Duplicate from robot-learning/v2w_ex1_ex2_all_lora_rank32_lr5.623e-05_bsz32_630062_2

Co-authored-by: Emerson Leonardo Azevedo Aguiar <eazevedo@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

checkpoints/model/iter_000000590.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:285b01d818b49ad793981af80624610d901a262cc0a76777cc8717b8ff54b28e
+size 4005911812

checkpoints/model/iter_000000590_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c3aca1735e21737f43a9685a4f47199dd6ba810dc111d95dc591ebd253e87b0
+size 3913047915

checkpoints/model/iter_000000595.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67103bc6eafe62ed2f380fddb2672f01798d44896829f0af0fe747d0f1610cf1
+size 4005911812

checkpoints/model/iter_000000595_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b0a04820dd39f018227a85652171d5b69377d9bd779dd2bfad1cbf363d09ca0
+size 3913047915

checkpoints/model/iter_000000600.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f7173c0eb6a232679fa14347b562c2c06678880698cacb9682e9eecb0c861dd
+size 4005911812

checkpoints/model/iter_000000600_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:178a887b416060f10459b44e52fae4441c2b1684866a145bb457987005436e8e
+size 3913047915

checkpoints/model/iter_000000605.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07d37d4d01cfd415a4bf66d98a00807446614494c07c97bb6a70d200ba9d4d17
+size 4005911812

checkpoints/model/iter_000000605_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6b93f3ee921e74156bc6940e5392f4243a1eebcca1cf472fe82b67b8d8b5fb8
+size 3913047915

checkpoints/model/iter_000000610.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da7c55c4ed0fd8f9d035af6f65a2a216952db27bef1da99bc1cbadf6ac0ecd04
+size 4005911812

checkpoints/model/iter_000000610_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83d5a53bd59d6c7fdb52c96f47bc40f7796fe8e248e915c09d71fdfda075c10a
+size 3913047915

checkpoints/optim/iter_000000590.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:017657786dc386c94d3a2bcc0b0b08b2b21ea0c9ce748db9ef71846c6efdf1a5
+size 1792982648

checkpoints/optim/iter_000000590_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5d17ff684fe2933d27ae654c75b0566ce974347c8fc0dfd99a3b426c83637f9
+size 1792991522

checkpoints/optim/iter_000000595.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77dbe181547d4dae4bd8a6ad9ad501225abd60680682ddfbc2b69bf962d27a13
+size 1792982648

checkpoints/optim/iter_000000595_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f94395e497b35a0284042299f9abaa84e5ca4808206cc61598295a58ef2969b
+size 1792991522

checkpoints/optim/iter_000000600.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d23ba5c7ef70de07244145c4200b73500ec9c95a3199d8d0276a7dafefa0642
+size 1792982648

checkpoints/optim/iter_000000600_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e9a885662f035ea7491d3f7c54968b23bb7f8eeead7e3cf34b7af4318a1ec31
+size 1792991522

checkpoints/optim/iter_000000605.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e499b7a202d4d6c7ce801ee20963df5e8a3f35da8502095467d1e2c4c771af88
+size 1792982648

checkpoints/optim/iter_000000605_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a41b1fd94d3572c8909d227c82eeaaff8e45c71487927e5d94121ea5d2147fc4
+size 1792991522

checkpoints/optim/iter_000000610.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f4e108d10e7e21aefd25c78273ebe8c12139f60826495ace929d9b3c5bb4d7a
+size 1792982648

checkpoints/optim/iter_000000610_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5da69730060a4cc685acdb8722be981aa93b0aa477fe170ab6c0f1777eec50ed
+size 1792991522

checkpoints/scheduler/iter_000000590.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:843d52709b0a7b2ce3e7c6fc6b6b9dd644d74a7226b530a723e37a2c3356e77a
+size 1602

checkpoints/scheduler/iter_000000590_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6e2f958451789d9d0acdfea937b0569cbce74d367bd49e0c7f6050ea95cdd30
+size 1638

checkpoints/scheduler/iter_000000595.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f1d703258fe6231f1dede549faa059a10546c8dd38de316e408971b690f4fd5d
+size 1602

checkpoints/scheduler/iter_000000595_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b8350861157f431535565e2daca2c18b0cf53835b1c936d1a4d0b5dbb6cd14ee
+size 1638

checkpoints/scheduler/iter_000000600.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d63df3fb8fee4688d10c1a3ebd7a65a9326f72bdf2958cd2413b1cac4007e839
+size 1602

checkpoints/scheduler/iter_000000600_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5cace1acf96a130c825274b1741a7b3e0d03ef331a133a30a5a88320585e7506
+size 1638

checkpoints/scheduler/iter_000000605.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3d01fa815d8198d98065158b901e9add4fed1194b12132384dfbdb6fb55557c
+size 1602

checkpoints/scheduler/iter_000000605_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59969cb88626c2a2c48e74a5555241b989e07b8a4f91f00de09c69a2e0acf036
+size 1638

checkpoints/scheduler/iter_000000610.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ca11ac803b0f8d8cf61848ad6ac950fb066a2376bed5647a966baef1c9866d2
+size 1602

checkpoints/scheduler/iter_000000610_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c9f90f51d77358e1286ad373b50a4c17acb964364aef917010a65f73d3dfce7
+size 1638

checkpoints/trainer/iter_000000590.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1eea0474f5fa546d584fd2317b7be832c4f9850d66ae5895c519fb9098f909e3
+size 892

checkpoints/trainer/iter_000000590_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00d148eb2716c82109bcc4d1fca29665da42bfb46ff7b7282462f6a60d2aafbd
+size 916

checkpoints/trainer/iter_000000595.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9431c415f6c7f62f69c166d72229b7ae41ef92ecb95ceacb39c913fa845c08ae
+size 892

checkpoints/trainer/iter_000000595_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92857fc8440d3c706d9327364f0a40eacaec745a5dd02fa7d7b4b2abb56f4e18
+size 916

checkpoints/trainer/iter_000000600.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d4241c8657960fc56d56e303b2f10176953d10a3c4c475826e14070d4f889bd
+size 892

checkpoints/trainer/iter_000000600_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21f10a27589d2d11e6e540100fff437f4d2eb8f47d51f65bda87706bc922ea66
+size 916

checkpoints/trainer/iter_000000605.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:842245e8cd923d3d58df3354a72d97222cde22a95ce40d471319d9d8afd8928d
+size 892

checkpoints/trainer/iter_000000605_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:41ebf6a2dc265de8465099c3bdc25349dfd2cf67493d44f83998111653495491
+size 916

checkpoints/trainer/iter_000000610.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c9d4b5845a309da593453ce13b71741af8e30885038ef179ca83c990d96e1c4
+size 892

checkpoints/trainer/iter_000000610_fused.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8e0ab481cb56269bb0fe7fe9e5ffe8558715dffce593003e03ad5d802025cb8
+size 916

config.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b8d893e2dd035cd2978c76de99bb1377dfac113d97b98931c712a1a7930d67d
+size 75030

config.yaml

@@ -0,0 +1,530 @@
+checkpoint:
+  broadcast_via_filesystem: 'False'
+  dcp_allow_mismatched_size: 'False'
+  dcp_async_mode_enabled: 'False'
+  jit:
+    device: cuda
+    dtype: bfloat16
+    enabled: 'False'
+    input_shape: null
+    strict: 'True'
+  keys_not_to_resume: []
+  load_ema_to_reg: 'False'
+  load_path: ''
+  load_training_state: 'False'
+  only_load_scheduler_state: 'False'
+  save_iter: '5'
+  strict_resume: 'True'
+  type:
+    _target_: <class 'cosmos_predict2.checkpointer.Checkpointer'>
+    callbacks: null
+  verbose: 'True'
+data_config: null
+dataloader_train:
+  _target_: <class 'torch.utils.data.dataloader.DataLoader'>
+  batch_sampler: null
+  batch_size: '2'
+  collate_fn: null
+  dataset:
+    _target_: <class 'cosmos_predict2.data.dataset_video.MultiDataset'>
+    ex1_all:
+      _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+      data_fps: '10.0'
+      dataset_dir: /home/nvidia/robot_learning_project/data/ex1_all-cosmos-video
+      exclude_with_substring: null
+      include_only_with_substrings: null
+      is_multi_img: 'False'
+      is_val: 'False'
+      num_frames: '61'
+      obs_history: '5'
+      val_ratio: '0.0'
+      video_size:
+      - '480'
+      - '640'
+    ex2_all:
+      _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+      data_fps: '10.0'
+      dataset_dir: /home/nvidia/robot_learning_project/data/ex2_all-cosmos-video
+      exclude_with_substring: null
+      include_only_with_substrings: null
+      is_multi_img: 'False'
+      is_val: 'False'
+      num_frames: '61'
+      obs_history: '5'
+      val_ratio: '0.0'
+      video_size:
+      - '480'
+      - '640'
+  drop_last: 'True'
+  generator: null
+  in_order: 'False'
+  multiprocessing_context: null
+  num_workers: '12'
+  persistent_workers: 'True'
+  pin_memory: 'True'
+  pin_memory_device: ''
+  prefetch_factor: '8'
+  sampler:
+    _target_: <function get_sampler at 0x7333e173a440>
+    dataset:
+      _target_: <class 'cosmos_predict2.data.dataset_video.MultiDataset'>
+      ex1_all:
+        _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+        data_fps: '10.0'
+        dataset_dir: /home/nvidia/robot_learning_project/data/ex1_all-cosmos-video
+        exclude_with_substring: null
+        include_only_with_substrings: null
+        is_multi_img: 'False'
+        is_val: 'False'
+        num_frames: '61'
+        obs_history: '5'
+        val_ratio: '0.0'
+        video_size:
+        - '480'
+        - '640'
+      ex2_all:
+        _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+        data_fps: '10.0'
+        dataset_dir: /home/nvidia/robot_learning_project/data/ex2_all-cosmos-video
+        exclude_with_substring: null
+        include_only_with_substrings: null
+        is_multi_img: 'False'
+        is_val: 'False'
+        num_frames: '61'
+        obs_history: '5'
+        val_ratio: '0.0'
+        video_size:
+        - '480'
+        - '640'
+  shuffle: null
+  timeout: '0'
+  worker_init_fn: null
+dataloader_val:
+  _target_: <class 'torch.utils.data.dataloader.DataLoader'>
+  batch_sampler: null
+  batch_size: '1'
+  collate_fn: null
+  dataset:
+    _target_: <class 'cosmos_predict2.data.dataset_video.MultiDataset'>
+    ex1_all:
+      _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+      data_fps: '10.0'
+      dataset_dir: /home/nvidia/robot_learning_project/data/ex1_all-cosmos-video
+      exclude_with_substring: null
+      include_only_with_substrings: null
+      is_multi_img: 'False'
+      is_val: 'True'
+      num_frames: '61'
+      obs_history: '5'
+      val_ratio: '0.0'
+      video_size:
+      - '480'
+      - '640'
+    ex2_all:
+      _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+      data_fps: '10.0'
+      dataset_dir: /home/nvidia/robot_learning_project/data/ex2_all-cosmos-video
+      exclude_with_substring: null
+      include_only_with_substrings: null
+      is_multi_img: 'False'
+      is_val: 'True'
+      num_frames: '61'
+      obs_history: '5'
+      val_ratio: '0.0'
+      video_size:
+      - '480'
+      - '640'
+  drop_last: 'False'
+  generator: null
+  in_order: 'False'
+  multiprocessing_context: null
+  num_workers: '0'
+  persistent_workers: 'False'
+  pin_memory: 'False'
+  pin_memory_device: ''
+  prefetch_factor: null
+  sampler:
+    _target_: <function get_sampler at 0x7333e173a440>
+    dataset:
+      _target_: <class 'cosmos_predict2.data.dataset_video.MultiDataset'>
+      ex1_all:
+        _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+        data_fps: '10.0'
+        dataset_dir: /home/nvidia/robot_learning_project/data/ex1_all-cosmos-video
+        exclude_with_substring: null
+        include_only_with_substrings: null
+        is_multi_img: 'False'
+        is_val: 'True'
+        num_frames: '61'
+        obs_history: '5'
+        val_ratio: '0.0'
+        video_size:
+        - '480'
+        - '640'
+      ex2_all:
+        _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+        data_fps: '10.0'
+        dataset_dir: /home/nvidia/robot_learning_project/data/ex2_all-cosmos-video
+        exclude_with_substring: null
+        include_only_with_substrings: null
+        is_multi_img: 'False'
+        is_val: 'True'
+        num_frames: '61'
+        obs_history: '5'
+        val_ratio: '0.0'
+        video_size:
+        - '480'
+        - '640'
+  shuffle: null
+  timeout: '0'
+  worker_init_fn: null
+defaults:
+- _self_
+- data_config: null
+- video_dataset_train: null
+- video_dataset_val: null
+- dataloader_train: null
+- dataloader_val: null
+- world2action_pipe: null
+- optimizer: fusedadamw
+- scheduler: constant
+- model: null
+- callbacks:
+  - basic
+- net: null
+- ema: null
+- checkpoint: null
+- ckpt_type: null
+- experiment: null
+job:
+  group: video2world
+  name: v2w_ex1_ex2_all_lora_rank32_lr5.623e-05_bsz32_630062
+  project: posttraining
+model:
+  _recursive_: 'False'
+  _target_: <class 'cosmos_predict2.models.video2world_model.Predict2Video2WorldModel'>
+  config:
+    adjust_video_noise: true
+    debug_without_randomness: false
+    fsdp_shard_size: 0
+    high_sigma_ratio: 0.05
+    init_lora_weights: true
+    input_image_key: images
+    input_video_key: video
+    lora_alpha: 32
+    lora_rank: 32
+    lora_target_modules: q_proj,k_proj,v_proj,output_proj,x_embedder.proj.1,linear_1,linear_2,mlp.layer1,mlp.layer2
+    loss_reduce: mean
+    loss_scale: 100.0
+    model_manager_config:
+      _target_: cosmos_predict2.models.video2world_model.Predict2ModelManagerConfig
+      dit_path: /home/nvidia/robot_learning_project/mimic-video/model/checkpoints/video_backbone//v2w_pretrained_cosmos.pt
+      text_encoder_path: ''
+    pipe_config:
+      adjust_video_noise: true
+      conditioner:
+        _target_: <class 'cosmos_predict2.conditioner.VideoConditioner'>
+        fps:
+          _target_: <class 'cosmos_predict2.conditioner.ReMapkey'>
+          dropout_rate: '0.0'
+          dtype: null
+          input_key: fps
+          output_key: fps
+        padding_mask:
+          _target_: <class 'cosmos_predict2.conditioner.ReMapkey'>
+          dropout_rate: '0.0'
+          dtype: null
+          input_key: padding_mask
+          output_key: padding_mask
+        text:
+          _target_: <class 'cosmos_predict2.conditioner.TextAttr'>
+          dropout_rate: '0.0'
+          input_key:
+          - obs/language_embedding
+        use_video_condition:
+          _target_: <class 'cosmos_predict2.conditioner.BooleanFlag'>
+          dropout_rate: '0.0'
+          input_key: fps
+          output_key: use_video_condition
+      conditioning_strategy: frame_replace
+      ema:
+        _target_: cosmos_predict2.configs.defaults.ema.EMAConfig
+        enabled: 'False'
+        iteration_shift: '0'
+        rate: '0.1'
+      guardrail_config:
+        checkpoint_dir: /home/nvidia/robot_learning_project/mimic-video/model/checkpoints
+        enabled: false
+        offload_model_to_cpu: true
+      input_image_key: images
+      input_video_key: video
+      max_num_conditional_frames: 2
+      min_num_conditional_frames: 1
+      net:
+        _target_: <class 'cosmos_predict2.models.video2world_dit.MinimalV1LVGDiT'>
+        adaln_lora_dim: '256'
+        atten_backend: minimal_a2a
+        concat_padding_mask: 'True'
+        extra_per_block_abs_pos_emb: 'False'
+        in_channels: '16'
+        max_frames: '128'
+        max_img_h: '240'
+        max_img_w: '240'
+        model_channels: '2048'
+        num_blocks: '28'
+        num_heads: '16'
+        out_channels: '16'
+        patch_spatial: '2'
+        patch_temporal: '1'
+        pos_emb_cls: rope3d
+        pos_emb_interpolation: crop
+        pos_emb_learnable: 'True'
+        rope_enable_fps_modulation: 'False'
+        rope_h_extrapolation_ratio: '3.0'
+        rope_t_extrapolation_ratio: '1.0'
+        rope_w_extrapolation_ratio: '3.0'
+        sac_config:
+          _target_: cosmos_predict2.models.text2image_dit.SACConfig
+          every_n_blocks: '1'
+          mode: predict2_2b_720
+        use_adaln_lora: 'True'
+      precision: bfloat16
+      rectified_flow_loss_weight_uniform: true
+      rectified_flow_t_scaling_factor: 1.0
+      resize_online: false
+      resolution: '480'
+      sigma_conditional: 0.0001
+      sigma_data: 1.0
+      state_ch: 16
+      state_t: 16
+      text_encoder:
+        cls: !!python/object/apply:imaginaire.constants.TextEncoderClass
+        - t5
+        t5:
+          ckpt_path: /home/nvidia/robot_learning_project/mimic-video/model/checkpoints/text_encoder/t5-11b
+          embed_dim: 1024
+          num_tokens: 512
+      timestamps:
+        is_forward: false
+        nfe: 35
+        order: 7.0
+        t_max: 80.0
+        t_min: 0.002
+      tokenizer:
+        _target_: <class 'cosmos_predict2.tokenizers.tokenizer.TokenizerInterface'>
+        chunk_duration: '81'
+        load_mean_std: 'False'
+        name: tokenizer
+        temporal_window: '16'
+        vae_pth: /home/nvidia/robot_learning_project/mimic-video/model/checkpoints/video_backbone//tokenizer/tokenizer.pth
+    precision: bfloat16
+    train_architecture: lora
+model_parallel:
+  _cpu_offloading_context: null
+  async_tensor_model_parallel_allreduce: false
+  autocast_dtype: torch.float32
+  barrier_with_L1_time: true
+  batch_p2p_comm: true
+  batch_p2p_sync: true
+  bf16: false
+  context_parallel_size: 1
+  cpu_offloading: false
+  cpu_offloading_activations: false
+  cpu_offloading_num_layers: 0
+  cpu_offloading_weights: false
+  cross_entropy_fusion_impl: native
+  cross_entropy_loss_fusion: false
+  deallocate_pipeline_outputs: false
+  defer_embedding_wgrad_compute: false
+  deterministic_mode: false
+  enable_autocast: false
+  expert_model_parallel_size: 1
+  expert_tensor_parallel_size: 1
+  finalize_model_grads_func: null
+  fp16: false
+  grad_scale_func: null
+  grad_sync_func: null
+  gradient_accumulation_fusion: false
+  hierarchical_context_parallel_sizes: null
+  microbatch_group_size_per_vp_stage: 1
+  moe_extended_tp: false
+  no_sync_func: null
+  num_microbatches_with_partial_activation_checkpoints: null
+  overlap_p2p_comm: false
+  overlap_p2p_comm_warmup_flush: false
+  param_sync_func: null
+  params_dtype: torch.float32
+  perform_initialization: true
+  pipeline_dtype: null
+  pipeline_model_parallel_comm_backend: null
+  pipeline_model_parallel_size: 1
+  pipeline_model_parallel_split_rank: null
+  sequence_parallel: false
+  tensor_model_parallel_size: 1
+  timers: null
+  tp_comm_atomic_ag: false
+  tp_comm_atomic_rs: false
+  tp_comm_bootstrap_backend: nccl
+  tp_comm_bulk_dgrad: true
+  tp_comm_bulk_wgrad: true
+  tp_comm_overlap: false
+  tp_comm_overlap_ag: true
+  tp_comm_overlap_disable_fc1: false
+  tp_comm_overlap_disable_qkv: false
+  tp_comm_overlap_rs: true
+  tp_comm_overlap_rs_dgrad: false
+  tp_comm_split_ag: true
+  tp_comm_split_rs: true
+  use_cpu_initialization: false
+  use_ring_exchange_p2p: false
+  use_te_rng_tracker: false
+  variable_seq_lengths: false
+  virtual_pipeline_model_parallel_size: null
+  wgrad_deferral_limit: 0
+optimizer:
+  _target_: <function get_base_optimizer at 0x7333d19b72e0>
+  betas:
+  - '0.9'
+  - '0.99'
+  capturable: 'True'
+  eps: 1e-08
+  lr: '5.623e-05'
+  master_weights: 'True'
+  model: null
+  optim_type: fusedadam
+  weight_decay: '0.1'
+scheduler:
+  _target_: <class 'cosmos_predict2.configs.defaults.scheduler.ConstantScheduler'>
+trainer:
+  callbacks:
+    device_monitor:
+      _target_: <class 'cosmos_predict2.callbacks.device_monitor.DeviceMonitor'>
+      every_n: '1'
+      log_memory_detail: 'True'
+      step_size: '1'
+    ema:
+      _target_: <class 'imaginaire.utils.callback.EMAModelCallback'>
+      config: null
+      trainer: null
+    grad_clip:
+      _target_: <class 'cosmos_predict2.callbacks.grad_clip.GradClip'>
+      clip_norm: '10.0'
+      force_finite: 'True'
+      log_wandb: 'False'
+    iter_speed:
+      _target_: <class 'cosmos_predict2.callbacks.iter_speed.IterSpeed'>
+      every_n: '1'
+      hit_thres: '5'
+    low_prec:
+      _target_: <class 'imaginaire.utils.callback.LowPrecisionCallback'>
+      config: null
+      trainer: null
+      update_iter: '1'
+    manual_gc:
+      _target_: <class 'imaginaire.callbacks.manual_gc.ManualGarbageCollection'>
+      every_n: '5'
+      warm_up: '5'
+    progress_bar:
+      _target_: <class 'imaginaire.utils.callback.ProgressBarCallback'>
+      config: null
+      trainer: null
+    video_eval:
+      _target_: <class 'cosmos_predict2.callbacks.video_eval.VideoEvalCallback'>
+      fuse_lora: 'True'
+    wandb:
+      _target_: <class 'cosmos_predict2.callbacks.wandb_logger.WandBLogger'>
+      enabled: 'True'
+      entity: robot-learning-wm
+      log_every_n: '1'
+      mode: online
+      project: cosmos-video-finetune
+  cudnn:
+    benchmark: 'True'
+    deterministic: 'False'
+  ddp:
+    broadcast_buffers: 'True'
+    find_unused_parameters: 'False'
+    static_graph: 'True'
+  distributed_parallelism: ddp
+  grad_accum_iter: '8'
+  grad_scaler_args:
+    enabled: 'False'
+  logging_iter: '1'
+  max_iter: '1000000'
+  max_val_iter: null
+  memory_format: torch.preserve_format
+  profiling:
+    enable_memory_snapshot: 'False'
+    enable_profiling: 'False'
+    first_n_rank: '4'
+    profile_freq: '1'
+    profile_memory: 'True'
+    record_shape: 'True'
+    with_modules: 'True'
+    with_stack: 'True'
+  run_validation: 'False'
+  seed: '0'
+  timeout_period: '999999999'
+  type: <class 'imaginaire.trainer.ImaginaireTrainer'>
+  validation_iter: '999999999'
+video_dataset_train:
+  _target_: <class 'cosmos_predict2.data.dataset_video.MultiDataset'>
+  ex1_all:
+    _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+    data_fps: '10.0'
+    dataset_dir: /home/nvidia/robot_learning_project/data/ex1_all-cosmos-video
+    exclude_with_substring: null
+    include_only_with_substrings: null
+    is_multi_img: 'False'
+    is_val: 'False'
+    num_frames: '61'
+    obs_history: '5'
+    val_ratio: '0.0'
+    video_size:
+    - '480'
+    - '640'
+  ex2_all:
+    _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+    data_fps: '10.0'
+    dataset_dir: /home/nvidia/robot_learning_project/data/ex2_all-cosmos-video
+    exclude_with_substring: null
+    include_only_with_substrings: null
+    is_multi_img: 'False'
+    is_val: 'False'
+    num_frames: '61'
+    obs_history: '5'
+    val_ratio: '0.0'
+    video_size:
+    - '480'
+    - '640'
+video_dataset_val:
+  _target_: <class 'cosmos_predict2.data.dataset_video.MultiDataset'>
+  ex1_all:
+    _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+    data_fps: '10.0'
+    dataset_dir: /home/nvidia/robot_learning_project/data/ex1_all-cosmos-video
+    exclude_with_substring: null
+    include_only_with_substrings: null
+    is_multi_img: 'False'
+    is_val: 'True'
+    num_frames: '61'
+    obs_history: '5'
+    val_ratio: '0.0'
+    video_size:
+    - '480'
+    - '640'
+  ex2_all:
+    _target_: <class 'cosmos_predict2.data.dataset_video.Dataset'>
+    data_fps: '10.0'
+    dataset_dir: /home/nvidia/robot_learning_project/data/ex2_all-cosmos-video
+    exclude_with_substring: null
+    include_only_with_substrings: null
+    is_multi_img: 'False'
+    is_val: 'True'
+    num_frames: '61'
+    obs_history: '5'
+    val_ratio: '0.0'
+    video_size:
+    - '480'
+    - '640'
+world2action_pipe: null