpointing.keep_top_k=null",
+    "checkpointing.metric_name=null",
+    "policy.train_global_batch_size=20",
+    "policy.generation_batch_size=20",
+    "policy.generation.temperature=0.0",
+    "policy.optimizer.kwargs.lr=3.0e-5",
+    "policy.optimizer.kwargs.weight_decay=0.01",
+    "policy.scheduler=[{name: \"torch.optim.lr_scheduler.ConstantLR\", kwargs: {factor: 1.0, total_iters: 10000000000}}, {milestones: []}]"
+  ],
+  "program": "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py",
+  "codePath": "mix_teachers/train/knowledge_opsd/run_opsd.py",
+  "codePathLocal": "run_opsd.py",
+  "git": {
+    "remote": "https://github.com/jiarui-liu/TeacherMix.git",
+    "commit": "9938bfa2b66c98512427baa8c5adbea81dbf6dc8"
+  },
+  "email": "jiaruil5@andrew.cmu.edu",
+  "root": "/home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb",
+  "host": "b0020",
+  "executable": "/home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3",
+  "cpu_count": 120,
+  "cpu_count_logical": 120,
+  "disk": {
+    "/": {
+      "total": "511749718016",
+      "used": "24012599296"
+    }
+  },
+  "memory": {
+    "total": "1081436749824"
+  },
+  "writerId": "4is3svrhq809lwa4dst4qbj79auqgiq9"
+}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb/wandb/run-20260502_210121-m9xk3p9p/files/wandb-summary.json

@@ -0,0 +1 @@
+{"ray/node.0.gpu.1.util":0,"ray/node.0.gpu.1.mem_gb":0.5,"_wandb":{"runtime":68},"ray/node.0.mem_total_gb":1007.1664581298828,"ray/node.0.gpu.2.util":0,"ray/node.0.gpu.2.mem_gb":0.5,"_step":0,"ray/ray_step":61,"_timestamp":1.777723355391118e+09,"_runtime":68.686869346,"ray/node.0.gpu.3.mem_gb":0.5,"ray/node.0.mem_gb":22.022918701171875,"ray/node.0.gpu.0.util":0,"ray/node.0.gpu.0.mem_gb":0.5,"ray/node.0.gpu.3.util":0}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb/wandb/run-20260502_210121-m9xk3p9p/logs/debug-core.log

@@ -0,0 +1,15 @@
+{"time":"2026-05-02T21:01:27.021617501+09:00","level":"INFO","msg":"main: starting server","port-filename":"/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/tmpfdc1r7hu/port-333.txt","pid":333,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-05-02T21:01:27.038960532+09:00","level":"INFO","msg":"server: will exit if parent process dies","ppid":333}
+{"time":"2026-05-02T21:01:27.039933738+09:00","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/wandb-333-8338-1677838298/socket","Net":"unix"}}
+{"time":"2026-05-02T21:01:27.050003734+09:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-05-02T21:01:27.084824833+09:00","level":"INFO","msg":"handleInformInit: received","streamId":"m9xk3p9p","id":"1(@)"}
+{"time":"2026-05-02T21:01:32.465471389+09:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"m9xk3p9p","id":"1(@)"}
+{"time":"2026-05-02T21:01:38.537250423+09:00","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"vlpea0bto5kg"}
+{"time":"2026-05-02T21:02:41.810472518+09:00","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
+{"time":"2026-05-02T21:02:41.810581455+09:00","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-05-02T21:02:41.810729078+09:00","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/wandb-333-8338-1677838298/socket","Net":"unix"}}
+{"time":"2026-05-02T21:02:41.810573067+09:00","level":"INFO","msg":"connection: closing","id":"1(@)"}
+{"time":"2026-05-02T21:02:41.811311314+09:00","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
+{"time":"2026-05-02T21:02:43.121448189+09:00","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
+{"time":"2026-05-02T21:02:43.12149348+09:00","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
+{"time":"2026-05-02T21:02:43.12151221+09:00","level":"INFO","msg":"server is closed"}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb/wandb/run-20260502_210121-m9xk3p9p/logs/debug-internal.log

@@ -0,0 +1,31 @@
+{"time":"2026-05-02T21:01:27.089287178+09:00","level":"INFO","msg":"wandb-core"}
+{"time":"2026-05-02T21:01:27.116487637+09:00","level":"INFO","msg":"stream: starting","core version":"0.25.1"}
+{"time":"2026-05-02T21:01:32.177703408+09:00","level":"ERROR","msg":"monitor: failed to initialize GPU resource: monitor: could not get GPU binary port: timeout reading portfile /tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/wandb-system-monitor-portfile-791433119"}
+{"time":"2026-05-02T21:01:32.463460082+09:00","level":"INFO","msg":"stream: created new stream","id":"m9xk3p9p"}
+{"time":"2026-05-02T21:01:32.464136548+09:00","level":"INFO","msg":"handler: started"}
+{"time":"2026-05-02T21:01:32.465451984+09:00","level":"INFO","msg":"stream: started"}
+{"time":"2026-05-02T21:01:32.466026932+09:00","level":"INFO","msg":"writer: started","stream_id":"m9xk3p9p"}
+{"time":"2026-05-02T21:01:32.467207947+09:00","level":"INFO","msg":"sender: started"}
+{"time":"2026-05-02T21:01:33.577986724+09:00","level":"INFO","msg":"filestream: sending request","total_files":1,"console_offset":0,"console_lines":1}
+{"time":"2026-05-02T21:01:34.004592056+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:01:48.576628897+09:00","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":0,"events_lines":1,"console_offset":0,"console_lines":16,"uploaded_len":2}
+{"time":"2026-05-02T21:01:49.073487601+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:02:03.57739858+09:00","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":1,"events_lines":1,"console_offset":11,"console_lines":1}
+{"time":"2026-05-02T21:02:03.995233262+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:02:18.577863349+09:00","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":2,"events_lines":1,"console_offset":16,"console_lines":15}
+{"time":"2026-05-02T21:02:18.859874591+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:02:33.577360034+09:00","level":"INFO","msg":"filestream: sending request","total_files":2,"events_offset":3,"events_lines":1,"console_offset":31,"console_lines":15}
+{"time":"2026-05-02T21:02:33.967842201+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:02:39.041087354+09:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":187}
+{"time":"2026-05-02T21:02:39.047778317+09:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":32}
+{"time":"2026-05-02T21:02:39.151241183+09:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":305}
+{"time":"2026-05-02T21:02:39.157195106+09:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":72}
+{"time":"2026-05-02T21:02:40.599169371+09:00","level":"INFO","msg":"flowcontrol: backed up, offloading to disk","recordNumber":477}
+{"time":"2026-05-02T21:02:40.614228567+09:00","level":"INFO","msg":"flowcontrol: unblocked","totalOffloaded":242}
+{"time":"2026-05-02T21:02:41.811243346+09:00","level":"INFO","msg":"stream: closing"}
+{"time":"2026-05-02T21:02:42.745957822+09:00","level":"INFO","msg":"fileTransfer: Close: file transfer manager closed"}
+{"time":"2026-05-02T21:02:42.747911197+09:00","level":"INFO","msg":"filestream: sending request","total_files":3,"history_offset":0,"history_lines":1,"console_offset":46,"console_lines":516,"uploaded_len":3,"complete":true,"exit_code":1}
+{"time":"2026-05-02T21:02:43.119821689+09:00","level":"INFO","msg":"filestream: request sent","status":"200 OK"}
+{"time":"2026-05-02T21:02:43.120030377+09:00","level":"INFO","msg":"handler: closed"}
+{"time":"2026-05-02T21:02:43.120442815+09:00","level":"INFO","msg":"sender: closed"}
+{"time":"2026-05-02T21:02:43.120456852+09:00","level":"INFO","msg":"stream: closed"}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/train_log_20260502.205927.txt

@@ -0,0 +1,718 @@
+grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/3rdparty/Megatron-LM-workspace/Megatron-LM/megatron/core/distributed/fsdp/src/megatron_fsdp/mixed_precision.py:114: UserWarning: Transformer Engine and Apex are not installed. Falling back to local implementations of multi_tensor_applier and multi_tensor_scale
+  warnings.warn(
+/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/tvm_ffi/_optional_torch_c_dlpack.py:181: UserWarning: Failed to JIT torch c dlpack extension, EnvTensorAllocator will not be enabled.
+We recommend installing via `pip install torch-c-dlpack-ext`
+  warnings.warn(
+/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/3rdparty/Megatron-LM-workspace/Megatron-LM/megatron/core/optimizer/__init__.py:24: UserWarning: Transformer Engine and Apex are not installed. Falling back to Torch optimizers.
+  warnings.warn(
+/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/3rdparty/Megatron-LM-workspace/Megatron-LM/megatron/core/optimizer/optimizer.py:29: UserWarning: Transformer Engine and Apex are not installed. Falling back to local implementations of multi_tensor_applier and multi_tensor_scale
+  warnings.warn(
+/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/3rdparty/Megatron-LM-workspace/Megatron-LM/megatron/core/optimizer/clip_grads.py:29: UserWarning: Transformer Engine and Apex are not installed. Falling back to local implementations of multi_tensor_applier, multi_tensor_l2norm, and multi_tensor_scale
+  warnings.warn(
+/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/3rdparty/Megatron-LM-workspace/Megatron-LM/megatron/core/models/backends.py:23: UserWarning: Apex is not installed. Falling back to Torch Norm
+  warnings.warn("Apex is not installed. Falling back to Torch Norm")
+[run_opsd] Using teacher-context prompt variant: knowledge_base_answer
+[INFO] Dataset processor knowledge_opsd registered
+Registered knowledge_opsd datasets from /home/pj24003162/ku40003404/jiarui/TeacherMix/data/knowledge_d5_e10/atomic_sft
+  -> Patching _build_cot_gt_texts with knowledge GT variant
+  -> Patching _build_teacher_inputs_with_cot_in_user_turn with smart truncation (max_new_tokens=8192, teacher input budget=1808)
+2026-05-02 21:01:15,485	INFO worker.py:2004 -- Started a local Ray instance. View the dashboard at http://127.0.0.1:8265 
+/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/worker.py:2052: FutureWarning: Tip: In future versions of Ray, Ray will no longer override accelerator visible devices env var if num_gpus=0 or num_gpus=None (default). To enable this behavior and turn off this error message, set RAY_ACCEL_ENV_VAR_OVERRIDE_ON_ZERO=0
+  warnings.warn(
+No chat template provided, using tokenizer's default
+
+▶ Setting up envs...
+
+▶ Setting up data...
+  - Loaded training dataset atomic_sft-train_messages with 100 samples.
+TaskDataSpec(task_name='atomic_sft-train_messages', prompt_file=None, system_prompt_file=None) <function _knowledge_opsd_processor at 0x14b3d2835f80>
+  ✓ Training dataset loaded with 100 samples.
+  - Loaded validation dataset atomic_sft-val_paraphrased_messages with 100 samples.
+  - Loaded validation dataset compositional_2step_sft-test_messages with 93 samples.
+  ✓ Validation dataset loaded with 193 samples.
+wandb: [wandb.login()] Loaded credentials for https://api.wandb.ai from WANDB_API_KEY.
+wandb: Currently logged in as: jiaruil5 (monitorability) to https://api.wandb.ai. Use `wandb login --relogin` to force relogin
+wandb: setting up run m9xk3p9p
+wandb: Tracking run with wandb version 0.25.1
+wandb: Run data is saved locally in /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb/wandb/run-20260502_210121-m9xk3p9p
+wandb: Run `wandb offline` to turn off syncing.
+wandb: Syncing run offpsd-knowledge-nll-gt-d5-e10-n1-t0-20260502-Llama-3.2-3B-Instruct
+wandb: ⭐️ View project at https://wandb.ai/mathrl-jiarui/mix-teachers-opsd
+wandb: 🚀 View run at https://wandb.ai/mathrl-jiarui/mix-teachers-opsd/runs/m9xk3p9p
+Unexpected error during git code logging: Artifact type 'code' is reserved for internal use. Please use a different type.
+No uncommitted changes found
+Error during git operations: Command '['git', 'diff', 'main']' returned non-zero exit status 128.
+Initialized WandbLogger for project mix-teachers-opsd, run offpsd-knowledge-nll-gt-d5-e10-n1-t0-20260502-Llama-3.2-3B-Instruct at /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb
+Initialized TensorboardLogger at /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/tensorboard
+GPU monitoring started with collection interval=10s, flush interval=10s
+  ✓ Training dataloader loaded with 100 samples
+  ✓ Validation dataloader loaded with 193 samples
+
+▶ Setting up compute cluster...
+  ✓ Ray cluster initialized with 1 nodes
+
+(IsolatedWorkerInitializer pid=9616) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(IsolatedWorkerInitializer pid=9619) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(IsolatedWorkerInitializer pid=9618) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+
+  ✓ 4 workers initialized in 13.56s
+(pid=10124) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(pid=10122) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(pid=10123) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(pid=10125) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(VllmGenerationWorker pid=10124) VLLM_RAY_BUNDLE_INDICES environment variable set to: 0,1,2,3
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [utils.py:253] non-default args: {'served_model_name': '/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct', 'skip_tokenizer_init': True, 'trust_remote_code': True, 'load_format': 'dummy', 'dtype': 'bfloat16', 'seed': 0, 'max_model_len': 10000, 'distributed_executor_backend': 'ray', 'tensor_parallel_size': 4, 'enable_prefix_caching': True, 'gpu_memory_utilization': 0.7, 'logprobs_mode': 'processed_logprobs', 'worker_extension_cls': 'nemo_rl.models.generation.vllm.vllm_backend.VllmInternalWorkerExtension', 'enable_sleep_mode': True, 'model': '/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct'}
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [model.py:631] Resolved architecture: LlamaForCausalLM
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [model.py:1745] Using max model len 10000
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [arg_utils.py:1443] Using ray runtime env (env vars redacted): {'py_executable': '/home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3', 'env_vars': {'AI_AGENT': '***', 'AVAILABLE_ADDR_LIST': '***', 'AVAILABLE_PORT_LIST': '***', 'BASH_FUNC__module_raw%%': '***', 'BASH_FUNC_ml%%': '***', 'BASH_FUNC_module%%': '***', 'BASH_FUNC_scl%%': '***', 'BASH_FUNC_switchml%%': '***', 'BASH_FUNC_which%%': '***', 'CLAUDECODE': '***', 'CLAUDE_CODE_ENTRYPOINT': '***', 'CLAUDE_CODE_EXECPATH': '***', 'CONDA_DEFAULT_ENV': '***', 'CONDA_EXE': '***', 'CONDA_PREFIX': '***', 'CONDA_PREFIX_1': '***', 'CONDA_PROMPT_MODIFIER': '***', 'CONDA_PYTHON_EXE': '***', 'CONDA_SHLVL': '***', 'COREPACK_ENABLE_AUTO_PIN': '***', 'CPATH': '***', 'CPATH_modshare': '***', 'CUDA_DIR': '***', 'CUDA_HOME': '***', 'CUDA_INSTALL_PATH': '***', 'CUDA_MPS_LOG_DIRECTORY': '***', 'CUDA_MPS_PIPE_DIRECTORY': '***', 'CUDA_PATH': '***', 'CUDA_TOP': '***', 'CYCLECLOUD_BOOTSTRAP': '***', 'CYCLECLOUD_HOME': '***', 'C_INCLUDE_PATH': '***', 'C_INCLUDE_PATH_modshare': '***', 'DBUS_SESSION_BUS_ADDRESS': '***', 'DISPLAY': '***', 'FJPNCKPT_CONFIG': '***', 'GDK_BACKEND': '***', 'GENKAI_CACHE_DIR': '***', 'GENKAI_CUDA_NAME': '***', 'GENKAI_CUDA_VER': '***', 'GENKAI_FAST_DIR': '***', 'GIT_EDITOR': '***', 'GLOBUS_GSSAPI_NAME_COMPATIBILITY': '***', 'HF_TOKEN': '***', 'HISTCONTROL': '***', 'HISTSIZE': '***', 'HOME': '***', 'HOSTNAME': '***', 'HYDRA_FULL_ERROR': '***', 'JOBID': '***', 'KMP_DUPLICATE_LIB_OK': '***', 'KMP_INIT_AT_FORK': '***', 'LANG': '***', 'LD_LIBRARY_PATH': '***', 'LD_LIBRARY_PATH_modshare': '***', 'LESSOPEN': '***', 'LIBRARY_PATH': '***', 'LIBRARY_PATH_modshare': '***', 'LOADEDMODULES': '***', 'LOADEDMODULES_modshare': '***', 'LOCAL_RANK': '***', 'LOGNAME': '***', 'LS_COLORS': '***', 'LUSTRE_JOBSTAT': '***', 'MAIL': '***', 'MANPATH': '***', 'MASTER_ADDR': '***', 'MASTER_PORT': '***', 'MATHEMATICA_HOME': '***', 'MIG_PARTED_CHECKPOINT_FILE': '***', 'MIG_PARTED_CONFIG_FILE': '***', 'MIG_PARTED_HOOKS_FILE': '***', 'MODULEPATH': '***', 'MODULEPATH_modshare': '***', 'MODULESHOME': '***', 'MODULES_CMD': '***', 'MODULES_LMCONFLICT': '***', 'MODULES_LMCONFLICT_modshare': '***', 'MODULES_RUN_QUARANTINE': '***', 'MOD_GIT_ROOTDIR': '***', 'NCCL_ASYNC_ERROR_HANDLING': '***', 'NCCL_DEBUG': '***', 'NCCL_SHM_USE_CUDA_MEMCPY': '***', 'NCCL_TIMEOUT': '***', 'NEMO_RL_PY_EXECUTABLES_SYSTEM': '***', 'NEMO_RL_VENV_DIR': '***', 'NODE_RANK': '***', 'NoDefaultCurrentDirectoryInExePath': '***', 'OLDPWD': '***', 'OMPI_CKPTCONFIG': '***', 'OTEL_EXPORTER_OTLP_METRICS_TEMPORALITY_PREFERENCE': '***', 'PATH': '***', 'PATH_modshare': '***', 'PJM_ASSIGN_LOGICAL_CPU': '***', 'PJM_CORE_MEM_LIMIT': '***', 'PJM_CUSTOM_RESOURCES': '***', 'PJM_DPREFIX': '***', 'PJM_ELAPSED_TIME_MODE': '***', 'PJM_ELAPSE_LIMIT': '***', 'PJM_ENVIRONMENT': '***', 'PJM_JOBDIR': '***', 'PJM_JOBID': '***', 'PJM_JOBNAME': '***', 'PJM_MAILOPTION': '***', 'PJM_MPI_PROC': '***', 'PJM_NET_ROUTE': '***', 'PJM_NODE': '***', 'PJM_NODE_CPUTIME_LIMIT': '***', 'PJM_O_HOME': '***', 'PJM_O_HOST': '***', 'PJM_O_LANG': '***', 'PJM_O_LOGNAME': '***', 'PJM_O_MAIL': '***', 'PJM_O_NODEINF': '***', 'PJM_O_PATH': '***', 'PJM_O_SHELL': '***', 'PJM_O_WORKDIR': '***', 'PJM_PROC_BY_NODE': '***', 'PJM_RSCGRP': '***', 'PJM_RSCUNIT': '***', 'PJM_SHELL': '***', 'PJM_SSD_DIR': '***', 'PJM_STDERR_PATH': '***', 'PJM_STDOUT_PATH': '***', 'PJM_SUBJOBID': '***', 'PJM_VNODE_MEM_LIMIT': '***', 'PLE_SCRIPT_TYPE': '***', 'PWD': '***', 'PYTHONBREAKPOINT': '***', 'PYTHONPATH': '***', 'PYTHONUNBUFFERED': '***', 'RANK': '***', 'RAY_CLIENT_MODE': '***', 'RAY_DEDUP_LOGS': '***', 'RAY_ENABLE_UV_RUN_RUNTIME_ENV': '***', 'RAY_EXPERIMENTAL_NOSET_CUDA_VISIBLE_DEVICES': '***', 'RAY_TMPDIR': '***', 'RAY_USAGE_STATS_ENABLED': '***', 'RAY_memory_monitor_refresh_ms': '***', 'RAY_worker_heartbeat_timeout_milliseconds': '***', 'RAY_worker_register_timeout_seconds': '***', 'SHLVL': '***', 'SSH_ASKPASS': '***', 'SSH_CLIENT': '***', 'SSH_CONNECTION': '***', 'SSH_TTY': '***', 'SWANLAB_API_HOST': '***', 'SWANLAB_RUNTIME': '***', 'SWANLAB_WEB_HOST': '***', 'S_COLORS': '***', 'TERM': '***', 'TMP': '***', 'TMPDIR': '***', 'TMUX': '***', 'TMUX_PANE': '***', 'TMUX_TMPDIR': '***', 'TORCHINDUCTOR_CACHE_DIR': '***', 'USER': '***', 'UV_PROJECT_ENVIRONMENT': '***', 'VIRTUAL_ENV': '***', 'VLLM_CACHE_ROOT': '***', 'VLLM_ENABLE_V1_MULTIPROCESSING': '***', 'VLLM_SKIP_P2P_CHECK': '***', 'VLLM_USE_V1': '***', 'WANDB_API_KEY': '***', 'WANDB_CONFIG_DIR': '***', 'WANDB_DIR': '***', 'WANDB_SERVICE': '***', 'WORLD_SIZE': '***', 'XDG_DATA_DIRS': '***', 'XDG_RUNTIME_DIR': '***', 'XDG_SESSION_ID': '***', '_': '***', '_CE_CONDA': '***', '_CE_M': '***', '_CONDA_EXE': '***', '_CONDA_ROOT': '***', '_LMFILES_': '***', '_LMFILES__modshare': '***', '_MLFLOW_TELEMETRY_SESSION_ID': '***', 'which_declare': '***'}}
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [scheduler.py:216] Chunked prefill is enabled with max_num_batched_tokens=16384.
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [core.py:93] Initializing a V1 LLM engine (v0.11.2) with config: model='/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct', speculative_config=None, tokenizer='/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct', skip_tokenizer_init=True, tokenizer_mode=auto, revision=None, tokenizer_revision=None, trust_remote_code=True, dtype=torch.bfloat16, max_seq_len=10000, download_dir=None, load_format=dummy, tensor_parallel_size=4, pipeline_parallel_size=1, data_parallel_size=1, disable_custom_all_reduce=False, quantization=None, enforce_eager=False, kv_cache_dtype=auto, device_config=cuda, structured_outputs_config=StructuredOutputsConfig(backend='auto', disable_fallback=False, disable_any_whitespace=False, disable_additional_properties=False, reasoning_parser='', reasoning_parser_plugin='', enable_in_reasoning=False), observability_config=ObservabilityConfig(show_hidden_metrics_for_version=None, otlp_traces_endpoint=None, collect_detailed_traces=None), seed=0, served_model_name=/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct, enable_prefix_caching=True, enable_chunked_prefill=True, pooler_config=None, compilation_config={'level': None, 'mode': <CompilationMode.VLLM_COMPILE: 3>, 'debug_dump_path': None, 'cache_dir': '', 'compile_cache_save_format': 'binary', 'backend': 'inductor', 'custom_ops': ['none'], 'splitting_ops': ['vllm::unified_attention', 'vllm::unified_attention_with_output', 'vllm::unified_mla_attention', 'vllm::unified_mla_attention_with_output', 'vllm::mamba_mixer2', 'vllm::mamba_mixer', 'vllm::short_conv', 'vllm::linear_attention', 'vllm::plamo2_mamba_mixer', 'vllm::gdn_attention_core', 'vllm::kda_attention', 'vllm::sparse_attn_indexer'], 'compile_mm_encoder': False, 'use_inductor': None, 'compile_sizes': [], 'inductor_compile_config': {'enable_auto_functionalized_v2': False, 'combo_kernels': True, 'benchmark_combo_kernel': True}, 'inductor_passes': {}, 'cudagraph_mode': <CUDAGraphMode.FULL_AND_PIECEWISE: (2, 1)>, 'cudagraph_num_of_warmups': 1, 'cudagraph_capture_sizes': [1, 2, 4, 8, 16, 24, 32, 40, 48, 56, 64, 72, 80, 88, 96, 104, 112, 120, 128, 136, 144, 152, 160, 168, 176, 184, 192, 200, 208, 216, 224, 232, 240, 248, 256, 272, 288, 304, 320, 336, 352, 368, 384, 400, 416, 432, 448, 464, 480, 496, 512], 'cudagraph_copy_inputs': False, 'cudagraph_specialize_lora': True, 'use_inductor_graph_partition': False, 'pass_config': {}, 'max_cudagraph_capture_size': 512, 'local_cache_dir': None}
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [ray_utils.py:343] Ray is already initialized. Skipping Ray initialization.
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:08 [ray_utils.py:374] Using the existing placement group
+(VllmGenerationWorker pid=10124) The argument `trust_remote_code` is to be used with Auto classes. It has no effect here and is ignored.
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:22 [ray_env.py:66] RAY_NON_CARRY_OVER_ENV_VARS from config: set()
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:22 [ray_env.py:69] Copying the following environment variables to workers: ['VLLM_ALLOW_INSECURE_SERIALIZATION', 'CUDA_HOME', 'VLLM_SKIP_P2P_CHECK', 'VLLM_CACHE_ROOT', 'VLLM_RAY_BUNDLE_INDICES', 'VLLM_ENABLE_V1_MULTIPROCESSING', 'RAY_ENABLE_UV_RUN_RUNTIME_ENV', 'HF_TOKEN', 'LD_LIBRARY_PATH', 'VLLM_RAY_PER_WORKER_GPUS']
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:22 [ray_env.py:74] If certain env vars should NOT be copied, add them to /home/pj24003162/ku40003404/.config/vllm/ray_non_carry_over_env_vars.json file
+(RayWorkerWrapper pid=10547) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(RayWorkerWrapper pid=10551) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(RayWorkerWrapper pid=10554) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(RayWorkerWrapper pid=10548) grouped_gemm is not available. Please run:pip install git+https://github.com/fanshiqing/grouped_gemm@v1.1.4
+(RayWorkerWrapper pid=10547) INFO 05-02 21:02:32 [worker_base.py:283] Injected <class 'nemo_rl.models.generation.vllm.vllm_backend.VllmInternalWorkerExtension'> into <class 'vllm.v1.worker.gpu_worker.Worker'> for extended collective_rpc calls ['_maybe_process_fp8_kv_cache', 'cleanup', 'get_zmq_address', 'init_collective', 'maybe_init_zmq', 'prepare_refit_info', 'report_device_id', 'start_gpu_profiling', 'stop_gpu_profiling', 'update_weights_from_collective', 'update_weights_via_ipc_zmq']
+(RayWorkerWrapper pid=10547) WARNING 05-02 21:02:32 [worker_base.py:301] Missing `shared_worker_lock` argument from executor. This argument is needed for mm_processor_cache_type='shm'.
+(RayWorkerWrapper pid=10551) INFO 05-02 21:02:32 [worker_base.py:283] Injected <class 'nemo_rl.models.generation.vllm.vllm_backend.VllmInternalWorkerExtension'> into <class 'vllm.v1.worker.gpu_worker.Worker'> for extended collective_rpc calls ['_maybe_process_fp8_kv_cache', 'cleanup', 'get_zmq_address', 'init_collective', 'maybe_init_zmq', 'prepare_refit_info', 'report_device_id', 'start_gpu_profiling', 'stop_gpu_profiling', 'update_weights_from_collective', 'update_weights_via_ipc_zmq']
+(RayWorkerWrapper pid=10551) WARNING 05-02 21:02:32 [worker_base.py:301] Missing `shared_worker_lock` argument from executor. This argument is needed for mm_processor_cache_type='shm'.
+(RayWorkerWrapper pid=10554) INFO 05-02 21:02:32 [worker_base.py:283] Injected <class 'nemo_rl.models.generation.vllm.vllm_backend.VllmInternalWorkerExtension'> into <class 'vllm.v1.worker.gpu_worker.Worker'> for extended collective_rpc calls ['_maybe_process_fp8_kv_cache', 'cleanup', 'get_zmq_address', 'init_collective', 'maybe_init_zmq', 'prepare_refit_info', 'report_device_id', 'start_gpu_profiling', 'stop_gpu_profiling', 'update_weights_from_collective', 'update_weights_via_ipc_zmq']
+(RayWorkerWrapper pid=10554) WARNING 05-02 21:02:32 [worker_base.py:301] Missing `shared_worker_lock` argument from executor. This argument is needed for mm_processor_cache_type='shm'.
+(RayWorkerWrapper pid=10548) INFO 05-02 21:02:32 [worker_base.py:283] Injected <class 'nemo_rl.models.generation.vllm.vllm_backend.VllmInternalWorkerExtension'> into <class 'vllm.v1.worker.gpu_worker.Worker'> for extended collective_rpc calls ['_maybe_process_fp8_kv_cache', 'cleanup', 'get_zmq_address', 'init_collective', 'maybe_init_zmq', 'prepare_refit_info', 'report_device_id', 'start_gpu_profiling', 'stop_gpu_profiling', 'update_weights_from_collective', 'update_weights_via_ipc_zmq']
+(RayWorkerWrapper pid=10548) WARNING 05-02 21:02:32 [worker_base.py:301] Missing `shared_worker_lock` argument from executor. This argument is needed for mm_processor_cache_type='shm'.
+(RayWorkerWrapper pid=10547) INFO 05-02 21:02:34 [parallel_state.py:1208] world_size=4 rank=1 local_rank=1 distributed_init_method=tcp://127.0.0.1:48355 backend=nccl
+(RayWorkerWrapper pid=10551) INFO 05-02 21:02:34 [parallel_state.py:1208] world_size=4 rank=0 local_rank=0 distributed_init_method=tcp://127.0.0.1:48355 backend=nccl
+(RayWorkerWrapper pid=10554) INFO 05-02 21:02:34 [parallel_state.py:1208] world_size=4 rank=2 local_rank=2 distributed_init_method=tcp://127.0.0.1:48355 backend=nccl
+(RayWorkerWrapper pid=10548) INFO 05-02 21:02:34 [parallel_state.py:1208] world_size=4 rank=3 local_rank=3 distributed_init_method=tcp://127.0.0.1:48355 backend=nccl
+(RayWorkerWrapper pid=10547) [Gloo] Rank 1 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10551) [Gloo] Rank 0 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10554) [Gloo] Rank 2 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10548) [Gloo] Rank 3 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10547) [Gloo] Rank 1 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10551) [Gloo] Rank 0 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10551) INFO 05-02 21:02:34 [pynccl.py:111] vLLM is using nccl==2.27.5
+(RayWorkerWrapper pid=10551) NCCL version 2.27.5+cuda12.9
+(RayWorkerWrapper pid=10554) [Gloo] Rank 2 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10548) [Gloo] Rank 3 is connected to 3 peer ranks. Expected number of connected peer ranks is : 3
+(RayWorkerWrapper pid=10547) WARNING 05-02 21:02:38 [symm_mem.py:107] SymmMemCommunicator: symmetric memory multicast operations are not supported.
+(RayWorkerWrapper pid=10551) WARNING 05-02 21:02:38 [symm_mem.py:107] SymmMemCommunicator: symmetric memory multicast operations are not supported.
+(RayWorkerWrapper pid=10554) WARNING 05-02 21:02:38 [symm_mem.py:107] SymmMemCommunicator: symmetric memory multicast operations are not supported.
+(RayWorkerWrapper pid=10548) WARNING 05-02 21:02:38 [symm_mem.py:107] SymmMemCommunicator: symmetric memory multicast operations are not supported.
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342] Error executing method 'init_device'. This might cause deadlock in distributed execution.
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342] Traceback (most recent call last):
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     return run_method(self, method, args, kwargs)
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/util/tracing/tracing_helper.py", line 461, in _resume_span
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     return method(self, *_args, **_kwargs)
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     self.worker.init_device()  # type: ignore
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     init_worker_distributed_environment(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     ensure_model_parallel_initialized(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     initialize_model_parallel(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     _TP = init_model_parallel_group(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     return GroupCoordinator(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 707, in create_from_process_group
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     buffer_io = MessageQueue(
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]                 ^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 281, in __init__
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     self.local_socket.bind(local_subscribe_addr)
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/zmq/sugar/socket.py", line 320, in bind
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]     super().bind(addr)
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342]   File "zmq/backend/cython/_zmq.py", line 1002, in zmq.backend.cython._zmq.Socket.bind
+(RayWorkerWrapper pid=10551) ERROR 05-02 21:02:39 [worker_base.py:342] zmq.error.ZMQError: ipc path "/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355" is longer than 107 characters (sizeof(sockaddr_un.sun_path)). zmq.IPC_PATH_MAX_LEN constant can be used to check addr length (if it is defined). (addr='ipc:///tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355')
+(VllmGenerationWorker pid=10124) INFO 05-02 21:02:39 [ray_executor.py:121] Shutting down Ray distributed executor. If you see error log from logging.cc regarding SIGTERM received, please ignore because this is the expected termination process in Ray.
+(VllmGenerationWorker pid=10124) Exception raised in creation task: The actor died because of an error raised in its creation task, ray::vllm_policy-grp0-0:VllmGenerationWorker.__init__() (pid=10124, ip=172.16.17.20, actor_id=788c98f21211d983db82fd8401000000, repr=VllmGenerationWorker)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 462, in __init__
+(VllmGenerationWorker pid=10124)     self._create_engine(llm_kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 579, in _create_engine
+(VllmGenerationWorker pid=10124)     self.llm = vllm.LLM(**llm_kwargs)
+(VllmGenerationWorker pid=10124)                ^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/entrypoints/llm.py", line 343, in __init__
+(VllmGenerationWorker pid=10124)     self.llm_engine = LLMEngine.from_engine_args(
+(VllmGenerationWorker pid=10124)                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 174, in from_engine_args
+(VllmGenerationWorker pid=10124)     return cls(
+(VllmGenerationWorker pid=10124)            ^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 108, in __init__
+(VllmGenerationWorker pid=10124)     self.engine_core = EngineCoreClient.make_client(
+(VllmGenerationWorker pid=10124)                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 95, in make_client
+(VllmGenerationWorker pid=10124)     return InprocClient(vllm_config, executor_class, log_stats)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 264, in __init__
+(VllmGenerationWorker pid=10124)     self.engine_core = EngineCore(*args, **kwargs)
+(VllmGenerationWorker pid=10124)                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core.py", line 102, in __init__
+(VllmGenerationWorker pid=10124)     self.model_executor = executor_class(vllm_config)
+(VllmGenerationWorker pid=10124)                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/abstract.py", line 101, in __init__
+(VllmGenerationWorker pid=10124)     self._init_executor()
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 97, in _init_executor
+(VllmGenerationWorker pid=10124)     self._init_workers_ray(placement_group, runtime_env={"py_executable": "/home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3"})
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 370, in _init_workers_ray
+(VllmGenerationWorker pid=10124)     self.collective_rpc("init_device")
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 493, in collective_rpc
+(VllmGenerationWorker pid=10124)     return ray.get(ray_worker_outputs, timeout=timeout)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)                                   ^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124) ray.exceptions.RayTaskError(ZMQError): ray::RayWorkerWrapper.execute_method() (pid=10551, ip=172.16.17.20, actor_id=503a40dba590cd200475f83501000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x1518ed3f8490>)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+(VllmGenerationWorker pid=10124)     raise e
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(VllmGenerationWorker pid=10124)     return run_method(self, method, args, kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(VllmGenerationWorker pid=10124)     self.worker.init_device()  # type: ignore
+(VllmGenerationWorker pid=10124)     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(VllmGenerationWorker pid=10124)     init_worker_distributed_environment(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(VllmGenerationWorker pid=10124)     ensure_model_parallel_initialized(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(VllmGenerationWorker pid=10124)     initialize_model_parallel(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(VllmGenerationWorker pid=10124)     _TP = init_model_parallel_group(
+(VllmGenerationWorker pid=10124)           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(VllmGenerationWorker pid=10124)     return GroupCoordinator(
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(VllmGenerationWorker pid=10124)     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(VllmGenerationWorker pid=10124)                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 707, in create_from_process_group
+(VllmGenerationWorker pid=10124)     buffer_io = MessageQueue(
+(VllmGenerationWorker pid=10124)                 ^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 281, in __init__
+(VllmGenerationWorker pid=10124)     self.local_socket.bind(local_subscribe_addr)
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/zmq/sugar/socket.py", line 320, in bind
+(VllmGenerationWorker pid=10124)     super().bind(addr)
+(VllmGenerationWorker pid=10124)   File "zmq/backend/cython/_zmq.py", line 1002, in zmq.backend.cython._zmq.Socket.bind
+(VllmGenerationWorker pid=10124) zmq.error.ZMQError: ipc path "/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355" is longer than 107 characters (sizeof(sockaddr_un.sun_path)). zmq.IPC_PATH_MAX_LEN constant can be used to check addr length (if it is defined). (addr='ipc:///tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355')
+(RayWorkerWrapper pid=10551) [rank0]:[W502 21:02:40.774631326 ProcessGroupNCCL.cpp:1524] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
+(VllmGenerationWorker pid=10124) Unhandled error (suppress with 'RAY_IGNORE_UNHANDLED_ERRORS=1'): ray::RayWorkerWrapper.execute_method() (pid=10554, ip=172.16.17.20, actor_id=3a062de3b6335cf114e4b89b01000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x14d9739f7c90>)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+(VllmGenerationWorker pid=10124)     raise e
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(VllmGenerationWorker pid=10124)     return run_method(self, method, args, kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(VllmGenerationWorker pid=10124)     self.worker.init_device()  # type: ignore
+(VllmGenerationWorker pid=10124)     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(VllmGenerationWorker pid=10124)     init_worker_distributed_environment(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(VllmGenerationWorker pid=10124)     ensure_model_parallel_initialized(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(VllmGenerationWorker pid=10124)     initialize_model_parallel(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(VllmGenerationWorker pid=10124)     _TP = init_model_parallel_group(
+(VllmGenerationWorker pid=10124)           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(VllmGenerationWorker pid=10124)     return GroupCoordinator(
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(VllmGenerationWorker pid=10124)     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(VllmGenerationWorker pid=10124)                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 724, in create_from_process_group
+(VllmGenerationWorker pid=10124)     dist.broadcast_object_list(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 3691, in broadcast_object_list
+(VllmGenerationWorker pid=10124)     broadcast(object_sizes_tensor, group_src=group_src, group=group)
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 2841, in broadcast
+(VllmGenerationWorker pid=10124)     work.wait()
+(VllmGenerationWorker pid=10124) RuntimeError: [/pytorch/third_party/gloo/gloo/transport/tcp/pair.cc:547] Connection closed by peer [172.16.17.20]:27801
+(VllmGenerationWorker pid=10124) Unhandled error (suppress with 'RAY_IGNORE_UNHANDLED_ERRORS=1'): ray::RayWorkerWrapper.execute_method() (pid=10547, ip=172.16.17.20, actor_id=afe18c46590e3690d9e9625e01000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x14338cc28f10>)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+(VllmGenerationWorker pid=10124)     raise e
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(VllmGenerationWorker pid=10124)     return run_method(self, method, args, kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(VllmGenerationWorker pid=10124)     self.worker.init_device()  # type: ignore
+(VllmGenerationWorker pid=10124)     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(VllmGenerationWorker pid=10124)     init_worker_distributed_environment(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(VllmGenerationWorker pid=10124)     ensure_model_parallel_initialized(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(VllmGenerationWorker pid=10124)     initialize_model_parallel(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(VllmGenerationWorker pid=10124)     _TP = init_model_parallel_group(
+(VllmGenerationWorker pid=10124)           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(VllmGenerationWorker pid=10124)     return GroupCoordinator(
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(VllmGenerationWorker pid=10124)     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(VllmGenerationWorker pid=10124)                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 724, in create_from_process_group
+(VllmGenerationWorker pid=10124)     dist.broadcast_object_list(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 3691, in broadcast_object_list
+(VllmGenerationWorker pid=10124)     broadcast(object_sizes_tensor, group_src=group_src, group=group)
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 2841, in broadcast
+(VllmGenerationWorker pid=10124)     work.wait()
+(VllmGenerationWorker pid=10124) RuntimeError: [/pytorch/third_party/gloo/gloo/transport/tcp/pair.cc:547] Connection closed by peer [172.16.17.20]:35480
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342] Error executing method 'init_device'. This might cause deadlock in distributed execution.
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342] Traceback (most recent call last):
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return run_method(self, method, args, kwargs)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/util/tracing/tracing_helper.py", line 461, in _resume_span
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return method(self, *_args, **_kwargs)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     self.worker.init_device()  # type: ignore
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     init_worker_distributed_environment(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     ensure_model_parallel_initialized(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     initialize_model_parallel(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     _TP = init_model_parallel_group(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return GroupCoordinator(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 724, in create_from_process_group
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     dist.broadcast_object_list(
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 3691, in broadcast_object_list
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     broadcast(object_sizes_tensor, group_src=group_src, group=group)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 2841, in broadcast
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342]     work.wait()
+(RayWorkerWrapper pid=10547) ERROR 05-02 21:02:40 [worker_base.py:342] RuntimeError: [/pytorch/third_party/gloo/gloo/transport/tcp/pair.cc:547] Connection closed by peer [172.16.17.20]:35480
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342] Error executing method 'init_device'. This might cause deadlock in distributed execution.
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342] Traceback (most recent call last):
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return run_method(self, method, args, kwargs)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/util/tracing/tracing_helper.py", line 461, in _resume_span
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return method(self, *_args, **_kwargs)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     self.worker.init_device()  # type: ignore
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     init_worker_distributed_environment(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     ensure_model_parallel_initialized(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     initialize_model_parallel(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     _TP = init_model_parallel_group(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return GroupCoordinator(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 724, in create_from_process_group
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     dist.broadcast_object_list(
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 3691, in broadcast_object_list
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     broadcast(object_sizes_tensor, group_src=group_src, group=group)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 2841, in broadcast
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342]     work.wait()
+(RayWorkerWrapper pid=10554) ERROR 05-02 21:02:40 [worker_base.py:342] RuntimeError: [/pytorch/third_party/gloo/gloo/transport/tcp/pair.cc:547] Connection closed by peer [172.16.17.20]:27801
+Traceback (most recent call last):
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py", line 205, in <module>
+    main()
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py", line 193, in main
+    run_opsd_training(
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/math_operations_opsd/opsd_common.py", line 394, in run_opsd_training
+    ) = setup(config, tokenizer, dataset, val_dataset)
+        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/algorithms/self_distillation.py", line 1038, in setup
+    student_generation = VllmGeneration(
+                         ^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_generation.py", line 235, in __init__
+    self._post_init()
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_generation.py", line 433, in _post_init
+    results = ray.get(futures)
+              ^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/auto_init_hook.py", line 22, in auto_init_wrapper
+    return fn(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/client_mode_hook.py", line 104, in wrapper
+    return func(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/worker.py", line 2981, in get
+    values, debugger_breakpoint = worker.get_objects(
+                                  ^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/worker.py", line 1014, in get_objects
+    raise value
+ray.exceptions.ActorDiedError: The actor died because of an error raised in its creation task, ray::vllm_policy-grp0-0:VllmGenerationWorker.__init__() (pid=10124, ip=172.16.17.20, actor_id=788c98f21211d983db82fd8401000000, repr=VllmGenerationWorker)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 462, in __init__
+    self._create_engine(llm_kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 579, in _create_engine
+    self.llm = vllm.LLM(**llm_kwargs)
+               ^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/entrypoints/llm.py", line 343, in __init__
+    self.llm_engine = LLMEngine.from_engine_args(
+                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 174, in from_engine_args
+    return cls(
+           ^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 108, in __init__
+    self.engine_core = EngineCoreClient.make_client(
+                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 95, in make_client
+    return InprocClient(vllm_config, executor_class, log_stats)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 264, in __init__
+    self.engine_core = EngineCore(*args, **kwargs)
+                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core.py", line 102, in __init__
+    self.model_executor = executor_class(vllm_config)
+                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/abstract.py", line 101, in __init__
+    self._init_executor()
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 97, in _init_executor
+    self._init_workers_ray(placement_group, runtime_env={"py_executable": "/home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3"})
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 370, in _init_workers_ray
+    self.collective_rpc("init_device")
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 493, in collective_rpc
+    return ray.get(ray_worker_outputs, timeout=timeout)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^
+                                  ^^^^^^^^^^^^^^^^^^^
+ray.exceptions.RayTaskError(ZMQError): ray::RayWorkerWrapper.execute_method() (pid=10551, ip=172.16.17.20, actor_id=503a40dba590cd200475f83501000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x1518ed3f8490>)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+    raise e
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+    return run_method(self, method, args, kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+    return func(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+    self.worker.init_device()  # type: ignore
+    ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+    init_worker_distributed_environment(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+    ensure_model_parallel_initialized(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+    initialize_model_parallel(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+    _TP = init_model_parallel_group(
+          ^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+    return GroupCoordinator(
+           ^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+    self.mq_broadcaster = MessageQueue.create_from_process_group(
+                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 707, in create_from_process_group
+    buffer_io = MessageQueue(
+                ^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 281, in __init__
+    self.local_socket.bind(local_subscribe_addr)
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/zmq/sugar/socket.py", line 320, in bind
+    super().bind(addr)
+  File "zmq/backend/cython/_zmq.py", line 1002, in zmq.backend.cython._zmq.Socket.bind
+zmq.error.ZMQError: ipc path "/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355" is longer than 107 characters (sizeof(sockaddr_un.sun_path)). zmq.IPC_PATH_MAX_LEN constant can be used to check addr length (if it is defined). (addr='ipc:///tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355')
+Traceback (most recent call last):
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py", line 205, in <module>
+    main()
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py", line 193, in main
+    run_opsd_training(
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/math_operations_opsd/opsd_common.py", line 394, in run_opsd_training
+    ) = setup(config, tokenizer, dataset, val_dataset)
+        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/algorithms/self_distillation.py", line 1038, in setup
+    student_generation = VllmGeneration(
+                         ^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_generation.py", line 235, in __init__
+    self._post_init()
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_generation.py", line 433, in _post_init
+    results = ray.get(futures)
+              ^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/auto_init_hook.py", line 22, in auto_init_wrapper
+    return fn(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/client_mode_hook.py", line 104, in wrapper
+    return func(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/worker.py", line 2981, in get
+    values, debugger_breakpoint = worker.get_objects(
+                                  ^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/_private/worker.py", line 1014, in get_objects
+    raise value
+ray.exceptions.ActorDiedError: The actor died because of an error raised in its creation task, ray::vllm_policy-grp0-0:VllmGenerationWorker.__init__() (pid=10124, ip=172.16.17.20, actor_id=788c98f21211d983db82fd8401000000, repr=VllmGenerationWorker)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 462, in __init__
+    self._create_engine(llm_kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/nemo-rl/nemo_rl/models/generation/vllm/vllm_worker.py", line 579, in _create_engine
+    self.llm = vllm.LLM(**llm_kwargs)
+               ^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/entrypoints/llm.py", line 343, in __init__
+    self.llm_engine = LLMEngine.from_engine_args(
+                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 174, in from_engine_args
+    return cls(
+           ^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/llm_engine.py", line 108, in __init__
+    self.engine_core = EngineCoreClient.make_client(
+                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 95, in make_client
+    return InprocClient(vllm_config, executor_class, log_stats)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core_client.py", line 264, in __init__
+    self.engine_core = EngineCore(*args, **kwargs)
+                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/engine/core.py", line 102, in __init__
+    self.model_executor = executor_class(vllm_config)
+                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/abstract.py", line 101, in __init__
+    self._init_executor()
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 97, in _init_executor
+    self._init_workers_ray(placement_group, runtime_env={"py_executable": "/home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3"})
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 370, in _init_workers_ray
+    self.collective_rpc("init_device")
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/executor/ray_executor.py", line 493, in collective_rpc
+    return ray.get(ray_worker_outputs, timeout=timeout)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^
+                                  ^^^^^^^^^^^^^^^^^^^
+ray.exceptions.RayTaskError(ZMQError): ray::RayWorkerWrapper.execute_method() (pid=10551, ip=172.16.17.20, actor_id=503a40dba590cd200475f83501000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x1518ed3f8490>)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+    raise e
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+    return run_method(self, method, args, kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+    return func(*args, **kwargs)
+           ^^^^^^^^^^^^^^^^^^^^^
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+    self.worker.init_device()  # type: ignore
+    ^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+    init_worker_distributed_environment(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+    ensure_model_parallel_initialized(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+    initialize_model_parallel(
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+    _TP = init_model_parallel_group(
+          ^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+    return GroupCoordinator(
+           ^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+    self.mq_broadcaster = MessageQueue.create_from_process_group(
+                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 707, in create_from_process_group
+    buffer_io = MessageQueue(
+                ^^^^^^^^^^^^^
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 281, in __init__
+    self.local_socket.bind(local_subscribe_addr)
+  File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/zmq/sugar/socket.py", line 320, in bind
+    super().bind(addr)
+  File "zmq/backend/cython/_zmq.py", line 1002, in zmq.backend.cython._zmq.Socket.bind
+zmq.error.ZMQError: ipc path "/tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355" is longer than 107 characters (sizeof(sockaddr_un.sun_path)). zmq.IPC_PATH_MAX_LEN constant can be used to check addr length (if it is defined). (addr='ipc:///tmp/5738900_b2ZmcHNkLWtub3ctbmxsLWd0LWQ1ZTEwLWxsYW1hMy0yLTNiLW4xLXQwCg==/40909801-4452-4e06-a3d2-f099b3864355')
+(VllmGenerationWorker pid=10124) Unhandled error (suppress with 'RAY_IGNORE_UNHANDLED_ERRORS=1'): ray::RayWorkerWrapper.execute_method() (pid=10548, ip=172.16.17.20, actor_id=0f999d2a68297575e6e4563701000000, repr=<vllm.v1.executor.ray_utils.RayWorkerWrapper object at 0x147bc3e3d110>)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 343, in execute_method
+(VllmGenerationWorker pid=10124)     raise e
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(VllmGenerationWorker pid=10124)     return run_method(self, method, args, kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(VllmGenerationWorker pid=10124)     self.worker.init_device()  # type: ignore
+(VllmGenerationWorker pid=10124)     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(VllmGenerationWorker pid=10124)     init_worker_distributed_environment(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+(VllmGenerationWorker pid=10124)     ensure_model_parallel_initialized(
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+WARNING:asyncio:selector_events.py:1054: socket.send() raised exception.
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(VllmGenerationWorker pid=10124)     initialize_model_parallel(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(VllmGenerationWorker pid=10124)     _TP = init_model_parallel_group(
+(VllmGenerationWorker pid=10124)           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(VllmGenerationWorker pid=10124)     return GroupCoordinator(
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(VllmGenerationWorker pid=10124)     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(VllmGenerationWorker pid=10124)                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 724, in create_from_process_group
+(VllmGenerationWorker pid=10124)     dist.broadcast_object_list(
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 3691, in broadcast_object_list
+(VllmGenerationWorker pid=10124)     broadcast(object_sizes_tensor, group_src=group_src, group=group)
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(VllmGenerationWorker pid=10124)     return func(*args, **kwargs)
+(VllmGenerationWorker pid=10124)            ^^^^^^^^^^^^^^^^^^^^^
+(VllmGenerationWorker pid=10124)   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 2841, in broadcast
+(VllmGenerationWorker pid=10124)     work.wait()
+(VllmGenerationWorker pid=10124) RuntimeError: [/pytorch/third_party/gloo/gloo/transport/tcp/pair.cc:547] Connection closed by peer [172.16.17.20]:12717
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342] Error executing method 'init_device'. This might cause deadlock in distributed execution.
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342] Traceback (most recent call last):
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 332, in execute_method
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     return run_method(self, method, args, kwargs)
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/serial_utils.py", line 479, in run_method
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/ray/util/tracing/tracing_helper.py", line 461, in _resume_span
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     return method(self, *_args, **_kwargs)
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/worker_base.py", line 324, in init_device
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     self.worker.init_device()  # type: ignore
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     ^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 224, in init_device
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     init_worker_distributed_environment(
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/v1/worker/gpu_worker.py", line 886, in init_worker_distributed_environment
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     ensure_model_parallel_initialized(
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1418, in ensure_model_parallel_initialized
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     initialize_model_parallel(
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1338, in initialize_model_parallel
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     _TP = init_model_parallel_group(
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]           ^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 1061, in init_model_parallel_group
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     return GroupCoordinator(
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/parallel_state.py", line 374, in __init__
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     self.mq_broadcaster = MessageQueue.create_from_process_group(
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/vllm/distributed/device_communicators/shm_broadcast.py", line 724, in create_from_process_group
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     dist.broadcast_object_list(
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 3691, in broadcast_object_list
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     broadcast(object_sizes_tensor, group_src=group_src, group=group)
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/c10d_logger.py", line 81, in wrapper
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     return func(*args, **kwargs)
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]            ^^^^^^^^^^^^^^^^^^^^^
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]   File "/home/pj24003162/ku40003404/miniconda3/envs/opsd/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py", line 2841, in broadcast
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342]     work.wait()
+(RayWorkerWrapper pid=10548) ERROR 05-02 21:02:41 [worker_base.py:342] RuntimeError: [/pytorch/third_party/gloo/gloo/transport/tcp/pair.cc:547] Connection closed by peer [172.16.17.20]:12717
+wandb: 
+wandb: 🚀 View run offpsd-knowledge-nll-gt-d5-e10-n1-t0-20260502-Llama-3.2-3B-Instruct at: https://wandb.ai/mathrl-jiarui/mix-teachers-opsd/runs/m9xk3p9p
+wandb: Find logs at: ../../../checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.205927/logs/exp_001/wandb/wandb/run-20260502_210121-m9xk3p9p/logs

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_200/config.yaml

@@ -0,0 +1,256 @@
+checkpointing:
+  checkpoint_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834
+  checkpoint_must_save_by: null
+  enabled: true
+  higher_is_better: true
+  keep_top_k: null
+  metric_name: null
+  model_save_format: safetensors
+  save_consolidated: false
+  save_optimizer: false
+  save_per_epoch: false
+  save_period: 20
+cluster:
+  gpus_per_node: 4
+  num_nodes: 1
+data:
+  default:
+    env_name: knowledge
+    prompt_file: null
+    system_prompt_file: null
+  max_input_seq_length: 10000
+  shuffle: true
+  train:
+  - dataset_name: knowledge_opsd
+    env_name: knowledge
+    processor: knowledge_opsd
+    prompt_file: null
+    system_prompt_file: null
+  validation:
+  - dataset_name: knowledge_opsd_val
+    env_name: knowledge
+    processor: knowledge_opsd
+    prompt_file: null
+    system_prompt_file: null
+  - dataset_name: knowledge_opsd_comp_val
+    env_name: knowledge
+    processor: knowledge_opsd
+    prompt_file: null
+    system_prompt_file: null
+distillation:
+  ema_decay: 1.0
+  max_num_epochs: 40
+  max_num_steps: 1000
+  max_rollout_turns: 1
+  max_val_samples: 200
+  num_generations_per_prompt: 1
+  num_prompts_per_step: 20
+  rollout_policy: teacher
+  seed: 42
+  topk_logits_k: 64
+  use_nll_loss: true
+  val_at_end: true
+  val_at_start: true
+  val_batch_size: 32
+  val_period: 20
+env:
+  knowledge:
+    num_workers: 8
+logger:
+  gpu_monitoring:
+    collection_interval: 10
+    flush_interval: 10
+  log_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/logs/exp_001
+  mlflow_enabled: false
+  monitor_gpus: true
+  num_val_samples_to_print: 5
+  swanlab_enabled: false
+  tensorboard:
+    log_dir: tb_logs-offpsd-knowledge-nll-gt-d5-e10-llama3-2-3b-n1-t0
+  tensorboard_enabled: true
+  wandb:
+    entity: mathrl-jiarui
+    name: offpsd-knowledge-nll-gt-d5-e10-n1-t0-20260502-Llama-3.2-3B-Instruct
+    project: mix-teachers-opsd
+  wandb_enabled: true
+loss_fn:
+  kl_type: forward
+  mixed_kl_weight: 0.5
+  zero_outside_topk: false
+paths:
+  base_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd
+  checkpoint_dir: /home/fs01/yw2349/jiarui/mathrl_data/mix_teachers/checkpoints/knowledge/opsd_nemo_rl
+  data_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/data/knowledge_d5_e10/atomic_sft
+  log_dir: /home/fs01/yw2349/jiarui/mathrl_data/mix_teachers/checkpoints/knowledge/opsd_nemo_rl/logs
+  model_search_dirs:
+  - /home/fs01/yw2349/jiarui/mathrl_data/models
+policy:
+  dtensor_cfg:
+    _v2: true
+    activation_checkpointing: true
+    context_parallel_size: 1
+    cpu_offload: true
+    custom_parallel_plan: null
+    enabled: true
+    sequence_parallel: false
+    tensor_parallel_size: 4
+  dynamic_batching:
+    enabled: true
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  generation:
+    _pad_token_id: 128009
+    backend: vllm
+    colocated:
+      enabled: true
+      resources:
+        gpus_per_node: null
+        num_nodes: null
+    max_new_tokens: 8192
+    model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+    stop_strings: null
+    stop_token_ids:
+    - 128009
+    temperature: 0.0
+    top_k: null
+    top_p: 1.0
+    vllm_cfg:
+      async_engine: false
+      distributed_executor_backend: null
+      enforce_eager: false
+      expert_parallel_size: 1
+      gpu_memory_utilization: 0.7
+      hf_overrides: {}
+      kv_cache_dtype: auto
+      load_format: dummy
+      max_model_len: 10000
+      num_first_layers_in_bf16: 0
+      num_last_layers_in_bf16: 0
+      pipeline_parallel_size: 1
+      precision: bfloat16
+      skip_tokenizer_init: true
+      tensor_parallel_size: 4
+      use_deep_gemm: false
+  generation_batch_size: 20
+  logprob_batch_size: 1
+  logprob_chunk_size: null
+  make_sequence_length_divisible_by: 8
+  max_grad_norm: 1.0
+  max_total_sequence_length: 10000
+  megatron_cfg:
+    enabled: false
+  model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+  offload_optimizer_for_logprob: false
+  optimizer:
+    kwargs:
+      betas:
+      - 0.9
+      - 0.95
+      eps: 1.0e-08
+      foreach: false
+      fused: false
+      lr: 3.0e-05
+      weight_decay: 0.01
+    name: torch.optim.AdamW
+  precision: bfloat16
+  scheduler:
+  - kwargs:
+      factor: 1.0
+      total_iters: 10000000000
+    name: torch.optim.lr_scheduler.ConstantLR
+  - milestones: []
+  sequence_packing:
+    algorithm: modified_first_fit_decreasing
+    enabled: false
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  tokenizer:
+    chat_template_kwargs:
+      enable_thinking: false
+    name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+  train_global_batch_size: 20
+  train_micro_batch_size: 1
+teacher:
+  dtensor_cfg:
+    _v2: true
+    activation_checkpointing: true
+    context_parallel_size: 1
+    cpu_offload: true
+    custom_parallel_plan: null
+    enabled: true
+    sequence_parallel: false
+    tensor_parallel_size: 4
+  dynamic_batching:
+    enabled: true
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  generation:
+    backend: vllm
+    colocated:
+      enabled: true
+      resources:
+        gpus_per_node: null
+        num_nodes: null
+    max_new_tokens: 8192
+    stop_strings: null
+    stop_token_ids: null
+    temperature: 1.0
+    top_k: null
+    top_p: 1.0
+    vllm_cfg:
+      async_engine: false
+      distributed_executor_backend: null
+      enforce_eager: false
+      expert_parallel_size: 1
+      gpu_memory_utilization: 0.7
+      kv_cache_dtype: auto
+      max_model_len: 10000
+      num_first_layers_in_bf16: 0
+      num_last_layers_in_bf16: 0
+      pipeline_parallel_size: 1
+      precision: bfloat16
+      tensor_parallel_size: 4
+      use_deep_gemm: false
+  generation_batch_size: 16
+  logprob_batch_size: 1
+  logprob_chunk_size: null
+  make_sequence_length_divisible_by: 8
+  max_grad_norm: 1.0
+  max_total_sequence_length: 10000
+  megatron_cfg:
+    enabled: false
+  model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+  offload_optimizer_for_logprob: false
+  optimizer:
+    kwargs:
+      betas:
+      - 0.9
+      - 0.95
+      eps: 1.0e-08
+      foreach: false
+      fused: false
+      lr: 1.0e-05
+      weight_decay: 0.01
+    name: torch.optim.AdamW
+  precision: bfloat16
+  scheduler:
+  - kwargs:
+      factor: 1.0
+      total_iters: 10000000000
+    name: torch.optim.lr_scheduler.ConstantLR
+  - milestones: []
+  sequence_packing:
+    algorithm: modified_first_fit_decreasing
+    enabled: false
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  tokenizer:
+    chat_template_kwargs:
+      enable_thinking: false
+    name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+  train_global_batch_size: 16
+  train_micro_batch_size: 1

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_200/policy/tokenizer/chat_template.jinja

@@ -0,0 +1,93 @@
+{{- bos_token }}
+{%- if custom_tools is defined %}
+    {%- set tools = custom_tools %}
+{%- endif %}
+{%- if not tools_in_user_message is defined %}
+    {%- set tools_in_user_message = true %}
+{%- endif %}
+{%- if not date_string is defined %}
+    {%- if strftime_now is defined %}
+        {%- set date_string = strftime_now("%d %b %Y") %}
+    {%- else %}
+        {%- set date_string = "26 Jul 2024" %}
+    {%- endif %}
+{%- endif %}
+{%- if not tools is defined %}
+    {%- set tools = none %}
+{%- endif %}
+
+{#- This block extracts the system message, so we can slot it into the right place. #}
+{%- if messages[0]['role'] == 'system' %}
+    {%- set system_message = messages[0]['content']|trim %}
+    {%- set messages = messages[1:] %}
+{%- else %}
+    {%- set system_message = "" %}
+{%- endif %}
+
+{#- System message #}
+{{- "<|start_header_id|>system<|end_header_id|>\n\n" }}
+{%- if tools is not none %}
+    {{- "Environment: ipython\n" }}
+{%- endif %}
+{{- "Cutting Knowledge Date: December 2023\n" }}
+{{- "Today Date: " + date_string + "\n\n" }}
+{%- if tools is not none and not tools_in_user_message %}
+    {{- "You have access to the following functions. To call a function, please respond with JSON for a function call." }}
+    {{- 'Respond in the format {"name": function name, "parameters": dictionary of argument name and its value}.' }}
+    {{- "Do not use variables.\n\n" }}
+    {%- for t in tools %}
+        {{- t | tojson(indent=4) }}
+        {{- "\n\n" }}
+    {%- endfor %}
+{%- endif %}
+{{- system_message }}
+{{- "<|eot_id|>" }}
+
+{#- Custom tools are passed in a user message with some extra guidance #}
+{%- if tools_in_user_message and not tools is none %}
+    {#- Extract the first user message so we can plug it in here #}
+    {%- if messages | length != 0 %}
+        {%- set first_user_message = messages[0]['content']|trim %}
+        {%- set messages = messages[1:] %}
+    {%- else %}
+        {{- raise_exception("Cannot put tools in the first user message when there's no first user message!") }}
+{%- endif %}
+    {{- '<|start_header_id|>user<|end_header_id|>\n\n' -}}
+    {{- "Given the following functions, please respond with a JSON for a function call " }}
+    {{- "with its proper arguments that best answers the given prompt.\n\n" }}
+    {{- 'Respond in the format {"name": function name, "parameters": dictionary of argument name and its value}.' }}
+    {{- "Do not use variables.\n\n" }}
+    {%- for t in tools %}
+        {{- t | tojson(indent=4) }}
+        {{- "\n\n" }}
+    {%- endfor %}
+    {{- first_user_message + "<|eot_id|>"}}
+{%- endif %}
+
+{%- for message in messages %}
+    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}
+        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' }}
+    {%- elif 'tool_calls' in message %}
+        {%- if not message.tool_calls|length == 1 %}
+            {{- raise_exception("This model only supports single tool-calls at once!") }}
+        {%- endif %}
+        {%- set tool_call = message.tool_calls[0].function %}
+        {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' -}}
+        {{- '{"name": "' + tool_call.name + '", ' }}
+        {{- '"parameters": ' }}
+        {{- tool_call.arguments | tojson }}
+        {{- "}" }}
+        {{- "<|eot_id|>" }}
+    {%- elif message.role == "tool" or message.role == "ipython" %}
+        {{- "<|start_header_id|>ipython<|end_header_id|>\n\n" }}
+        {%- if message.content is mapping or message.content is iterable %}
+            {{- message.content | tojson }}
+        {%- else %}
+            {{- message.content }}
+        {%- endif %}
+        {{- "<|eot_id|>" }}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' }}
+{%- endif %}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_200/policy/tokenizer/special_tokens_map.json

@@ -0,0 +1,17 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<|eot_id|>"
+}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_200/policy/tokenizer/tokenizer_config.json

@@ -0,0 +1,2063 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "<|eot_id|>",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_200/policy/weights/model/.hf_metadata/config.json

@@ -0,0 +1,41 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "dtype": "float32",
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "nemo_version": "0.2.0",
+  "num_attention_heads": 24,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 8,
+  "pad_token_id": 128009,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.1",
+  "use_cache": false,
+  "vocab_size": 128256
+}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_200/policy/weights/model/.hf_metadata/fqn_to_file_index_mapping.json

@@ -0,0 +1,256 @@
+{
+  "model.embed_tokens.weight": 1,
+  "model.layers.0.input_layernorm.weight": 1,
+  "model.layers.0.mlp.down_proj.weight": 1,
+  "model.layers.0.mlp.gate_proj.weight": 1,
+  "model.layers.0.mlp.up_proj.weight": 1,
+  "model.layers.0.post_attention_layernorm.weight": 1,
+  "model.layers.0.self_attn.k_proj.weight": 1,
+  "model.layers.0.self_attn.o_proj.weight": 1,
+  "model.layers.0.self_attn.q_proj.weight": 1,
+  "model.layers.0.self_attn.v_proj.weight": 1,
+  "model.layers.1.input_layernorm.weight": 1,
+  "model.layers.1.mlp.down_proj.weight": 1,
+  "model.layers.1.mlp.gate_proj.weight": 1,
+  "model.layers.1.mlp.up_proj.weight": 1,
+  "model.layers.1.post_attention_layernorm.weight": 1,
+  "model.layers.1.self_attn.k_proj.weight": 1,
+  "model.layers.1.self_attn.o_proj.weight": 1,
+  "model.layers.1.self_attn.q_proj.weight": 1,
+  "model.layers.1.self_attn.v_proj.weight": 1,
+  "model.layers.10.input_layernorm.weight": 1,
+  "model.layers.10.mlp.down_proj.weight": 1,
+  "model.layers.10.mlp.gate_proj.weight": 1,
+  "model.layers.10.mlp.up_proj.weight": 1,
+  "model.layers.10.post_attention_layernorm.weight": 1,
+  "model.layers.10.self_attn.k_proj.weight": 1,
+  "model.layers.10.self_attn.o_proj.weight": 1,
+  "model.layers.10.self_attn.q_proj.weight": 1,
+  "model.layers.10.self_attn.v_proj.weight": 1,
+  "model.layers.11.input_layernorm.weight": 1,
+  "model.layers.11.mlp.down_proj.weight": 1,
+  "model.layers.11.mlp.gate_proj.weight": 1,
+  "model.layers.11.mlp.up_proj.weight": 1,
+  "model.layers.11.post_attention_layernorm.weight": 1,
+  "model.layers.11.self_attn.k_proj.weight": 1,
+  "model.layers.11.self_attn.o_proj.weight": 1,
+  "model.layers.11.self_attn.q_proj.weight": 1,
+  "model.layers.11.self_attn.v_proj.weight": 1,
+  "model.layers.12.input_layernorm.weight": 1,
+  "model.layers.12.mlp.down_proj.weight": 1,
+  "model.layers.12.mlp.gate_proj.weight": 1,
+  "model.layers.12.mlp.up_proj.weight": 1,
+  "model.layers.12.post_attention_layernorm.weight": 1,
+  "model.layers.12.self_attn.k_proj.weight": 1,
+  "model.layers.12.self_attn.o_proj.weight": 1,
+  "model.layers.12.self_attn.q_proj.weight": 1,
+  "model.layers.12.self_attn.v_proj.weight": 1,
+  "model.layers.13.input_layernorm.weight": 1,
+  "model.layers.13.mlp.down_proj.weight": 1,
+  "model.layers.13.mlp.gate_proj.weight": 1,
+  "model.layers.13.mlp.up_proj.weight": 1,
+  "model.layers.13.post_attention_layernorm.weight": 1,
+  "model.layers.13.self_attn.k_proj.weight": 1,
+  "model.layers.13.self_attn.o_proj.weight": 1,
+  "model.layers.13.self_attn.q_proj.weight": 1,
+  "model.layers.13.self_attn.v_proj.weight": 1,
+  "model.layers.14.input_layernorm.weight": 1,
+  "model.layers.14.mlp.down_proj.weight": 1,
+  "model.layers.14.mlp.gate_proj.weight": 1,
+  "model.layers.14.mlp.up_proj.weight": 1,
+  "model.layers.14.post_attention_layernorm.weight": 1,
+  "model.layers.14.self_attn.k_proj.weight": 1,
+  "model.layers.14.self_attn.o_proj.weight": 1,
+  "model.layers.14.self_attn.q_proj.weight": 1,
+  "model.layers.14.self_attn.v_proj.weight": 1,
+  "model.layers.15.input_layernorm.weight": 1,
+  "model.layers.15.mlp.down_proj.weight": 1,
+  "model.layers.15.mlp.gate_proj.weight": 1,
+  "model.layers.15.mlp.up_proj.weight": 1,
+  "model.layers.15.post_attention_layernorm.weight": 1,
+  "model.layers.15.self_attn.k_proj.weight": 1,
+  "model.layers.15.self_attn.o_proj.weight": 1,
+  "model.layers.15.self_attn.q_proj.weight": 1,
+  "model.layers.15.self_attn.v_proj.weight": 1,
+  "model.layers.16.input_layernorm.weight": 1,
+  "model.layers.16.mlp.down_proj.weight": 1,
+  "model.layers.16.mlp.gate_proj.weight": 1,
+  "model.layers.16.mlp.up_proj.weight": 1,
+  "model.layers.16.post_attention_layernorm.weight": 1,
+  "model.layers.16.self_attn.k_proj.weight": 1,
+  "model.layers.16.self_attn.o_proj.weight": 1,
+  "model.layers.16.self_attn.q_proj.weight": 1,
+  "model.layers.16.self_attn.v_proj.weight": 1,
+  "model.layers.17.input_layernorm.weight": 1,
+  "model.layers.17.mlp.down_proj.weight": 1,
+  "model.layers.17.mlp.gate_proj.weight": 1,
+  "model.layers.17.mlp.up_proj.weight": 1,
+  "model.layers.17.post_attention_layernorm.weight": 1,
+  "model.layers.17.self_attn.k_proj.weight": 1,
+  "model.layers.17.self_attn.o_proj.weight": 1,
+  "model.layers.17.self_attn.q_proj.weight": 1,
+  "model.layers.17.self_attn.v_proj.weight": 1,
+  "model.layers.18.input_layernorm.weight": 1,
+  "model.layers.18.mlp.down_proj.weight": 1,
+  "model.layers.18.mlp.gate_proj.weight": 1,
+  "model.layers.18.mlp.up_proj.weight": 1,
+  "model.layers.18.post_attention_layernorm.weight": 1,
+  "model.layers.18.self_attn.k_proj.weight": 1,
+  "model.layers.18.self_attn.o_proj.weight": 1,
+  "model.layers.18.self_attn.q_proj.weight": 1,
+  "model.layers.18.self_attn.v_proj.weight": 1,
+  "model.layers.19.input_layernorm.weight": 1,
+  "model.layers.19.mlp.down_proj.weight": 1,
+  "model.layers.19.mlp.gate_proj.weight": 1,
+  "model.layers.19.mlp.up_proj.weight": 1,
+  "model.layers.19.post_attention_layernorm.weight": 1,
+  "model.layers.19.self_attn.k_proj.weight": 1,
+  "model.layers.19.self_attn.o_proj.weight": 1,
+  "model.layers.19.self_attn.q_proj.weight": 1,
+  "model.layers.19.self_attn.v_proj.weight": 1,
+  "model.layers.2.input_layernorm.weight": 1,
+  "model.layers.2.mlp.down_proj.weight": 1,
+  "model.layers.2.mlp.gate_proj.weight": 1,
+  "model.layers.2.mlp.up_proj.weight": 1,
+  "model.layers.2.post_attention_layernorm.weight": 1,
+  "model.layers.2.self_attn.k_proj.weight": 1,
+  "model.layers.2.self_attn.o_proj.weight": 1,
+  "model.layers.2.self_attn.q_proj.weight": 1,
+  "model.layers.2.self_attn.v_proj.weight": 1,
+  "model.layers.20.input_layernorm.weight": 2,
+  "model.layers.20.mlp.down_proj.weight": 2,
+  "model.layers.20.mlp.gate_proj.weight": 1,
+  "model.layers.20.mlp.up_proj.weight": 1,
+  "model.layers.20.post_attention_layernorm.weight": 2,
+  "model.layers.20.self_attn.k_proj.weight": 1,
+  "model.layers.20.self_attn.o_proj.weight": 1,
+  "model.layers.20.self_attn.q_proj.weight": 1,
+  "model.layers.20.self_attn.v_proj.weight": 1,
+  "model.layers.21.input_layernorm.weight": 2,
+  "model.layers.21.mlp.down_proj.weight": 2,
+  "model.layers.21.mlp.gate_proj.weight": 2,
+  "model.layers.21.mlp.up_proj.weight": 2,
+  "model.layers.21.post_attention_layernorm.weight": 2,
+  "model.layers.21.self_attn.k_proj.weight": 2,
+  "model.layers.21.self_attn.o_proj.weight": 2,
+  "model.layers.21.self_attn.q_proj.weight": 2,
+  "model.layers.21.self_attn.v_proj.weight": 2,
+  "model.layers.22.input_layernorm.weight": 2,
+  "model.layers.22.mlp.down_proj.weight": 2,
+  "model.layers.22.mlp.gate_proj.weight": 2,
+  "model.layers.22.mlp.up_proj.weight": 2,
+  "model.layers.22.post_attention_layernorm.weight": 2,
+  "model.layers.22.self_attn.k_proj.weight": 2,
+  "model.layers.22.self_attn.o_proj.weight": 2,
+  "model.layers.22.self_attn.q_proj.weight": 2,
+  "model.layers.22.self_attn.v_proj.weight": 2,
+  "model.layers.23.input_layernorm.weight": 2,
+  "model.layers.23.mlp.down_proj.weight": 2,
+  "model.layers.23.mlp.gate_proj.weight": 2,
+  "model.layers.23.mlp.up_proj.weight": 2,
+  "model.layers.23.post_attention_layernorm.weight": 2,
+  "model.layers.23.self_attn.k_proj.weight": 2,
+  "model.layers.23.self_attn.o_proj.weight": 2,
+  "model.layers.23.self_attn.q_proj.weight": 2,
+  "model.layers.23.self_attn.v_proj.weight": 2,
+  "model.layers.24.input_layernorm.weight": 2,
+  "model.layers.24.mlp.down_proj.weight": 2,
+  "model.layers.24.mlp.gate_proj.weight": 2,
+  "model.layers.24.mlp.up_proj.weight": 2,
+  "model.layers.24.post_attention_layernorm.weight": 2,
+  "model.layers.24.self_attn.k_proj.weight": 2,
+  "model.layers.24.self_attn.o_proj.weight": 2,
+  "model.layers.24.self_attn.q_proj.weight": 2,
+  "model.layers.24.self_attn.v_proj.weight": 2,
+  "model.layers.25.input_layernorm.weight": 2,
+  "model.layers.25.mlp.down_proj.weight": 2,
+  "model.layers.25.mlp.gate_proj.weight": 2,
+  "model.layers.25.mlp.up_proj.weight": 2,
+  "model.layers.25.post_attention_layernorm.weight": 2,
+  "model.layers.25.self_attn.k_proj.weight": 2,
+  "model.layers.25.self_attn.o_proj.weight": 2,
+  "model.layers.25.self_attn.q_proj.weight": 2,
+  "model.layers.25.self_attn.v_proj.weight": 2,
+  "model.layers.26.input_layernorm.weight": 2,
+  "model.layers.26.mlp.down_proj.weight": 2,
+  "model.layers.26.mlp.gate_proj.weight": 2,
+  "model.layers.26.mlp.up_proj.weight": 2,
+  "model.layers.26.post_attention_layernorm.weight": 2,
+  "model.layers.26.self_attn.k_proj.weight": 2,
+  "model.layers.26.self_attn.o_proj.weight": 2,
+  "model.layers.26.self_attn.q_proj.weight": 2,
+  "model.layers.26.self_attn.v_proj.weight": 2,
+  "model.layers.27.input_layernorm.weight": 2,
+  "model.layers.27.mlp.down_proj.weight": 2,
+  "model.layers.27.mlp.gate_proj.weight": 2,
+  "model.layers.27.mlp.up_proj.weight": 2,
+  "model.layers.27.post_attention_layernorm.weight": 2,
+  "model.layers.27.self_attn.k_proj.weight": 2,
+  "model.layers.27.self_attn.o_proj.weight": 2,
+  "model.layers.27.self_attn.q_proj.weight": 2,
+  "model.layers.27.self_attn.v_proj.weight": 2,
+  "model.layers.3.input_layernorm.weight": 1,
+  "model.layers.3.mlp.down_proj.weight": 1,
+  "model.layers.3.mlp.gate_proj.weight": 1,
+  "model.layers.3.mlp.up_proj.weight": 1,
+  "model.layers.3.post_attention_layernorm.weight": 1,
+  "model.layers.3.self_attn.k_proj.weight": 1,
+  "model.layers.3.self_attn.o_proj.weight": 1,
+  "model.layers.3.self_attn.q_proj.weight": 1,
+  "model.layers.3.self_attn.v_proj.weight": 1,
+  "model.layers.4.input_layernorm.weight": 1,
+  "model.layers.4.mlp.down_proj.weight": 1,
+  "model.layers.4.mlp.gate_proj.weight": 1,
+  "model.layers.4.mlp.up_proj.weight": 1,
+  "model.layers.4.post_attention_layernorm.weight": 1,
+  "model.layers.4.self_attn.k_proj.weight": 1,
+  "model.layers.4.self_attn.o_proj.weight": 1,
+  "model.layers.4.self_attn.q_proj.weight": 1,
+  "model.layers.4.self_attn.v_proj.weight": 1,
+  "model.layers.5.input_layernorm.weight": 1,
+  "model.layers.5.mlp.down_proj.weight": 1,
+  "model.layers.5.mlp.gate_proj.weight": 1,
+  "model.layers.5.mlp.up_proj.weight": 1,
+  "model.layers.5.post_attention_layernorm.weight": 1,
+  "model.layers.5.self_attn.k_proj.weight": 1,
+  "model.layers.5.self_attn.o_proj.weight": 1,
+  "model.layers.5.self_attn.q_proj.weight": 1,
+  "model.layers.5.self_attn.v_proj.weight": 1,
+  "model.layers.6.input_layernorm.weight": 1,
+  "model.layers.6.mlp.down_proj.weight": 1,
+  "model.layers.6.mlp.gate_proj.weight": 1,
+  "model.layers.6.mlp.up_proj.weight": 1,
+  "model.layers.6.post_attention_layernorm.weight": 1,
+  "model.layers.6.self_attn.k_proj.weight": 1,
+  "model.layers.6.self_attn.o_proj.weight": 1,
+  "model.layers.6.self_attn.q_proj.weight": 1,
+  "model.layers.6.self_attn.v_proj.weight": 1,
+  "model.layers.7.input_layernorm.weight": 1,
+  "model.layers.7.mlp.down_proj.weight": 1,
+  "model.layers.7.mlp.gate_proj.weight": 1,
+  "model.layers.7.mlp.up_proj.weight": 1,
+  "model.layers.7.post_attention_layernorm.weight": 1,
+  "model.layers.7.self_attn.k_proj.weight": 1,
+  "model.layers.7.self_attn.o_proj.weight": 1,
+  "model.layers.7.self_attn.q_proj.weight": 1,
+  "model.layers.7.self_attn.v_proj.weight": 1,
+  "model.layers.8.input_layernorm.weight": 1,
+  "model.layers.8.mlp.down_proj.weight": 1,
+  "model.layers.8.mlp.gate_proj.weight": 1,
+  "model.layers.8.mlp.up_proj.weight": 1,
+  "model.layers.8.post_attention_layernorm.weight": 1,
+  "model.layers.8.self_attn.k_proj.weight": 1,
+  "model.layers.8.self_attn.o_proj.weight": 1,
+  "model.layers.8.self_attn.q_proj.weight": 1,
+  "model.layers.8.self_attn.v_proj.weight": 1,
+  "model.layers.9.input_layernorm.weight": 1,
+  "model.layers.9.mlp.down_proj.weight": 1,
+  "model.layers.9.mlp.gate_proj.weight": 1,
+  "model.layers.9.mlp.up_proj.weight": 1,
+  "model.layers.9.post_attention_layernorm.weight": 1,
+  "model.layers.9.self_attn.k_proj.weight": 1,
+  "model.layers.9.self_attn.o_proj.weight": 1,
+  "model.layers.9.self_attn.q_proj.weight": 1,
+  "model.layers.9.self_attn.v_proj.weight": 1,
+  "model.norm.weight": 2
+}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_200/policy/weights/model/.hf_metadata/generation_config.json

@@ -0,0 +1,12 @@
+{
+  "bos_token_id": 128000,
+  "do_sample": true,
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "temperature": 0.6,
+  "top_p": 0.9,
+  "transformers_version": "4.57.1"
+}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_200/training_info.json

@@ -0,0 +1 @@
+{"current_epoch": 39, "current_step": 5, "total_steps": 200, "val_reward": 0.3125, "consumed_samples": 4000, "total_valid_tokens": 582771.0}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_80/config.yaml

@@ -0,0 +1,256 @@
+checkpointing:
+  checkpoint_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834
+  checkpoint_must_save_by: null
+  enabled: true
+  higher_is_better: true
+  keep_top_k: null
+  metric_name: null
+  model_save_format: safetensors
+  save_consolidated: false
+  save_optimizer: false
+  save_per_epoch: false
+  save_period: 20
+cluster:
+  gpus_per_node: 4
+  num_nodes: 1
+data:
+  default:
+    env_name: knowledge
+    prompt_file: null
+    system_prompt_file: null
+  max_input_seq_length: 10000
+  shuffle: true
+  train:
+  - dataset_name: knowledge_opsd
+    env_name: knowledge
+    processor: knowledge_opsd
+    prompt_file: null
+    system_prompt_file: null
+  validation:
+  - dataset_name: knowledge_opsd_val
+    env_name: knowledge
+    processor: knowledge_opsd
+    prompt_file: null
+    system_prompt_file: null
+  - dataset_name: knowledge_opsd_comp_val
+    env_name: knowledge
+    processor: knowledge_opsd
+    prompt_file: null
+    system_prompt_file: null
+distillation:
+  ema_decay: 1.0
+  max_num_epochs: 40
+  max_num_steps: 1000
+  max_rollout_turns: 1
+  max_val_samples: 200
+  num_generations_per_prompt: 1
+  num_prompts_per_step: 20
+  rollout_policy: teacher
+  seed: 42
+  topk_logits_k: 64
+  use_nll_loss: true
+  val_at_end: true
+  val_at_start: true
+  val_batch_size: 32
+  val_period: 20
+env:
+  knowledge:
+    num_workers: 8
+logger:
+  gpu_monitoring:
+    collection_interval: 10
+    flush_interval: 10
+  log_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/logs/exp_001
+  mlflow_enabled: false
+  monitor_gpus: true
+  num_val_samples_to_print: 5
+  swanlab_enabled: false
+  tensorboard:
+    log_dir: tb_logs-offpsd-knowledge-nll-gt-d5-e10-llama3-2-3b-n1-t0
+  tensorboard_enabled: true
+  wandb:
+    entity: mathrl-jiarui
+    name: offpsd-knowledge-nll-gt-d5-e10-n1-t0-20260502-Llama-3.2-3B-Instruct
+    project: mix-teachers-opsd
+  wandb_enabled: true
+loss_fn:
+  kl_type: forward
+  mixed_kl_weight: 0.5
+  zero_outside_topk: false
+paths:
+  base_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd
+  checkpoint_dir: /home/fs01/yw2349/jiarui/mathrl_data/mix_teachers/checkpoints/knowledge/opsd_nemo_rl
+  data_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/data/knowledge_d5_e10/atomic_sft
+  log_dir: /home/fs01/yw2349/jiarui/mathrl_data/mix_teachers/checkpoints/knowledge/opsd_nemo_rl/logs
+  model_search_dirs:
+  - /home/fs01/yw2349/jiarui/mathrl_data/models
+policy:
+  dtensor_cfg:
+    _v2: true
+    activation_checkpointing: true
+    context_parallel_size: 1
+    cpu_offload: true
+    custom_parallel_plan: null
+    enabled: true
+    sequence_parallel: false
+    tensor_parallel_size: 4
+  dynamic_batching:
+    enabled: true
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  generation:
+    _pad_token_id: 128009
+    backend: vllm
+    colocated:
+      enabled: true
+      resources:
+        gpus_per_node: null
+        num_nodes: null
+    max_new_tokens: 8192
+    model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+    stop_strings: null
+    stop_token_ids:
+    - 128009
+    temperature: 0.0
+    top_k: null
+    top_p: 1.0
+    vllm_cfg:
+      async_engine: false
+      distributed_executor_backend: null
+      enforce_eager: false
+      expert_parallel_size: 1
+      gpu_memory_utilization: 0.7
+      hf_overrides: {}
+      kv_cache_dtype: auto
+      load_format: dummy
+      max_model_len: 10000
+      num_first_layers_in_bf16: 0
+      num_last_layers_in_bf16: 0
+      pipeline_parallel_size: 1
+      precision: bfloat16
+      skip_tokenizer_init: true
+      tensor_parallel_size: 4
+      use_deep_gemm: false
+  generation_batch_size: 20
+  logprob_batch_size: 1
+  logprob_chunk_size: null
+  make_sequence_length_divisible_by: 8
+  max_grad_norm: 1.0
+  max_total_sequence_length: 10000
+  megatron_cfg:
+    enabled: false
+  model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+  offload_optimizer_for_logprob: false
+  optimizer:
+    kwargs:
+      betas:
+      - 0.9
+      - 0.95
+      eps: 1.0e-08
+      foreach: false
+      fused: false
+      lr: 3.0e-05
+      weight_decay: 0.01
+    name: torch.optim.AdamW
+  precision: bfloat16
+  scheduler:
+  - kwargs:
+      factor: 1.0
+      total_iters: 10000000000
+    name: torch.optim.lr_scheduler.ConstantLR
+  - milestones: []
+  sequence_packing:
+    algorithm: modified_first_fit_decreasing
+    enabled: false
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  tokenizer:
+    chat_template_kwargs:
+      enable_thinking: false
+    name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+  train_global_batch_size: 20
+  train_micro_batch_size: 1
+teacher:
+  dtensor_cfg:
+    _v2: true
+    activation_checkpointing: true
+    context_parallel_size: 1
+    cpu_offload: true
+    custom_parallel_plan: null
+    enabled: true
+    sequence_parallel: false
+    tensor_parallel_size: 4
+  dynamic_batching:
+    enabled: true
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  generation:
+    backend: vllm
+    colocated:
+      enabled: true
+      resources:
+        gpus_per_node: null
+        num_nodes: null
+    max_new_tokens: 8192
+    stop_strings: null
+    stop_token_ids: null
+    temperature: 1.0
+    top_k: null
+    top_p: 1.0
+    vllm_cfg:
+      async_engine: false
+      distributed_executor_backend: null
+      enforce_eager: false
+      expert_parallel_size: 1
+      gpu_memory_utilization: 0.7
+      kv_cache_dtype: auto
+      max_model_len: 10000
+      num_first_layers_in_bf16: 0
+      num_last_layers_in_bf16: 0
+      pipeline_parallel_size: 1
+      precision: bfloat16
+      tensor_parallel_size: 4
+      use_deep_gemm: false
+  generation_batch_size: 16
+  logprob_batch_size: 1
+  logprob_chunk_size: null
+  make_sequence_length_divisible_by: 8
+  max_grad_norm: 1.0
+  max_total_sequence_length: 10000
+  megatron_cfg:
+    enabled: false
+  model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+  offload_optimizer_for_logprob: false
+  optimizer:
+    kwargs:
+      betas:
+      - 0.9
+      - 0.95
+      eps: 1.0e-08
+      foreach: false
+      fused: false
+      lr: 1.0e-05
+      weight_decay: 0.01
+    name: torch.optim.AdamW
+  precision: bfloat16
+  scheduler:
+  - kwargs:
+      factor: 1.0
+      total_iters: 10000000000
+    name: torch.optim.lr_scheduler.ConstantLR
+  - milestones: []
+  sequence_packing:
+    algorithm: modified_first_fit_decreasing
+    enabled: false
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  tokenizer:
+    chat_template_kwargs:
+      enable_thinking: false
+    name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-3B-Instruct
+  train_global_batch_size: 16
+  train_micro_batch_size: 1

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_80/policy/tokenizer/special_tokens_map.json

@@ -0,0 +1,17 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<|eot_id|>"
+}

checkpoints/knowledge/offpsd_nll_gt_d5_e10_llama3_2_3b_n1_t0_20260502/run_20260502.214834/step_80/training_info.json

@@ -0,0 +1 @@
+{"current_epoch": 15, "current_step": 5, "total_steps": 80, "val_reward": 0.3854166666666667, "consumed_samples": 1600, "total_valid_tokens": 226709.0}

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs/exp_001/wandb/wandb/run-20260502_204022-lec8l6gk/files/config.yaml

@@ -0,0 +1,398 @@
+_wandb:
+    value:
+        cli_version: 0.25.1
+        e:
+            8pnbx653goz1fz52mzj1yeg80v3inpgq:
+                args:
+                    - --config
+                    - /home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/opsd_distillation_config.yaml
+                    - policy.model_name=/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct
+                    - policy.dtensor_cfg.tensor_parallel_size=4
+                    - policy.dtensor_cfg.context_parallel_size=1
+                    - policy.max_total_sequence_length=10000
+                    - policy.generation.max_new_tokens=8192
+                    - policy.generation.vllm_cfg.tensor_parallel_size=4
+                    - policy.generation.vllm_cfg.gpu_memory_utilization=0.7
+                    - policy.logprob_batch_size=1
+                    - policy.dtensor_cfg.cpu_offload=true
+                    - policy.dtensor_cfg.activation_checkpointing=true
+                    - paths.base_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd
+                    - paths.data_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/data/knowledge_d5_e10/atomic_sft
+                    - checkpointing.checkpoint_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716
+                    - logger.log_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs
+                    - logger.wandb.project=mix-teachers-opsd
+                    - logger.wandb.name=opsd-knowledge-forward-kl-gt-d5-e10-20260502-Llama-3.2-1B-Instruct
+                    - logger.tensorboard.log_dir=tb_logs-opsd-knowledge-forward-kl-gt-d5-e10-llama3-2-1b
+                    - cluster.gpus_per_node=4
+                    - cluster.num_nodes=1
+                    - loss_fn.kl_type=forward
+                    - loss_fn.zero_outside_topk=false
+                    - distillation.ema_decay=1.0
+                    - distillation.topk_logits_k=64
+                    - distillation.num_prompts_per_step=20
+                    - distillation.num_generations_per_prompt=8
+                    - distillation.val_period=20
+                    - distillation.max_num_epochs=40
+                    - checkpointing.save_per_epoch=false
+                    - checkpointing.save_period=20
+                    - checkpointing.save_optimizer=false
+                    - checkpointing.keep_top_k=null
+                    - checkpointing.metric_name=null
+                    - policy.train_global_batch_size=20
+                    - policy.generation_batch_size=20
+                    - policy.generation.temperature=0.8
+                    - policy.optimizer.kwargs.lr=3.0e-5
+                    - policy.optimizer.kwargs.weight_decay=0.01
+                    - 'policy.scheduler=[{name: "torch.optim.lr_scheduler.ConstantLR", kwargs: {factor: 1.0, total_iters: 10000000000}}, {milestones: []}]'
+                codePath: mix_teachers/train/knowledge_opsd/run_opsd.py
+                codePathLocal: run_opsd.py
+                cpu_count: 120
+                cpu_count_logical: 120
+                cudaVersion: "13.2"
+                disk:
+                    /:
+                        total: "511749718016"
+                        used: "25329385472"
+                email: jiaruil5@andrew.cmu.edu
+                executable: /home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3
+                git:
+                    commit: 9938bfa2b66c98512427baa8c5adbea81dbf6dc8
+                    remote: https://github.com/jiarui-liu/TeacherMix.git
+                gpu: NVIDIA H100
+                gpu_count: 4
+                gpu_nvidia:
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "100485038080"
+                      name: NVIDIA H100
+                      uuid: GPU-f879d657-3497-66cd-daca-a4f8d1d73359
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "100485038080"
+                      name: NVIDIA H100
+                      uuid: GPU-4622e484-3226-d178-7e48-7aae340aad13
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "100485038080"
+                      name: NVIDIA H100
+                      uuid: GPU-ad99b84e-fcc6-7236-3d91-6cbaa83e680b
+                    - architecture: Hopper
+                      cudaCores: 16896
+                      memoryTotal: "100485038080"
+                      name: NVIDIA H100
+                      uuid: GPU-2681ca36-95a6-471b-e939-d814aa0a1dd3
+                host: b0017
+                memory:
+                    total: "1081436749824"
+                os: Linux-4.18.0-477.27.1.el8_8.x86_64-x86_64-with-glibc2.28
+                program: /home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py
+                python: CPython 3.11.15
+                root: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs/exp_001/wandb
+                startedAt: "2026-05-02T11:40:22.035189Z"
+                writerId: 8pnbx653goz1fz52mzj1yeg80v3inpgq
+        m:
+            - "2": ray/*
+              "5": 2
+              "6":
+                - 1
+              "7": []
+            - "1": ray/ray_step
+              "6": []
+              "7": []
+        python_version: 3.11.15
+        t:
+            "1":
+                - 1
+                - 5
+                - 11
+                - 30
+                - 41
+                - 49
+                - 50
+                - 51
+                - 53
+                - 71
+            "2":
+                - 1
+                - 5
+                - 11
+                - 30
+                - 41
+                - 49
+                - 50
+                - 51
+                - 53
+                - 71
+            "3":
+                - 7
+                - 13
+                - 41
+                - 61
+            "4": 3.11.15
+            "5": 0.25.1
+            "6": 4.57.1
+            "12": 0.25.1
+            "13": linux-x86_64
+checkpointing:
+    value:
+        checkpoint_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716
+        checkpoint_must_save_by: null
+        enabled: true
+        higher_is_better: true
+        keep_top_k: null
+        metric_name: null
+        model_save_format: safetensors
+        save_consolidated: false
+        save_optimizer: false
+        save_per_epoch: false
+        save_period: 20
+cluster:
+    value:
+        gpus_per_node: 4
+        num_nodes: 1
+data:
+    value:
+        default:
+            env_name: knowledge
+            prompt_file: null
+            system_prompt_file: null
+        max_input_seq_length: 10000
+        shuffle: true
+        train:
+            - dataset_name: knowledge_opsd
+              env_name: knowledge
+              processor: knowledge_opsd
+              prompt_file: null
+              system_prompt_file: null
+        validation:
+            - dataset_name: knowledge_opsd_val
+              env_name: knowledge
+              processor: knowledge_opsd
+              prompt_file: null
+              system_prompt_file: null
+            - dataset_name: knowledge_opsd_comp_val
+              env_name: knowledge
+              processor: knowledge_opsd
+              prompt_file: null
+              system_prompt_file: null
+distillation:
+    value:
+        ema_decay: 1
+        max_num_epochs: 40
+        max_num_steps: 1000
+        max_rollout_turns: 1
+        max_val_samples: 200
+        num_generations_per_prompt: 8
+        num_prompts_per_step: 20
+        rollout_policy: student
+        seed: 42
+        topk_logits_k: 64
+        val_at_end: true
+        val_at_start: true
+        val_batch_size: 32
+        val_period: 20
+env:
+    value:
+        knowledge:
+            num_workers: 8
+logger:
+    value:
+        gpu_monitoring:
+            collection_interval: 10
+            flush_interval: 10
+        log_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs/exp_001
+        mlflow_enabled: false
+        monitor_gpus: true
+        num_val_samples_to_print: 5
+        swanlab_enabled: false
+        tensorboard:
+            log_dir: tb_logs-opsd-knowledge-forward-kl-gt-d5-e10-llama3-2-1b
+        tensorboard_enabled: true
+        wandb:
+            entity: mathrl-jiarui
+            name: opsd-knowledge-forward-kl-gt-d5-e10-20260502-Llama-3.2-1B-Instruct
+            project: mix-teachers-opsd
+        wandb_enabled: true
+loss_fn:
+    value:
+        kl_type: forward
+        mixed_kl_weight: 0.5
+        zero_outside_topk: false
+paths:
+    value:
+        base_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd
+        checkpoint_dir: /home/fs01/yw2349/jiarui/mathrl_data/mix_teachers/checkpoints/knowledge/opsd_nemo_rl
+        data_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/data/knowledge_d5_e10/atomic_sft
+        log_dir: /home/fs01/yw2349/jiarui/mathrl_data/mix_teachers/checkpoints/knowledge/opsd_nemo_rl/logs
+        model_search_dirs:
+            - /home/fs01/yw2349/jiarui/mathrl_data/models
+policy:
+    value:
+        dtensor_cfg:
+            _v2: true
+            activation_checkpointing: true
+            context_parallel_size: 1
+            cpu_offload: true
+            custom_parallel_plan: null
+            enabled: true
+            sequence_parallel: false
+            tensor_parallel_size: 4
+        dynamic_batching:
+            enabled: true
+            logprob_mb_tokens: 10000
+            sequence_length_round: 64
+            train_mb_tokens: 10000
+        generation:
+            _pad_token_id: 128009
+            backend: vllm
+            colocated:
+                enabled: true
+                resources:
+                    gpus_per_node: null
+                    num_nodes: null
+            max_new_tokens: 8192
+            stop_strings: null
+            stop_token_ids:
+                - 128009
+            temperature: 0.8
+            top_k: null
+            top_p: 1
+            vllm_cfg:
+                async_engine: false
+                distributed_executor_backend: null
+                enforce_eager: false
+                expert_parallel_size: 1
+                gpu_memory_utilization: 0.7
+                kv_cache_dtype: auto
+                load_format: dummy
+                max_model_len: 10000
+                num_first_layers_in_bf16: 0
+                num_last_layers_in_bf16: 0
+                pipeline_parallel_size: 1
+                precision: bfloat16
+                skip_tokenizer_init: true
+                tensor_parallel_size: 4
+                use_deep_gemm: false
+        generation_batch_size: 20
+        logprob_batch_size: 1
+        logprob_chunk_size: null
+        make_sequence_length_divisible_by: 8
+        max_grad_norm: 1
+        max_total_sequence_length: 10000
+        megatron_cfg:
+            enabled: false
+        model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct
+        offload_optimizer_for_logprob: false
+        optimizer:
+            kwargs:
+                betas:
+                    - 0.9
+                    - 0.95
+                eps: 1e-08
+                foreach: false
+                fused: false
+                lr: 3e-05
+                weight_decay: 0.01
+            name: torch.optim.AdamW
+        precision: bfloat16
+        scheduler:
+            - kwargs:
+                factor: 1
+                total_iters: 10000000000
+              name: torch.optim.lr_scheduler.ConstantLR
+            - milestones: []
+        sequence_packing:
+            algorithm: modified_first_fit_decreasing
+            enabled: false
+            logprob_mb_tokens: 10000
+            sequence_length_round: 64
+            train_mb_tokens: 10000
+        tokenizer:
+            chat_template_kwargs:
+                enable_thinking: false
+            name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct
+        train_global_batch_size: 20
+        train_micro_batch_size: 1
+teacher:
+    value:
+        dtensor_cfg:
+            _v2: true
+            activation_checkpointing: true
+            context_parallel_size: 1
+            cpu_offload: true
+            custom_parallel_plan: null
+            enabled: true
+            sequence_parallel: false
+            tensor_parallel_size: 4
+        dynamic_batching:
+            enabled: true
+            logprob_mb_tokens: 10000
+            sequence_length_round: 64
+            train_mb_tokens: 10000
+        generation:
+            backend: vllm
+            colocated:
+                enabled: true
+                resources:
+                    gpus_per_node: null
+                    num_nodes: null
+            max_new_tokens: 8192
+            stop_strings: null
+            stop_token_ids: null
+            temperature: 1
+            top_k: null
+            top_p: 1
+            vllm_cfg:
+                async_engine: false
+                distributed_executor_backend: null
+                enforce_eager: false
+                expert_parallel_size: 1
+                gpu_memory_utilization: 0.7
+                kv_cache_dtype: auto
+                max_model_len: 10000
+                num_first_layers_in_bf16: 0
+                num_last_layers_in_bf16: 0
+                pipeline_parallel_size: 1
+                precision: bfloat16
+                tensor_parallel_size: 4
+                use_deep_gemm: false
+        generation_batch_size: 16
+        logprob_batch_size: 1
+        logprob_chunk_size: null
+        make_sequence_length_divisible_by: 8
+        max_grad_norm: 1
+        max_total_sequence_length: 10000
+        megatron_cfg:
+            enabled: false
+        model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct
+        offload_optimizer_for_logprob: false
+        optimizer:
+            kwargs:
+                betas:
+                    - 0.9
+                    - 0.95
+                eps: 1e-08
+                foreach: false
+                fused: false
+                lr: 1e-05
+                weight_decay: 0.01
+            name: torch.optim.AdamW
+        precision: bfloat16
+        scheduler:
+            - kwargs:
+                factor: 1
+                total_iters: 10000000000
+              name: torch.optim.lr_scheduler.ConstantLR
+            - milestones: []
+        sequence_packing:
+            algorithm: modified_first_fit_decreasing
+            enabled: false
+            logprob_mb_tokens: 10000
+            sequence_length_round: 64
+            train_mb_tokens: 10000
+        tokenizer:
+            chat_template_kwargs:
+                enable_thinking: false
+            name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct
+        train_global_batch_size: 16
+        train_micro_batch_size: 1

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs/exp_001/wandb/wandb/run-20260502_204022-lec8l6gk/files/requirements.txt

@@ -0,0 +1,290 @@
+colorama==0.4.6
+psutil==7.2.2
+hydra-core==1.3.2
+tensorboard==2.20.0
+dnspython==2.8.0
+mdurl==0.1.2
+httpx==0.28.1
+ninja==1.13.0
+nvidia-ml-py==13.595.45
+google-api-core==2.30.3
+mistral_common==1.11.0
+mlflow-skinny==3.11.1
+python-multipart==0.0.27
+lark==1.2.2
+python-dotenv==1.2.2
+huggingface_hub==0.36.2
+cupy-cuda12x==14.0.1
+openai-harmony==0.0.8
+math-verify==0.9.0
+pyasn1_modules==0.4.2
+Mako==1.3.12
+email-validator==2.3.0
+pyparsing==3.3.2
+xformers==0.0.33.post1
+nvidia-nvjitlink-cu12==12.8.93
+dill==0.4.1
+graphql-core==3.2.8
+graphql-relay==3.2.0
+jiter==0.14.0
+sentencepiece==0.2.1
+msgpack==1.1.2
+multiprocess==0.70.19
+click==8.3.3
+nvidia-cufft-cu12==11.3.3.83
+protobuf==6.33.6
+interegular==0.3.3
+asttokens==3.0.1
+cuda-bindings==13.2.0
+blake3==1.0.8
+py-spy==0.4.2
+python-json-logger==4.1.0
+outlines_core==0.2.11
+openai==2.32.0
+contourpy==1.3.3
+mpmath==1.3.0
+annotated-doc==0.0.4
+nvidia-nccl-cu12==2.27.5
+distro==1.9.0
+wandb==0.25.1
+aiohttp==3.13.5
+opentelemetry-semantic-conventions==0.62b1
+pyecharts==2.1.0
+aiohttp-cors==0.8.1
+triton==3.5.0
+kiwisolver==1.5.0
+shellingham==1.5.4
+einops==0.8.2
+setproctitle==1.3.7
+wrapt==2.1.2
+aiohappyeyeballs==2.6.1
+annotated-types==0.7.0
+nvidia-cuda-nvrtc-cu12==12.8.93
+watchfiles==1.1.1
+nvidia-cuda-cupti-cu12==12.8.90
+scipy==1.17.1
+matplotlib==3.10.8
+stack-data==0.6.3
+absl-py==2.4.0
+antlr4-python3-runtime==4.9.3
+tzdata==2026.2
+distlib==0.4.0
+comm==0.2.3
+parso==0.8.6
+tensorboard-data-server==0.7.2
+ptyprocess==0.7.0
+opencv-python-headless==4.13.0.92
+msgspec==0.21.1
+uvicorn==0.46.0
+nvidia-cuda-runtime-cu12==12.8.90
+datasets==4.8.4
+requests==2.33.1
+Flask==3.1.3
+fastar==0.11.0
+executing==2.2.1
+Werkzeug==3.1.8
+nvidia-cutlass-dsl==4.4.2
+blobfile==3.2.0
+decord==0.6.0
+jsonschema==4.26.0
+llvmlite==0.44.0
+uvloop==0.22.1
+nvidia-cusparselt-cu12==0.7.1
+model-hosting-container-standards==0.1.14
+databricks-sdk==0.105.0
+nest-asyncio==1.6.0
+sentry-sdk==2.58.0
+cuda-python==13.2.0
+wcwidth==0.6.0
+Jinja2==3.1.6
+depyf==0.20.0
+nvidia-cufile-cu12==1.13.1.3
+colorful==0.5.8
+loguru==0.7.3
+blinker==1.9.0
+debugpy==1.8.20
+pycryptodomex==3.23.0
+regex==2026.4.4
+nvidia-cutlass-dsl-libs-base==4.4.2
+jedi==0.19.2
+python-discovery==1.2.2
+pydantic-extra-types==2.11.1
+mlflow-tracing==3.11.1
+opentelemetry-sdk==1.41.1
+anyio==4.13.0
+GitPython==3.1.48
+setuptools==82.0.1
+PyYAML==6.0.3
+starlette==0.52.1
+multidict==6.7.1
+attrs==26.1.0
+pip==26.0.1
+sympy==1.14.0
+pyasn1==0.6.3
+sqlparse==0.5.5
+opentelemetry-proto==1.41.1
+astor==0.8.1
+peft==0.18.1
+torchdata==0.11.0
+compressed-tensors==0.12.2
+Pygments==2.20.0
+transformers==4.57.1
+cryptography==46.0.7
+plotly==6.7.0
+typing-inspection==0.4.2
+flashinfer-python==0.5.2
+nvtx==0.2.15
+ipython_pygments_lexers==1.1.1
+rpds-py==0.30.0
+pillow==12.2.0
+accelerate==1.13.0
+yarl==1.23.0
+grpcio==1.80.0
+numba==0.61.2
+omegaconf==2.3.0
+googleapis-common-protos==1.74.0
+joblib==1.5.3
+graphene==3.4.3
+num2words==0.5.14
+docstring_parser==0.18.0
+referencing==0.37.0
+lxml==6.1.0
+smmap==5.0.3
+jsonschema-specifications==2025.9.1
+MarkupSafe==3.0.3
+opencensus-context==0.1.3
+charset-normalizer==3.4.7
+torchaudio==2.9.0
+nvidia-cusolver-cu12==11.7.3.90
+safetensors==0.7.0
+sniffio==1.3.1
+gguf==0.18.0
+threadpoolctl==3.6.0
+filelock==3.29.0
+pexpect==4.9.0
+networkx==3.6.1
+idna==3.13
+aiosignal==1.4.0
+fastapi-cloud-cli==0.17.1
+py-cpuinfo==9.0.0
+skops==0.14.0
+nvidia-cusparse-cu12==12.5.8.93
+flask-cors==6.0.2
+decorator==5.2.1
+nvidia-nvtx-cu12==12.8.90
+importlib_metadata==8.7.1
+rich==13.9.4
+opentelemetry-exporter-prometheus==0.62b1
+fastapi-cli==0.0.24
+pybase64==1.4.3
+pydantic-settings==2.14.0
+nvidia-cublas-cu12==12.8.4.1
+wheel==0.47.0
+smart_open==7.6.0
+xgrammar==0.1.25
+llguidance==1.3.0
+jupyter_client==8.8.0
+mlflow==3.11.1
+fastapi==0.136.1
+jmespath==1.1.0
+ipython==9.13.0
+colored==2.2.3
+pycparser==3.0
+anthropic==0.71.0
+pyzmq==27.1.0
+proto-plus==1.27.2
+prompt_toolkit==3.0.52
+python-dateutil==2.9.0.post0
+SQLAlchemy==2.0.49
+psutil==7.2.2
+urllib3==2.6.3
+httptools==0.7.1
+google-auth==2.49.2
+tornado==6.5.5
+gitdb==4.0.12
+cachetools==7.0.6
+narwhals==2.20.0
+rignore==0.7.6
+latex2sympy2_extended==1.11.0
+tabulate==0.10.0
+pytz==2026.1.post1
+prometheus_client==0.25.0
+h11==0.16.0
+partial-json-parser==0.2.1.1.post7
+gunicorn==25.3.0
+torch==2.9.0
+cloudpickle==3.1.2
+fonttools==4.62.1
+tiktoken==0.12.0
+scikit-learn==1.8.0
+cuda-pathfinder==1.5.4
+greenlet==3.5.0
+pycountry==26.2.16
+tokenizers==0.22.2
+propcache==0.4.1
+nvidia-cudnn-frontend==1.22.1
+virtualenv==21.3.0
+zipp==3.23.1
+ray==2.54.1
+alembic==1.18.4
+docker==7.1.0
+opentelemetry-api==1.41.1
+swanlab==0.7.16
+packaging==26.2
+diskcache==5.6.3
+numpy==2.2.6
+vllm==0.11.2
+typer==0.25.0
+pandas==2.3.3
+fsspec==2026.2.0
+tqdm==4.67.3
+nvidia-cudnn-cu12==9.10.2.21
+torchvision==0.24.0
+cycler==0.12.1
+nvidia-nvshmem-cu12==3.3.20
+typing_extensions==4.15.0
+hf-xet==1.4.3
+pyarrow==23.0.1
+prometheus-fastapi-instrumentator==7.1.0
+traitlets==5.14.3
+itsdangerous==2.2.0
+platformdirs==4.9.6
+cffi==2.0.0
+opencensus==0.11.4
+ipykernel==7.2.0
+lm-format-enforcer==0.11.3
+certifi==2026.4.22
+supervisor==4.3.0
+pydantic_core==2.46.3
+markdown-it-py==4.0.0
+httpcore==1.0.9
+docopt==0.6.2
+matplotlib-inline==0.2.1
+Markdown==3.10.2
+six==1.17.0
+cbor2==5.9.0
+jupyter_core==5.9.1
+pydantic==2.13.3
+frozenlist==1.8.0
+websockets==16.0
+rich-toolkit==0.19.7
+xxhash==3.7.0
+simplejson==4.1.1
+hf_transfer==0.1.9
+huey==2.6.0
+pure_eval==0.2.3
+apache-tvm-ffi==0.1.10
+nvidia-curand-cu12==10.3.9.90
+prettytable==3.17.0
+zipp==3.23.0
+tomli==2.4.0
+jaraco.text==4.0.0
+backports.tarfile==1.2.0
+more-itertools==10.8.0
+packaging==26.0
+autocommand==2.2.2
+platformdirs==4.4.0
+importlib_metadata==8.7.1
+wheel==0.46.3
+jaraco.context==6.1.0
+jaraco.functools==4.4.0

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs/exp_001/wandb/wandb/run-20260502_204022-lec8l6gk/files/wandb-metadata.json

@@ -0,0 +1,103 @@
+{
+  "os": "Linux-4.18.0-477.27.1.el8_8.x86_64-x86_64-with-glibc2.28",
+  "python": "CPython 3.11.15",
+  "startedAt": "2026-05-02T11:40:22.035189Z",
+  "args": [
+    "--config",
+    "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/opsd_distillation_config.yaml",
+    "policy.model_name=/home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct",
+    "policy.dtensor_cfg.tensor_parallel_size=4",
+    "policy.dtensor_cfg.context_parallel_size=1",
+    "policy.max_total_sequence_length=10000",
+    "policy.generation.max_new_tokens=8192",
+    "policy.generation.vllm_cfg.tensor_parallel_size=4",
+    "policy.generation.vllm_cfg.gpu_memory_utilization=0.7",
+    "policy.logprob_batch_size=1",
+    "policy.dtensor_cfg.cpu_offload=true",
+    "policy.dtensor_cfg.activation_checkpointing=true",
+    "paths.base_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd",
+    "paths.data_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/data/knowledge_d5_e10/atomic_sft",
+    "checkpointing.checkpoint_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716",
+    "logger.log_dir=/home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs",
+    "logger.wandb.project=mix-teachers-opsd",
+    "logger.wandb.name=opsd-knowledge-forward-kl-gt-d5-e10-20260502-Llama-3.2-1B-Instruct",
+    "logger.tensorboard.log_dir=tb_logs-opsd-knowledge-forward-kl-gt-d5-e10-llama3-2-1b",
+    "cluster.gpus_per_node=4",
+    "cluster.num_nodes=1",
+    "loss_fn.kl_type=forward",
+    "loss_fn.zero_outside_topk=false",
+    "distillation.ema_decay=1.0",
+    "distillation.topk_logits_k=64",
+    "distillation.num_prompts_per_step=20",
+    "distillation.num_generations_per_prompt=8",
+    "distillation.val_period=20",
+    "distillation.max_num_epochs=40",
+    "checkpointing.save_per_epoch=false",
+    "checkpointing.save_period=20",
+    "checkpointing.save_optimizer=false",
+    "checkpointing.keep_top_k=null",
+    "checkpointing.metric_name=null",
+    "policy.train_global_batch_size=20",
+    "policy.generation_batch_size=20",
+    "policy.generation.temperature=0.8",
+    "policy.optimizer.kwargs.lr=3.0e-5",
+    "policy.optimizer.kwargs.weight_decay=0.01",
+    "policy.scheduler=[{name: \"torch.optim.lr_scheduler.ConstantLR\", kwargs: {factor: 1.0, total_iters: 10000000000}}, {milestones: []}]"
+  ],
+  "program": "/home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd/run_opsd.py",
+  "codePath": "mix_teachers/train/knowledge_opsd/run_opsd.py",
+  "codePathLocal": "run_opsd.py",
+  "git": {
+    "remote": "https://github.com/jiarui-liu/TeacherMix.git",
+    "commit": "9938bfa2b66c98512427baa8c5adbea81dbf6dc8"
+  },
+  "email": "jiaruil5@andrew.cmu.edu",
+  "root": "/home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs/exp_001/wandb",
+  "host": "b0017",
+  "executable": "/home/pj24003162/ku40003404/miniconda3/envs/opsd/bin/python3",
+  "cpu_count": 120,
+  "cpu_count_logical": 120,
+  "gpu": "NVIDIA H100",
+  "gpu_count": 4,
+  "disk": {
+    "/": {
+      "total": "511749718016",
+      "used": "25329385472"
+    }
+  },
+  "memory": {
+    "total": "1081436749824"
+  },
+  "gpu_nvidia": [
+    {
+      "name": "NVIDIA H100",
+      "memoryTotal": "100485038080",
+      "cudaCores": 16896,
+      "architecture": "Hopper",
+      "uuid": "GPU-f879d657-3497-66cd-daca-a4f8d1d73359"
+    },
+    {
+      "name": "NVIDIA H100",
+      "memoryTotal": "100485038080",
+      "cudaCores": 16896,
+      "architecture": "Hopper",
+      "uuid": "GPU-4622e484-3226-d178-7e48-7aae340aad13"
+    },
+    {
+      "name": "NVIDIA H100",
+      "memoryTotal": "100485038080",
+      "cudaCores": 16896,
+      "architecture": "Hopper",
+      "uuid": "GPU-ad99b84e-fcc6-7236-3d91-6cbaa83e680b"
+    },
+    {
+      "name": "NVIDIA H100",
+      "memoryTotal": "100485038080",
+      "cudaCores": 16896,
+      "architecture": "Hopper",
+      "uuid": "GPU-2681ca36-95a6-471b-e939-d814aa0a1dd3"
+    }
+  ],
+  "cudaVersion": "13.2",
+  "writerId": "8pnbx653goz1fz52mzj1yeg80v3inpgq"
+}

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs/exp_001/wandb/wandb/run-20260502_204022-lec8l6gk/files/wandb-summary.json

@@ -0,0 +1 @@
+{"_runtime":34424.903274576,"ray/node.0.mem_total_gb":1007.1664581298828,"ray/node.0.gpu.1.mem_gb":68.9658203125,"timing/train/get_topk_logits/submit_topk_logits_futures":0.02265331894159317,"timing/train/teacher_logprob_inference":20.40494916960597,"timing/train/prepare_for_generation/transfer_and_update_weights":1.1734567312523723,"train/total_turns":160,"timing/train/valid_tokens_per_sec_per_gpu":41.724503181979614,"train/loss":0.009389259281306295,"train/mean_total_tokens_per_sample":2391.518798828125,"timing/train/policy_training/submit_training_futures":2.6412849347107112,"timing/validation/total_validation_time":136.29101202869788,"train/lr":3.000000000000001e-05,"train/avg_turns_per_sample":1,"train/num_valid_samples":20,"ray/node.0.mem_gb":100.28120040893555,"train/mean_env_tokens_per_sample":3,"train/total_num_tokens":392331,"train/global_valid_toks":50598.01587301587,"train/tokens_per_sec_per_gpu":323.52683747465545,"train/natural_termination_rate":1,"timing/train/policy_training":91.18469234928489,"timing/train/checkpointing":2.385032492224127,"ray/node.0.gpu.2.mem_gb":69.1435546875,"validation/avg_length":2686.1770833333335,"train/max_gen_tokens_per_sample":8192,"timing/train/generation":44.99103088118136,"_timestamp":1.7777564564408863e+09,"ray/node.0.gpu.0.mem_gb":68.9658203125,"validation/accuracy/compositional_2step_sft-test_messages":0,"timing/train/prepare_for_generation":2.858971363864839,"timing/train/data_processing":0.02707233652472496,"ray/node.0.gpu.3.mem_gb":68.9658203125,"ray/node.0.gpu.1.util":46,"timing/train/teacher_logprob_inference_prep":0.4878837880678475,"train/mean_prompt_length":60.55,"train/grad_norm":0.5598791837692261,"validation/accuracy/atomic_sft-val_paraphrased_messages":0.04,"train/accuracy":0.06875,"ray/node.0.gpu.0.util":45,"ray/node.0.gpu.2.util":44,"validation/accuracy":0.020833333333333332,"train/global_valid_seqs":20,"ray/node.0.gpu.3.util":45,"timing/train/get_topk_logits/shard_data":0.0374897588044405,"timing/train/policy_training/sharding_data":4.439121197909117,"ray/ray_step":34412,"train/max_turns_reached_rate":0,"timing/train/training_prep":0.9035933441482484,"timing/train/total_step_time":303.1672759070061,"train/truncation_rate":0.03750000149011612,"train/mean_gen_tokens_per_sample":2388.518798828125,"train/max_turns_per_sample":1,"_wandb":{"runtime":34424},"_step":200}

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs/exp_001/wandb/wandb/run-20260502_204022-lec8l6gk/logs/debug-core.log

@@ -0,0 +1,22 @@
+{"time":"2026-05-02T20:40:30.123059211+09:00","level":"INFO","msg":"main: starting server","port-filename":"/tmp/5738897_b3BzZC1rbm93LWZrbC1ndC1kNWUxMC1sbGFtYTMtMi0xYgo=/tmpw348nrd9/port-333.txt","pid":333,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-05-02T20:40:30.123068758+09:00","level":"INFO","msg":"main: starting server","port-filename":"/tmp/5738898_b3BzZC1rbm93LWZrbC1ndC1kNWUxMC1sbGFtYTMtMi0zYgo=/tmpivg8qor2/port-333.txt","pid":333,"log-level":0,"disable-analytics":false,"shutdown-on-parent-exit":false,"enable-dcgm-profiling":false}
+{"time":"2026-05-02T20:40:30.13831872+09:00","level":"INFO","msg":"server: will exit if parent process dies","ppid":333}
+{"time":"2026-05-02T20:40:30.13880721+09:00","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/5738897_b3BzZC1rbm93LWZrbC1ndC1kNWUxMC1sbGFtYTMtMi0xYgo=/wandb-333-8717-1688308034/socket","Net":"unix"}}
+{"time":"2026-05-02T20:40:30.143907114+09:00","level":"INFO","msg":"server: will exit if parent process dies","ppid":333}
+{"time":"2026-05-02T20:40:30.144705238+09:00","level":"INFO","msg":"server: accepting connections","addr":{"Name":"/tmp/5738898_b3BzZC1rbm93LWZrbC1ndC1kNWUxMC1sbGFtYTMtMi0zYgo=/wandb-333-8379-3397703323/socket","Net":"unix"}}
+{"time":"2026-05-02T20:40:30.319208816+09:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-05-02T20:40:30.319210098+09:00","level":"INFO","msg":"connection: ManageConnectionData: new connection created","id":"1(@)"}
+{"time":"2026-05-02T20:40:30.396810719+09:00","level":"INFO","msg":"handleInformInit: received","streamId":"lec8l6gk","id":"1(@)"}
+{"time":"2026-05-02T20:40:30.396852027+09:00","level":"INFO","msg":"handleInformInit: received","streamId":"jaynkchc","id":"1(@)"}
+{"time":"2026-05-02T20:40:32.153930035+09:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"lec8l6gk","id":"1(@)"}
+{"time":"2026-05-02T20:40:32.162051171+09:00","level":"INFO","msg":"handleInformInit: stream started","streamId":"jaynkchc","id":"1(@)"}
+{"time":"2026-05-02T20:40:40.959491447+09:00","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"ydkgyjqlflnz"}
+{"time":"2026-05-02T20:40:40.961544217+09:00","level":"INFO","msg":"connection: cancelling request","id":"1(@)","requestId":"gg18tydj9vtd"}
+{"time":"2026-05-03T06:14:17.389642695+09:00","level":"INFO","msg":"handleInformTeardown: server teardown initiated","id":"1(@)"}
+{"time":"2026-05-03T06:14:17.396598276+09:00","level":"INFO","msg":"server is shutting down"}
+{"time":"2026-05-03T06:14:17.403620033+09:00","level":"INFO","msg":"server: listener closed","addr":{"Name":"/tmp/5738897_b3BzZC1rbm93LWZrbC1ndC1kNWUxMC1sbGFtYTMtMi0xYgo=/wandb-333-8717-1688308034/socket","Net":"unix"}}
+{"time":"2026-05-03T06:14:17.39659316+09:00","level":"INFO","msg":"connection: closing","id":"1(@)"}
+{"time":"2026-05-03T06:14:17.416424237+09:00","level":"INFO","msg":"connection: closed successfully","id":"1(@)"}
+{"time":"2026-05-03T06:14:21.480547528+09:00","level":"INFO","msg":"handleInformTeardown: server shutdown complete","id":"1(@)"}
+{"time":"2026-05-03T06:14:21.48125889+09:00","level":"INFO","msg":"connection: ManageConnectionData: connection closed","id":"1(@)"}
+{"time":"2026-05-03T06:14:21.481864166+09:00","level":"INFO","msg":"server is closed"}

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/step_120/config.yaml

@@ -0,0 +1,255 @@
+checkpointing:
+  checkpoint_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716
+  checkpoint_must_save_by: null
+  enabled: true
+  higher_is_better: true
+  keep_top_k: null
+  metric_name: null
+  model_save_format: safetensors
+  save_consolidated: false
+  save_optimizer: false
+  save_per_epoch: false
+  save_period: 20
+cluster:
+  gpus_per_node: 4
+  num_nodes: 1
+data:
+  default:
+    env_name: knowledge
+    prompt_file: null
+    system_prompt_file: null
+  max_input_seq_length: 10000
+  shuffle: true
+  train:
+  - dataset_name: knowledge_opsd
+    env_name: knowledge
+    processor: knowledge_opsd
+    prompt_file: null
+    system_prompt_file: null
+  validation:
+  - dataset_name: knowledge_opsd_val
+    env_name: knowledge
+    processor: knowledge_opsd
+    prompt_file: null
+    system_prompt_file: null
+  - dataset_name: knowledge_opsd_comp_val
+    env_name: knowledge
+    processor: knowledge_opsd
+    prompt_file: null
+    system_prompt_file: null
+distillation:
+  ema_decay: 1.0
+  max_num_epochs: 40
+  max_num_steps: 1000
+  max_rollout_turns: 1
+  max_val_samples: 200
+  num_generations_per_prompt: 8
+  num_prompts_per_step: 20
+  rollout_policy: student
+  seed: 42
+  topk_logits_k: 64
+  val_at_end: true
+  val_at_start: true
+  val_batch_size: 32
+  val_period: 20
+env:
+  knowledge:
+    num_workers: 8
+logger:
+  gpu_monitoring:
+    collection_interval: 10
+    flush_interval: 10
+  log_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/logs/exp_001
+  mlflow_enabled: false
+  monitor_gpus: true
+  num_val_samples_to_print: 5
+  swanlab_enabled: false
+  tensorboard:
+    log_dir: tb_logs-opsd-knowledge-forward-kl-gt-d5-e10-llama3-2-1b
+  tensorboard_enabled: true
+  wandb:
+    entity: mathrl-jiarui
+    name: opsd-knowledge-forward-kl-gt-d5-e10-20260502-Llama-3.2-1B-Instruct
+    project: mix-teachers-opsd
+  wandb_enabled: true
+loss_fn:
+  kl_type: forward
+  mixed_kl_weight: 0.5
+  zero_outside_topk: false
+paths:
+  base_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/mix_teachers/train/knowledge_opsd
+  checkpoint_dir: /home/fs01/yw2349/jiarui/mathrl_data/mix_teachers/checkpoints/knowledge/opsd_nemo_rl
+  data_dir: /home/pj24003162/ku40003404/jiarui/TeacherMix/data/knowledge_d5_e10/atomic_sft
+  log_dir: /home/fs01/yw2349/jiarui/mathrl_data/mix_teachers/checkpoints/knowledge/opsd_nemo_rl/logs
+  model_search_dirs:
+  - /home/fs01/yw2349/jiarui/mathrl_data/models
+policy:
+  dtensor_cfg:
+    _v2: true
+    activation_checkpointing: true
+    context_parallel_size: 1
+    cpu_offload: true
+    custom_parallel_plan: null
+    enabled: true
+    sequence_parallel: false
+    tensor_parallel_size: 4
+  dynamic_batching:
+    enabled: true
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  generation:
+    _pad_token_id: 128009
+    backend: vllm
+    colocated:
+      enabled: true
+      resources:
+        gpus_per_node: null
+        num_nodes: null
+    max_new_tokens: 8192
+    model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct
+    stop_strings: null
+    stop_token_ids:
+    - 128009
+    temperature: 0.8
+    top_k: null
+    top_p: 1.0
+    vllm_cfg:
+      async_engine: false
+      distributed_executor_backend: null
+      enforce_eager: false
+      expert_parallel_size: 1
+      gpu_memory_utilization: 0.7
+      hf_overrides: {}
+      kv_cache_dtype: auto
+      load_format: dummy
+      max_model_len: 10000
+      num_first_layers_in_bf16: 0
+      num_last_layers_in_bf16: 0
+      pipeline_parallel_size: 1
+      precision: bfloat16
+      skip_tokenizer_init: true
+      tensor_parallel_size: 4
+      use_deep_gemm: false
+  generation_batch_size: 20
+  logprob_batch_size: 1
+  logprob_chunk_size: null
+  make_sequence_length_divisible_by: 8
+  max_grad_norm: 1.0
+  max_total_sequence_length: 10000
+  megatron_cfg:
+    enabled: false
+  model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct
+  offload_optimizer_for_logprob: false
+  optimizer:
+    kwargs:
+      betas:
+      - 0.9
+      - 0.95
+      eps: 1.0e-08
+      foreach: false
+      fused: false
+      lr: 3.0e-05
+      weight_decay: 0.01
+    name: torch.optim.AdamW
+  precision: bfloat16
+  scheduler:
+  - kwargs:
+      factor: 1.0
+      total_iters: 10000000000
+    name: torch.optim.lr_scheduler.ConstantLR
+  - milestones: []
+  sequence_packing:
+    algorithm: modified_first_fit_decreasing
+    enabled: false
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  tokenizer:
+    chat_template_kwargs:
+      enable_thinking: false
+    name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct
+  train_global_batch_size: 20
+  train_micro_batch_size: 1
+teacher:
+  dtensor_cfg:
+    _v2: true
+    activation_checkpointing: true
+    context_parallel_size: 1
+    cpu_offload: true
+    custom_parallel_plan: null
+    enabled: true
+    sequence_parallel: false
+    tensor_parallel_size: 4
+  dynamic_batching:
+    enabled: true
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  generation:
+    backend: vllm
+    colocated:
+      enabled: true
+      resources:
+        gpus_per_node: null
+        num_nodes: null
+    max_new_tokens: 8192
+    stop_strings: null
+    stop_token_ids: null
+    temperature: 1.0
+    top_k: null
+    top_p: 1.0
+    vllm_cfg:
+      async_engine: false
+      distributed_executor_backend: null
+      enforce_eager: false
+      expert_parallel_size: 1
+      gpu_memory_utilization: 0.7
+      kv_cache_dtype: auto
+      max_model_len: 10000
+      num_first_layers_in_bf16: 0
+      num_last_layers_in_bf16: 0
+      pipeline_parallel_size: 1
+      precision: bfloat16
+      tensor_parallel_size: 4
+      use_deep_gemm: false
+  generation_batch_size: 16
+  logprob_batch_size: 1
+  logprob_chunk_size: null
+  make_sequence_length_divisible_by: 8
+  max_grad_norm: 1.0
+  max_total_sequence_length: 10000
+  megatron_cfg:
+    enabled: false
+  model_name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct
+  offload_optimizer_for_logprob: false
+  optimizer:
+    kwargs:
+      betas:
+      - 0.9
+      - 0.95
+      eps: 1.0e-08
+      foreach: false
+      fused: false
+      lr: 1.0e-05
+      weight_decay: 0.01
+    name: torch.optim.AdamW
+  precision: bfloat16
+  scheduler:
+  - kwargs:
+      factor: 1.0
+      total_iters: 10000000000
+    name: torch.optim.lr_scheduler.ConstantLR
+  - milestones: []
+  sequence_packing:
+    algorithm: modified_first_fit_decreasing
+    enabled: false
+    logprob_mb_tokens: 10000
+    sequence_length_round: 64
+    train_mb_tokens: 10000
+  tokenizer:
+    chat_template_kwargs:
+      enable_thinking: false
+    name: /home/pj24003162/ku40003404/jiarui/models/Llama-3.2-1B-Instruct
+  train_global_batch_size: 16
+  train_micro_batch_size: 1

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/step_120/policy/tokenizer/chat_template.jinja

@@ -0,0 +1,93 @@
+{{- bos_token }}
+{%- if custom_tools is defined %}
+    {%- set tools = custom_tools %}
+{%- endif %}
+{%- if not tools_in_user_message is defined %}
+    {%- set tools_in_user_message = true %}
+{%- endif %}
+{%- if not date_string is defined %}
+    {%- if strftime_now is defined %}
+        {%- set date_string = strftime_now("%d %b %Y") %}
+    {%- else %}
+        {%- set date_string = "26 Jul 2024" %}
+    {%- endif %}
+{%- endif %}
+{%- if not tools is defined %}
+    {%- set tools = none %}
+{%- endif %}
+
+{#- This block extracts the system message, so we can slot it into the right place. #}
+{%- if messages[0]['role'] == 'system' %}
+    {%- set system_message = messages[0]['content']|trim %}
+    {%- set messages = messages[1:] %}
+{%- else %}
+    {%- set system_message = "" %}
+{%- endif %}
+
+{#- System message #}
+{{- "<|start_header_id|>system<|end_header_id|>\n\n" }}
+{%- if tools is not none %}
+    {{- "Environment: ipython\n" }}
+{%- endif %}
+{{- "Cutting Knowledge Date: December 2023\n" }}
+{{- "Today Date: " + date_string + "\n\n" }}
+{%- if tools is not none and not tools_in_user_message %}
+    {{- "You have access to the following functions. To call a function, please respond with JSON for a function call." }}
+    {{- 'Respond in the format {"name": function name, "parameters": dictionary of argument name and its value}.' }}
+    {{- "Do not use variables.\n\n" }}
+    {%- for t in tools %}
+        {{- t | tojson(indent=4) }}
+        {{- "\n\n" }}
+    {%- endfor %}
+{%- endif %}
+{{- system_message }}
+{{- "<|eot_id|>" }}
+
+{#- Custom tools are passed in a user message with some extra guidance #}
+{%- if tools_in_user_message and not tools is none %}
+    {#- Extract the first user message so we can plug it in here #}
+    {%- if messages | length != 0 %}
+        {%- set first_user_message = messages[0]['content']|trim %}
+        {%- set messages = messages[1:] %}
+    {%- else %}
+        {{- raise_exception("Cannot put tools in the first user message when there's no first user message!") }}
+{%- endif %}
+    {{- '<|start_header_id|>user<|end_header_id|>\n\n' -}}
+    {{- "Given the following functions, please respond with a JSON for a function call " }}
+    {{- "with its proper arguments that best answers the given prompt.\n\n" }}
+    {{- 'Respond in the format {"name": function name, "parameters": dictionary of argument name and its value}.' }}
+    {{- "Do not use variables.\n\n" }}
+    {%- for t in tools %}
+        {{- t | tojson(indent=4) }}
+        {{- "\n\n" }}
+    {%- endfor %}
+    {{- first_user_message + "<|eot_id|>"}}
+{%- endif %}
+
+{%- for message in messages %}
+    {%- if not (message.role == 'ipython' or message.role == 'tool' or 'tool_calls' in message) %}
+        {{- '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' }}
+    {%- elif 'tool_calls' in message %}
+        {%- if not message.tool_calls|length == 1 %}
+            {{- raise_exception("This model only supports single tool-calls at once!") }}
+        {%- endif %}
+        {%- set tool_call = message.tool_calls[0].function %}
+        {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' -}}
+        {{- '{"name": "' + tool_call.name + '", ' }}
+        {{- '"parameters": ' }}
+        {{- tool_call.arguments | tojson }}
+        {{- "}" }}
+        {{- "<|eot_id|>" }}
+    {%- elif message.role == "tool" or message.role == "ipython" %}
+        {{- "<|start_header_id|>ipython<|end_header_id|>\n\n" }}
+        {%- if message.content is mapping or message.content is iterable %}
+            {{- message.content | tojson }}
+        {%- else %}
+            {{- message.content }}
+        {%- endif %}
+        {{- "<|eot_id|>" }}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|start_header_id|>assistant<|end_header_id|>\n\n' }}
+{%- endif %}

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/step_120/policy/tokenizer/special_tokens_map.json

@@ -0,0 +1,17 @@
+{
+  "bos_token": {
+    "content": "<|begin_of_text|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|eot_id|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<|eot_id|>"
+}

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/step_120/policy/tokenizer/tokenizer_config.json

@@ -0,0 +1,2063 @@
+{
+  "added_tokens_decoder": {
+    "128000": {
+      "content": "<|begin_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128001": {
+      "content": "<|end_of_text|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128002": {
+      "content": "<|reserved_special_token_0|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128003": {
+      "content": "<|reserved_special_token_1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128004": {
+      "content": "<|finetune_right_pad_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128005": {
+      "content": "<|reserved_special_token_2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128006": {
+      "content": "<|start_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128007": {
+      "content": "<|end_header_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128008": {
+      "content": "<|eom_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128009": {
+      "content": "<|eot_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128010": {
+      "content": "<|python_tag|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128011": {
+      "content": "<|reserved_special_token_3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128012": {
+      "content": "<|reserved_special_token_4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128013": {
+      "content": "<|reserved_special_token_5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128014": {
+      "content": "<|reserved_special_token_6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128015": {
+      "content": "<|reserved_special_token_7|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128016": {
+      "content": "<|reserved_special_token_8|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128017": {
+      "content": "<|reserved_special_token_9|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128018": {
+      "content": "<|reserved_special_token_10|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128019": {
+      "content": "<|reserved_special_token_11|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128020": {
+      "content": "<|reserved_special_token_12|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128021": {
+      "content": "<|reserved_special_token_13|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128022": {
+      "content": "<|reserved_special_token_14|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128023": {
+      "content": "<|reserved_special_token_15|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128024": {
+      "content": "<|reserved_special_token_16|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128025": {
+      "content": "<|reserved_special_token_17|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128026": {
+      "content": "<|reserved_special_token_18|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128027": {
+      "content": "<|reserved_special_token_19|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128028": {
+      "content": "<|reserved_special_token_20|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128029": {
+      "content": "<|reserved_special_token_21|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128030": {
+      "content": "<|reserved_special_token_22|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128031": {
+      "content": "<|reserved_special_token_23|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128032": {
+      "content": "<|reserved_special_token_24|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128033": {
+      "content": "<|reserved_special_token_25|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128034": {
+      "content": "<|reserved_special_token_26|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128035": {
+      "content": "<|reserved_special_token_27|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128036": {
+      "content": "<|reserved_special_token_28|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128037": {
+      "content": "<|reserved_special_token_29|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128038": {
+      "content": "<|reserved_special_token_30|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128039": {
+      "content": "<|reserved_special_token_31|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128040": {
+      "content": "<|reserved_special_token_32|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128041": {
+      "content": "<|reserved_special_token_33|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128042": {
+      "content": "<|reserved_special_token_34|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128043": {
+      "content": "<|reserved_special_token_35|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128044": {
+      "content": "<|reserved_special_token_36|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128045": {
+      "content": "<|reserved_special_token_37|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128046": {
+      "content": "<|reserved_special_token_38|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128047": {
+      "content": "<|reserved_special_token_39|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128048": {
+      "content": "<|reserved_special_token_40|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128049": {
+      "content": "<|reserved_special_token_41|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128050": {
+      "content": "<|reserved_special_token_42|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128051": {
+      "content": "<|reserved_special_token_43|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128052": {
+      "content": "<|reserved_special_token_44|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128053": {
+      "content": "<|reserved_special_token_45|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128054": {
+      "content": "<|reserved_special_token_46|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128055": {
+      "content": "<|reserved_special_token_47|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128056": {
+      "content": "<|reserved_special_token_48|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128057": {
+      "content": "<|reserved_special_token_49|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128058": {
+      "content": "<|reserved_special_token_50|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128059": {
+      "content": "<|reserved_special_token_51|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128060": {
+      "content": "<|reserved_special_token_52|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128061": {
+      "content": "<|reserved_special_token_53|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128062": {
+      "content": "<|reserved_special_token_54|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128063": {
+      "content": "<|reserved_special_token_55|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128064": {
+      "content": "<|reserved_special_token_56|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128065": {
+      "content": "<|reserved_special_token_57|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128066": {
+      "content": "<|reserved_special_token_58|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128067": {
+      "content": "<|reserved_special_token_59|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128068": {
+      "content": "<|reserved_special_token_60|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128069": {
+      "content": "<|reserved_special_token_61|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128070": {
+      "content": "<|reserved_special_token_62|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128071": {
+      "content": "<|reserved_special_token_63|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128072": {
+      "content": "<|reserved_special_token_64|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128073": {
+      "content": "<|reserved_special_token_65|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128074": {
+      "content": "<|reserved_special_token_66|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128075": {
+      "content": "<|reserved_special_token_67|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128076": {
+      "content": "<|reserved_special_token_68|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128077": {
+      "content": "<|reserved_special_token_69|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128078": {
+      "content": "<|reserved_special_token_70|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128079": {
+      "content": "<|reserved_special_token_71|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128080": {
+      "content": "<|reserved_special_token_72|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128081": {
+      "content": "<|reserved_special_token_73|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128082": {
+      "content": "<|reserved_special_token_74|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128083": {
+      "content": "<|reserved_special_token_75|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128084": {
+      "content": "<|reserved_special_token_76|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128085": {
+      "content": "<|reserved_special_token_77|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128086": {
+      "content": "<|reserved_special_token_78|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128087": {
+      "content": "<|reserved_special_token_79|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128088": {
+      "content": "<|reserved_special_token_80|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128089": {
+      "content": "<|reserved_special_token_81|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128090": {
+      "content": "<|reserved_special_token_82|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128091": {
+      "content": "<|reserved_special_token_83|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128092": {
+      "content": "<|reserved_special_token_84|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128093": {
+      "content": "<|reserved_special_token_85|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128094": {
+      "content": "<|reserved_special_token_86|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128095": {
+      "content": "<|reserved_special_token_87|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128096": {
+      "content": "<|reserved_special_token_88|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128097": {
+      "content": "<|reserved_special_token_89|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128098": {
+      "content": "<|reserved_special_token_90|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128099": {
+      "content": "<|reserved_special_token_91|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128100": {
+      "content": "<|reserved_special_token_92|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128101": {
+      "content": "<|reserved_special_token_93|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128102": {
+      "content": "<|reserved_special_token_94|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128103": {
+      "content": "<|reserved_special_token_95|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128104": {
+      "content": "<|reserved_special_token_96|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128105": {
+      "content": "<|reserved_special_token_97|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128106": {
+      "content": "<|reserved_special_token_98|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128107": {
+      "content": "<|reserved_special_token_99|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128108": {
+      "content": "<|reserved_special_token_100|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128109": {
+      "content": "<|reserved_special_token_101|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128110": {
+      "content": "<|reserved_special_token_102|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128111": {
+      "content": "<|reserved_special_token_103|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128112": {
+      "content": "<|reserved_special_token_104|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128113": {
+      "content": "<|reserved_special_token_105|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128114": {
+      "content": "<|reserved_special_token_106|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128115": {
+      "content": "<|reserved_special_token_107|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128116": {
+      "content": "<|reserved_special_token_108|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128117": {
+      "content": "<|reserved_special_token_109|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128118": {
+      "content": "<|reserved_special_token_110|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128119": {
+      "content": "<|reserved_special_token_111|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128120": {
+      "content": "<|reserved_special_token_112|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128121": {
+      "content": "<|reserved_special_token_113|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128122": {
+      "content": "<|reserved_special_token_114|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128123": {
+      "content": "<|reserved_special_token_115|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128124": {
+      "content": "<|reserved_special_token_116|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128125": {
+      "content": "<|reserved_special_token_117|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128126": {
+      "content": "<|reserved_special_token_118|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128127": {
+      "content": "<|reserved_special_token_119|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128128": {
+      "content": "<|reserved_special_token_120|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128129": {
+      "content": "<|reserved_special_token_121|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128130": {
+      "content": "<|reserved_special_token_122|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128131": {
+      "content": "<|reserved_special_token_123|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128132": {
+      "content": "<|reserved_special_token_124|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128133": {
+      "content": "<|reserved_special_token_125|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128134": {
+      "content": "<|reserved_special_token_126|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128135": {
+      "content": "<|reserved_special_token_127|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128136": {
+      "content": "<|reserved_special_token_128|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128137": {
+      "content": "<|reserved_special_token_129|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128138": {
+      "content": "<|reserved_special_token_130|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128139": {
+      "content": "<|reserved_special_token_131|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128140": {
+      "content": "<|reserved_special_token_132|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128141": {
+      "content": "<|reserved_special_token_133|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128142": {
+      "content": "<|reserved_special_token_134|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128143": {
+      "content": "<|reserved_special_token_135|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128144": {
+      "content": "<|reserved_special_token_136|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128145": {
+      "content": "<|reserved_special_token_137|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128146": {
+      "content": "<|reserved_special_token_138|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128147": {
+      "content": "<|reserved_special_token_139|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128148": {
+      "content": "<|reserved_special_token_140|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128149": {
+      "content": "<|reserved_special_token_141|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128150": {
+      "content": "<|reserved_special_token_142|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128151": {
+      "content": "<|reserved_special_token_143|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128152": {
+      "content": "<|reserved_special_token_144|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128153": {
+      "content": "<|reserved_special_token_145|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128154": {
+      "content": "<|reserved_special_token_146|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128155": {
+      "content": "<|reserved_special_token_147|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128156": {
+      "content": "<|reserved_special_token_148|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128157": {
+      "content": "<|reserved_special_token_149|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128158": {
+      "content": "<|reserved_special_token_150|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128159": {
+      "content": "<|reserved_special_token_151|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128160": {
+      "content": "<|reserved_special_token_152|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128161": {
+      "content": "<|reserved_special_token_153|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128162": {
+      "content": "<|reserved_special_token_154|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128163": {
+      "content": "<|reserved_special_token_155|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128164": {
+      "content": "<|reserved_special_token_156|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128165": {
+      "content": "<|reserved_special_token_157|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128166": {
+      "content": "<|reserved_special_token_158|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128167": {
+      "content": "<|reserved_special_token_159|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128168": {
+      "content": "<|reserved_special_token_160|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128169": {
+      "content": "<|reserved_special_token_161|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128170": {
+      "content": "<|reserved_special_token_162|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128171": {
+      "content": "<|reserved_special_token_163|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128172": {
+      "content": "<|reserved_special_token_164|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128173": {
+      "content": "<|reserved_special_token_165|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128174": {
+      "content": "<|reserved_special_token_166|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128175": {
+      "content": "<|reserved_special_token_167|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128176": {
+      "content": "<|reserved_special_token_168|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128177": {
+      "content": "<|reserved_special_token_169|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128178": {
+      "content": "<|reserved_special_token_170|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128179": {
+      "content": "<|reserved_special_token_171|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128180": {
+      "content": "<|reserved_special_token_172|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128181": {
+      "content": "<|reserved_special_token_173|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128182": {
+      "content": "<|reserved_special_token_174|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128183": {
+      "content": "<|reserved_special_token_175|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128184": {
+      "content": "<|reserved_special_token_176|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128185": {
+      "content": "<|reserved_special_token_177|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128186": {
+      "content": "<|reserved_special_token_178|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128187": {
+      "content": "<|reserved_special_token_179|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128188": {
+      "content": "<|reserved_special_token_180|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128189": {
+      "content": "<|reserved_special_token_181|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128190": {
+      "content": "<|reserved_special_token_182|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128191": {
+      "content": "<|reserved_special_token_183|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128192": {
+      "content": "<|reserved_special_token_184|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128193": {
+      "content": "<|reserved_special_token_185|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128194": {
+      "content": "<|reserved_special_token_186|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128195": {
+      "content": "<|reserved_special_token_187|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128196": {
+      "content": "<|reserved_special_token_188|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128197": {
+      "content": "<|reserved_special_token_189|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128198": {
+      "content": "<|reserved_special_token_190|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128199": {
+      "content": "<|reserved_special_token_191|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128200": {
+      "content": "<|reserved_special_token_192|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128201": {
+      "content": "<|reserved_special_token_193|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128202": {
+      "content": "<|reserved_special_token_194|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128203": {
+      "content": "<|reserved_special_token_195|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128204": {
+      "content": "<|reserved_special_token_196|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128205": {
+      "content": "<|reserved_special_token_197|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128206": {
+      "content": "<|reserved_special_token_198|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128207": {
+      "content": "<|reserved_special_token_199|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128208": {
+      "content": "<|reserved_special_token_200|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128209": {
+      "content": "<|reserved_special_token_201|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128210": {
+      "content": "<|reserved_special_token_202|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128211": {
+      "content": "<|reserved_special_token_203|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128212": {
+      "content": "<|reserved_special_token_204|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128213": {
+      "content": "<|reserved_special_token_205|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128214": {
+      "content": "<|reserved_special_token_206|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128215": {
+      "content": "<|reserved_special_token_207|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128216": {
+      "content": "<|reserved_special_token_208|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128217": {
+      "content": "<|reserved_special_token_209|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128218": {
+      "content": "<|reserved_special_token_210|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128219": {
+      "content": "<|reserved_special_token_211|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128220": {
+      "content": "<|reserved_special_token_212|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128221": {
+      "content": "<|reserved_special_token_213|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128222": {
+      "content": "<|reserved_special_token_214|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128223": {
+      "content": "<|reserved_special_token_215|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128224": {
+      "content": "<|reserved_special_token_216|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128225": {
+      "content": "<|reserved_special_token_217|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128226": {
+      "content": "<|reserved_special_token_218|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128227": {
+      "content": "<|reserved_special_token_219|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128228": {
+      "content": "<|reserved_special_token_220|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128229": {
+      "content": "<|reserved_special_token_221|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128230": {
+      "content": "<|reserved_special_token_222|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128231": {
+      "content": "<|reserved_special_token_223|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128232": {
+      "content": "<|reserved_special_token_224|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128233": {
+      "content": "<|reserved_special_token_225|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128234": {
+      "content": "<|reserved_special_token_226|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128235": {
+      "content": "<|reserved_special_token_227|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128236": {
+      "content": "<|reserved_special_token_228|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128237": {
+      "content": "<|reserved_special_token_229|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128238": {
+      "content": "<|reserved_special_token_230|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128239": {
+      "content": "<|reserved_special_token_231|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128240": {
+      "content": "<|reserved_special_token_232|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128241": {
+      "content": "<|reserved_special_token_233|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128242": {
+      "content": "<|reserved_special_token_234|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128243": {
+      "content": "<|reserved_special_token_235|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128244": {
+      "content": "<|reserved_special_token_236|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<|reserved_special_token_237|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128246": {
+      "content": "<|reserved_special_token_238|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128247": {
+      "content": "<|reserved_special_token_239|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128248": {
+      "content": "<|reserved_special_token_240|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128249": {
+      "content": "<|reserved_special_token_241|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128250": {
+      "content": "<|reserved_special_token_242|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128251": {
+      "content": "<|reserved_special_token_243|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128252": {
+      "content": "<|reserved_special_token_244|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128253": {
+      "content": "<|reserved_special_token_245|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128254": {
+      "content": "<|reserved_special_token_246|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128255": {
+      "content": "<|reserved_special_token_247|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|begin_of_text|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|eot_id|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 131072,
+  "pad_token": "<|eot_id|>",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/step_120/policy/weights/model/.hf_metadata/config.json

@@ -0,0 +1,41 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "dtype": "float32",
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "head_dim": 64,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "max_position_embeddings": 131072,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "nemo_version": "0.2.0",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 16,
+  "num_key_value_heads": 8,
+  "pad_token_id": 128009,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 32.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": true,
+  "transformers_version": "4.57.1",
+  "use_cache": false,
+  "vocab_size": 128256
+}

checkpoints/knowledge/opsd_forward_kl_gt_d5_e10_llama3_2_1b_20260502/run_20260502.203716/step_120/policy/weights/model/.hf_metadata/fqn_to_file_index_mapping.json

@@ -0,0 +1,148 @@
+{
+  "model.embed_tokens.weight": 1,
+  "model.layers.0.input_layernorm.weight": 1,
+  "model.layers.0.mlp.down_proj.weight": 1,
+  "model.layers.0.mlp.gate_proj.weight": 1,
+  "model.layers.0.mlp.up_proj.weight": 1,
+  "model.layers.0.post_attention_layernorm.weight": 1,
+  "model.layers.0.self_attn.k_proj.weight": 1,
+  "model.layers.0.self_attn.o_proj.weight": 1,
+  "model.layers.0.self_attn.q_proj.weight": 1,
+  "model.layers.0.self_attn.v_proj.weight": 1,
+  "model.layers.1.input_layernorm.weight": 1,
+  "model.layers.1.mlp.down_proj.weight": 1,
+  "model.layers.1.mlp.gate_proj.weight": 1,
+  "model.layers.1.mlp.up_proj.weight": 1,
+  "model.layers.1.post_attention_layernorm.weight": 1,
+  "model.layers.1.self_attn.k_proj.weight": 1,
+  "model.layers.1.self_attn.o_proj.weight": 1,
+  "model.layers.1.self_attn.q_proj.weight": 1,
+  "model.layers.1.self_attn.v_proj.weight": 1,
+  "model.layers.10.input_layernorm.weight": 1,
+  "model.layers.10.mlp.down_proj.weight": 1,
+  "model.layers.10.mlp.gate_proj.weight": 1,
+  "model.layers.10.mlp.up_proj.weight": 1,
+  "model.layers.10.post_attention_layernorm.weight": 1,
+  "model.layers.10.self_attn.k_proj.weight": 1,
+  "model.layers.10.self_attn.o_proj.weight": 1,
+  "model.layers.10.self_attn.q_proj.weight": 1,
+  "model.layers.10.self_attn.v_proj.weight": 1,
+  "model.layers.11.input_layernorm.weight": 1,
+  "model.layers.11.mlp.down_proj.weight": 1,
+  "model.layers.11.mlp.gate_proj.weight": 1,
+  "model.layers.11.mlp.up_proj.weight": 1,
+  "model.layers.11.post_attention_layernorm.weight": 1,
+  "model.layers.11.self_attn.k_proj.weight": 1,
+  "model.layers.11.self_attn.o_proj.weight": 1,
+  "model.layers.11.self_attn.q_proj.weight": 1,
+  "model.layers.11.self_attn.v_proj.weight": 1,
+  "model.layers.12.input_layernorm.weight": 1,
+  "model.layers.12.mlp.down_proj.weight": 1,
+  "model.layers.12.mlp.gate_proj.weight": 1,
+  "model.layers.12.mlp.up_proj.weight": 1,
+  "model.layers.12.post_attention_layernorm.weight": 1,
+  "model.layers.12.self_attn.k_proj.weight": 1,
+  "model.layers.12.self_attn.o_proj.weight": 1,
+  "model.layers.12.self_attn.q_proj.weight": 1,
+  "model.layers.12.self_attn.v_proj.weight": 1,
+  "model.layers.13.input_layernorm.weight": 1,
+  "model.layers.13.mlp.down_proj.weight": 1,
+  "model.layers.13.mlp.gate_proj.weight": 1,
+  "model.layers.13.mlp.up_proj.weight": 1,
+  "model.layers.13.post_attention_layernorm.weight": 1,
+  "model.layers.13.self_attn.k_proj.weight": 1,
+  "model.layers.13.self_attn.o_proj.weight": 1,
+  "model.layers.13.self_attn.q_proj.weight": 1,
+  "model.layers.13.self_attn.v_proj.weight": 1,
+  "model.layers.14.input_layernorm.weight": 1,
+  "model.layers.14.mlp.down_proj.weight": 1,
+  "model.layers.14.mlp.gate_proj.weight": 1,
+  "model.layers.14.mlp.up_proj.weight": 1,
+  "model.layers.14.post_attention_layernorm.weight": 1,
+  "model.layers.14.self_attn.k_proj.weight": 1,
+  "model.layers.14.self_attn.o_proj.weight": 1,
+  "model.layers.14.self_attn.q_proj.weight": 1,
+  "model.layers.14.self_attn.v_proj.weight": 1,
+  "model.layers.15.input_layernorm.weight": 1,
+  "model.layers.15.mlp.down_proj.weight": 1,
+  "model.layers.15.mlp.gate_proj.weight": 1,
+  "model.layers.15.mlp.up_proj.weight": 1,
+  "model.layers.15.post_attention_layernorm.weight": 1,
+  "model.layers.15.self_attn.k_proj.weight": 1,
+  "model.layers.15.self_attn.o_proj.weight": 1,
+  "model.layers.15.self_attn.q_proj.weight": 1,
+  "model.layers.15.self_attn.v_proj.weight": 1,
+  "model.layers.2.input_layernorm.weight": 1,
+  "model.layers.2.mlp.down_proj.weight": 1,
+  "model.layers.2.mlp.gate_proj.weight": 1,
+  "model.layers.2.mlp.up_proj.weight": 1,
+  "model.layers.2.post_attention_layernorm.weight": 1,
+  "model.layers.2.self_attn.k_proj.weight": 1,
+  "model.layers.2.self_attn.o_proj.weight": 1,
+  "model.layers.2.self_attn.q_proj.weight": 1,
+  "model.layers.2.self_attn.v_proj.weight": 1,
+  "model.layers.3.input_layernorm.weight": 1,
+  "model.layers.3.mlp.down_proj.weight": 1,
+  "model.layers.3.mlp.gate_proj.weight": 1,
+  "model.layers.3.mlp.up_proj.weight": 1,
+  "model.layers.3.post_attention_layernorm.weight": 1,
+  "model.layers.3.self_attn.k_proj.weight": 1,
+  "model.layers.3.self_attn.o_proj.weight": 1,
+  "model.layers.3.self_attn.q_proj.weight": 1,
+  "model.layers.3.self_attn.v_proj.weight": 1,
+  "model.layers.4.input_layernorm.weight": 1,
+  "model.layers.4.mlp.down_proj.weight": 1,
+  "model.layers.4.mlp.gate_proj.weight": 1,
+  "model.layers.4.mlp.up_proj.weight": 1,
+  "model.layers.4.post_attention_layernorm.weight": 1,
+  "model.layers.4.self_attn.k_proj.weight": 1,
+  "model.layers.4.self_attn.o_proj.weight": 1,
+  "model.layers.4.self_attn.q_proj.weight": 1,
+  "model.layers.4.self_attn.v_proj.weight": 1,
+  "model.layers.5.input_layernorm.weight": 1,
+  "model.layers.5.mlp.down_proj.weight": 1,
+  "model.layers.5.mlp.gate_proj.weight": 1,
+  "model.layers.5.mlp.up_proj.weight": 1,
+  "model.layers.5.post_attention_layernorm.weight": 1,
+  "model.layers.5.self_attn.k_proj.weight": 1,
+  "model.layers.5.self_attn.o_proj.weight": 1,
+  "model.layers.5.self_attn.q_proj.weight": 1,
+  "model.layers.5.self_attn.v_proj.weight": 1,
+  "model.layers.6.input_layernorm.weight": 1,
+  "model.layers.6.mlp.down_proj.weight": 1,
+  "model.layers.6.mlp.gate_proj.weight": 1,
+  "model.layers.6.mlp.up_proj.weight": 1,
+  "model.layers.6.post_attention_layernorm.weight": 1,
+  "model.layers.6.self_attn.k_proj.weight": 1,
+  "model.layers.6.self_attn.o_proj.weight": 1,
+  "model.layers.6.self_attn.q_proj.weight": 1,
+  "model.layers.6.self_attn.v_proj.weight": 1,
+  "model.layers.7.input_layernorm.weight": 1,
+  "model.layers.7.mlp.down_proj.weight": 1,
+  "model.layers.7.mlp.gate_proj.weight": 1,
+  "model.layers.7.mlp.up_proj.weight": 1,
+  "model.layers.7.post_attention_layernorm.weight": 1,
+  "model.layers.7.self_attn.k_proj.weight": 1,
+  "model.layers.7.self_attn.o_proj.weight": 1,
+  "model.layers.7.self_attn.q_proj.weight": 1,
+  "model.layers.7.self_attn.v_proj.weight": 1,
+  "model.layers.8.input_layernorm.weight": 1,
+  "model.layers.8.mlp.down_proj.weight": 1,
+  "model.layers.8.mlp.gate_proj.weight": 1,
+  "model.layers.8.mlp.up_proj.weight": 1,
+  "model.layers.8.post_attention_layernorm.weight": 1,
+  "model.layers.8.self_attn.k_proj.weight": 1,
+  "model.layers.8.self_attn.o_proj.weight": 1,
+  "model.layers.8.self_attn.q_proj.weight": 1,
+  "model.layers.8.self_attn.v_proj.weight": 1,
+  "model.layers.9.input_layernorm.weight": 1,
+  "model.layers.9.mlp.down_proj.weight": 1,
+  "model.layers.9.mlp.gate_proj.weight": 1,
+  "model.layers.9.mlp.up_proj.weight": 1,
+  "model.layers.9.post_attention_layernorm.weight": 1,
+  "model.layers.9.self_attn.k_proj.weight": 1,
+  "model.layers.9.self_attn.o_proj.weight": 1,
+  "model.layers.9.self_attn.q_proj.weight": 1,
+  "model.layers.9.self_attn.v_proj.weight": 1,
+  "model.norm.weight": 1
+}