Upload logs segment: .

logs/dreamer_cont/0408/073043/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_real_cup
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/0408/073043
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 146
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: meat_on_grill
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_real_cup_bs_16_bl_64_eps_300
+weight_decay: 0.0

logs/dreamer_cont/0408/073220/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_real_cup
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/0408/073220
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 146
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: meat_on_grill
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_real_cup_bs_16_bl_64_eps_300
+weight_decay: 0.0

logs/dreamer_cont/0408/081628/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_meat_on_grill
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/0408/081628
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: meat_on_grill
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_meat_on_grill_bs16_bl2
+weight_decay: 0.0

logs/dreamer_cont/0408/092217/config.yaml

@@ -0,0 +1,150 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_meat_on_grill
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/0408/092217
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: meat_on_grill
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_meat_on_grill_bs16_bl2
+weight_decay: 0.0
+_decay: 0.0

logs/dreamer_cont/0408/094241/best_pretrain_joint_0_00.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51455862225ea014d91ce20ad2c7cafb53a6f2ea8fee5aeedd6fc6cd04e1d90e
+size 213894142

logs/dreamer_cont/0408/094241/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_place_wine_at_rack_location
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/0408/094241
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: place_wine_at_rack_location
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_place_wine_at_rack_location_bs16_bl2
+weight_decay: 0.0

logs/dreamer_cont/0408/094241/pretrain_joint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52b9583ff40b025d2d2edfaac1b7445be1eaf36d7da3c804d3c3e9c0e15d37c2
+size 213890918

logs/dreamer_cont/meat_on_grill/0408/102201/best_pretrain_joint_0_00.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5753ae71ea1d32087341ffe70f2efdacda9d548b1b5b99085a823253bf61ca2
+size 213894142

logs/dreamer_cont/meat_on_grill/0408/102201/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_meat_on_grill
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/meat_on_grill/0408/102201
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: meat_on_grill
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_meat_on_grill_bs16_bl2
+weight_decay: 0.0

logs/dreamer_cont/meat_on_grill/0408/102201/pretrain_joint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7908ed3a2e36e101a5b677af5c0192f6899c802e2227da0fcfa89439c9341b18
+size 213890918

logs/dreamer_cont/move_hanger/0408/102201/best_pretrain_joint_0_00.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b3eb1ccd73c47deff97044a5c2b6a24d989daf2d02e52bd50388f318a0402eab
+size 213894142

logs/dreamer_cont/move_hanger/0408/102201/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_move_hanger
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/move_hanger/0408/102201
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: move_hanger
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_move_hanger_bs16_bl2
+weight_decay: 0.0

logs/dreamer_cont/move_hanger/0408/102201/pretrain_joint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc93054057bb426e799eccc177ca94277bdd3944260a8dadf4940443bb476e0e
+size 213890918

logs/dreamer_cont/open_drawer/0408/102201/best_pretrain_joint_0_00.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6500da333f8182974db47105609f15692837b5754304f29e6190400d2d014b36
+size 213894142

logs/dreamer_cont/open_drawer/0408/102201/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_open_drawer
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/open_drawer/0408/102201
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: open_drawer
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_open_drawer_bs16_bl2
+weight_decay: 0.0

logs/dreamer_cont/open_drawer/0408/102201/pretrain_joint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9c802783c1dfa5361233e9bfdd38171957df4c00880be0398c429fdb987ba46
+size 213890918

logs/dreamer_cont/place_wine_at_rack_location/0408/102201/best_pretrain_joint_0_00.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da816ff009ea8097eae2d0ab48217aa8c4d60ad7b87d8a5df21f82cdf72f886d
+size 213894142

logs/dreamer_cont/place_wine_at_rack_location/0408/102201/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_place_wine_at_rack_location
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/place_wine_at_rack_location/0408/102201
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: place_wine_at_rack_location
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_place_wine_at_rack_location_bs16_bl2
+weight_decay: 0.0

logs/dreamer_cont/place_wine_at_rack_location/0408/102201/pretrain_joint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd8592a1492ab4cc5dbc411b65544686ac91f4cc05ca83ec2aa536116a5b9e56
+size 213890918

logs/dreamer_cont/put_money_in_safe/0408/102201/best_pretrain_joint_0_00.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8fb9e7a9f7429d8c065ca8b154fa93a61a36763cf5acc9e3241f9f9754699518
+size 213894142

logs/dreamer_cont/put_money_in_safe/0408/102201/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_put_money_in_safe
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/put_money_in_safe/0408/102201
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: put_money_in_safe
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_put_money_in_safe_bs16_bl2
+weight_decay: 0.0

logs/dreamer_cont/put_money_in_safe/0408/102201/pretrain_joint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c31ccef63e713710bdaafe1f7c9c53db0a031f3b6bf14d688b8e219319ae88c0
+size 213890918

logs/dreamer_cont/scoop_with_spatula/0408/153050/best_pretrain_joint_0_00.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a7668a2170118465087190b3f754c41c20dc720fb1a1f1a478068bcd19d3941
+size 213894142

logs/dreamer_cont/scoop_with_spatula/0408/153050/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_scoop_with_spatula
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/scoop_with_spatula/0408/153050
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: scoop_with_spatula
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_scoop_with_spatula_bs16_bl2_0408
+weight_decay: 0.0

logs/dreamer_cont/scoop_with_spatula/0408/153050/pretrain_joint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:569d3b067ffd86a115c1f818bf7c31a25df2d49c0129846c71f7a05feee9160c
+size 213890918

logs/dreamer_cont/setup_chess/0408/153050/best_pretrain_joint_0_00.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc29ca1e03de9bba82135b429491cac12fae5c1697614302d032125307c877bb
+size 213894142

logs/dreamer_cont/setup_chess/0408/153050/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_setup_chess
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/setup_chess/0408/153050
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: setup_chess
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_setup_chess_bs16_bl2_0408
+weight_decay: 0.0

logs/dreamer_cont/setup_chess/0408/153050/pretrain_joint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b66f9979b35f678ac70245437664827bac8541f4657501acfaa2fb6f242e5f0
+size 213890918

logs/dreamer_cont/stack_cups/0408/153050/best_pretrain_joint_0_00.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6475057a3092e7f3d7822586fd0b17dbd44d37355ea520a66ce1709bbe76f9a2
+size 213894142

logs/dreamer_cont/stack_cups/0408/153050/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_stack_cups
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/stack_cups/0408/153050
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: stack_cups
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_stack_cups_bs16_bl2_0408
+weight_decay: 0.0

logs/dreamer_cont/stack_cups/0408/153050/pretrain_joint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ede064c1cfd86f31c822395b71a70a1a59cf191b24a5cf8d11f23e5975eae9c
+size 213890918

logs/dreamer_cont/turn_oven_on/0408/153050/best_pretrain_joint_0_00.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3a4487eba8dde310642d79502b4b23d14d3c472d7d98454a266631884209df5
+size 213894142

logs/dreamer_cont/turn_oven_on/0408/153050/config.yaml

@@ -0,0 +1,149 @@
+act: SiLU
+action_repeat: 1
+action_space: [8]
+action_type: abs
+actor: {dist: normal, entropy: 0.0003, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05,
+  max_std: 1.0, min_std: 0.1, outscale: 1.0, std: learned, temp: 0.1, unimix_ratio: 0.01}
+actor_only_scale: 0.0
+augment_images: false
+batch_length: 2
+batch_size: 16
+bc_loss_only: false
+bc_reg: true
+bc_reg_wd: 1.0
+bc_reg_weight: 1.0
+classifier_batch_length: 1500
+classifier_batch_size: 32
+classifier_epoch: 50
+classifier_latent_mode: all
+classifier_mode: split
+classifier_padding_mode: last
+compile: true
+cont_ensemble_size: 1
+cont_ensemble_subsample: 1
+cont_head: {layers: 2, loss_scale: 1.0, outscale: 1.0}
+critic: {dist: symlog_disc, eps: 1e-05, grad_clip: 100.0, layers: 2, lr: 3e-05, outscale: 0.0,
+  slow_target: true, slow_target_fraction: 0.02, slow_target_update: 1}
+critic_ensemble_size: 1
+critic_ensemble_subsample: 2
+critic_reset_every: 0
+dataset_size: 1000000
+debug: false
+decay_model_lr: false
+decoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, cnn_sigmoid: false, image_dist: mse,
+  kernel_size: 4, minres: 4, mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true,
+  outscale: 1.0, vector_dist: symlog_mse}
+deterministic_run: true
+device: cuda:0
+discount: 0.997
+discount_lambda: 0.95
+done_mode: 0
+dropout_recur_in_residuals: false
+dropout_recurrent_prob: 0
+dyn_deter: 512
+dyn_discrete: 0
+dyn_hidden: 512
+dyn_mean_act: none
+dyn_min_std: 0.1
+dyn_rec_depth: 1
+dyn_scale: 0.5
+dyn_std_act: sigmoid2
+dyn_stoch: 32
+ema_decay: 0.999
+ema_power: 0.8
+encoder: {act: SiLU, cnn_depth: 32, cnn_keys: .*image.*, kernel_size: 4, minres: 4,
+  mlp_keys: state, mlp_layers: 5, mlp_units: 1024, norm: true, symlog_inputs: true}
+ensemble_residuals: false
+envs: 1
+eval_every: 500
+eval_num_seeds: 50
+eval_per_seed: 1
+evaldir: null
+exp_name: wm_turn_oven_on
+freeze_encoder: true
+from_ckpt: null
+from_prefill_dir: null
+grad_clip: 1000
+grad_heads: [decoder, cont]
+grayscale: false
+hybrid_critic_fitting: false
+hybrid_training: true
+ignore_base_policy: false
+image_size: 64
+initial: learned
+initial_joint_train_steps: 0
+kl_free: 1.0
+log_every: 100
+log_every_video: 2000
+logdir: logs/dreamer_cont/turn_oven_on/0408/153050
+mask_recur: false
+model_lr: 0.0001
+model_only_scale: 0.0
+multi_task_data: false
+no_joint_steps: false
+norm: true
+num_exp_trajs: 300
+obs_keys: [cam_front_view_image, cam_wrist_view_image]
+obs_lr: 0.001
+obs_step: 1
+observation_space:
+  cam_front_view_image: [64, 64, 3]
+  cam_wrist_view_image: [64, 64, 3]
+  discount: [1]
+  state: [8]
+offline_evaldir: ''
+offline_traindir: ''
+opt: adam
+opt_eps: 1e-08
+parallel: false
+precision: 32
+prefill: 5000
+pretrain: 100
+pretrain_actor_steps: 0
+pretrain_annealing: null
+pretrain_bc_loss_only: true
+pretrain_ema: false
+pretrain_joint_steps: 150000
+pretrain_loss: ce
+pretrain_on_random: false
+pretrain_on_random_mixed: false
+pretrain_reward_val: null
+pretrain_separately: false
+recon_pretrain: true
+rep_scale: 0.1
+reset_every: 0
+residual_discount: 0.95
+residual_init_zeros: false
+reward_EMA: true
+reward_ensemble_size: 1
+reward_ensemble_subsample: 1
+reward_head: {dist: symlog_disc, layers: 2, loss_scale: 1.0, outscale: 0.0}
+root_dir: /data/colosseum_dataset/wm_data
+sample_freq: 2
+seed: 0
+seed_in_batches: true
+separate_reward_training: true
+shape_rewards: true
+shift_rewards: false
+size: [128, 128]
+state_keys: [state]
+steps: 1000000.0
+steps_per_batch: 100
+task: turn_oven_on
+time_limit: 100
+train_ratio: 1024
+train_residuals: true
+traindir: null
+transformer: {activation: gelu, attn_dropout: 0.05, block_output_dropout: 0.05, causal: true,
+  context_length: 1500, emb_dropout: 0.05, embed_dim: 64, enabled: true, nn_parameter_for_timesteps: false,
+  num_heads: 1, num_layers: 1, pred_future_acs: false, sinusoidal_embedding: true,
+  supervise_all_steps: false}
+unimix_ratio: 0.01
+units: 512
+use_additional_projection: false
+use_reward: false
+utd_ratio: 1
+validation_mse_trajs: 36
+video_pred_log: true
+wandb_name: wm_turn_oven_on_bs16_bl2_0408
+weight_decay: 0.0

logs/dreamer_cont/turn_oven_on/0408/153050/pretrain_joint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4322d5056d5bdb274e48d0eac9f97dc69acee94a55e74f7b7d3cfec6e932d08
+size 213890918

logs/master_vlm_runner.log

@@ -0,0 +1,19 @@
+nohup: ignoring input
+==========================================================
+🚀 [阶段一] 开始启动第一波 (老任务 5个)
+==========================================================
+🔥 启动任务: meat_on_grill -> GPU: 0 (时间戳: 102201)
+🔥 启动任务: move_hanger -> GPU: 1 (时间戳: 102201)
+🔥 启动任务: open_drawer -> GPU: 2 (时间戳: 102201)
+🔥 启动任务: put_money_in_safe -> GPU: 3 (时间戳: 102201)
+🔥 启动任务: place_wine_at_rack_location -> GPU: 4 (时间戳: 102201)
+⏳ 第一波任务正在后台运行。脚本将挂起等待它们全部完成...
+✅ 第一波任务全部圆满结束，显存已释放！
+==========================================================
+🚀 [阶段二] 开始启动第二波 (新任务 4个)
+==========================================================
+🔥 启动任务: scoop_with_spatula -> GPU: 0 (时间戳: 153050)
+🔥 启动任务: setup_chess -> GPU: 1 (时间戳: 153050)
+🔥 启动任务: stack_cups -> GPU: 2 (时间戳: 153050)
+🔥 启动任务: turn_oven_on -> GPU: 3 (时间戳: 153050)
+🎉 第二波任务已发射！您可以去休息了，所有工作都在自动推进。

logs/vlm_train_stack_cups_gpu2.log

@@ -0,0 +1,87 @@
+ROBOMIMIC WARNING(
+    No private macro file found!
+    It is recommended to use a private macro file
+    To setup, run: python /home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/robomimic/scripts/setup_macros.py
+)
+Warning: unknown parameter wm_ckpt
+Warning: unknown parameter target_modules
+wandb: Using wandb-core as the SDK backend. Please refer to https://wandb.me/wandb-core for more information.
+wandb: Currently logged in as: yukizhang0527 (yukizhang0527-harbin-institute-of-technology). Use `wandb login --relogin` to force relogin
+wandb: - Waiting for wandb.init()...
+wandb: Run data is saved locally in /home/zxn/Forewarn/wandb/run-20260409_101818-q6fxa9lv
+wandb: Run `wandb offline` to turn off syncing.
+wandb: Syncing run vlm_lora_stack_cups
+wandb: ⭐️ View project at https://wandb.ai/yukizhang0527-harbin-institute-of-technology/llama_recipes
+wandb: 🚀 View run at https://wandb.ai/yukizhang0527-harbin-institute-of-technology/llama_recipes/runs/q6fxa9lv
+Warning: custom_dataset does not accept parameter: custom_dataset.task_name
+
+--> Model /data/colosseum_dataset/wm_data/mllama_base/Llama-3.2-11B-Vision-Instruct/custom
+
+--> /data/colosseum_dataset/wm_data/mllama_base/Llama-3.2-11B-Vision-Instruct/custom has 9808.885776 Million params
+
+loading world model from ckpt path None
+Traceback (most recent call last):
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 540, in _check_seekable
+    f.seek(f.tell())
+AttributeError: 'NoneType' object has no attribute 'seek'
+
+During handling of the above exception, another exception occurred:
+
+Traceback (most recent call last):
+  File "/home/zxn/Forewarn/vlm/llama-recipes/recipes/quickstart/finetuning/finetuning_wm.py", line 9, in <module>
+    fire.Fire(main)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/fire/core.py", line 135, in Fire
+    component_trace = _Fire(component, args, parsed_flag_args, context, name)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/fire/core.py", line 468, in _Fire
+    component, remaining_args = _CallAndUpdateTrace(
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/fire/core.py", line 684, in _CallAndUpdateTrace
+    component = fn(*varargs, **kwargs)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/llama_recipes/finetuning_wm.py", line 355, in main
+    model.initialize_vision_model()
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/llama_recipes/models/mllama_model.py", line 140, in initialize_vision_model
+    self.wm_model = Dreamer.from_pretrained(path = wm_config.from_ckpt, obs_space = obs_space,
+  File "/home/zxn/Forewarn/model_based_irl_torch/dreamer/dreamer.py", line 62, in from_pretrained
+    ckpt = torch.load(path)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 997, in load
+    with _open_file_like(f, 'rb') as opened_file:
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 449, in _open_file_like
+    return _open_buffer_reader(name_or_buffer)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 434, in __init__
+    _check_seekable(buffer)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 543, in _check_seekable
+    raise_err_msg(["seek", "tell"], e)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 536, in raise_err_msg
+    raise type(e)(msg)
+AttributeError: 'NoneType' object has no attribute 'seek'. You can only torch.load from a file that is seekable. Please pre-load the data into a buffer like io.BytesIO and try to load from it instead.
+wandb: 🚀 View run vlm_lora_stack_cups at: https://wandb.ai/yukizhang0527-harbin-institute-of-technology/llama_recipes/runs/q6fxa9lv
+wandb: Find logs at: wandb/run-20260409_101818-q6fxa9lv/logs
+E0409 10:18:37.779480 140104379844416 torch/distributed/elastic/multiprocessing/api.py:826] failed (exitcode: 1) local_rank: 0 (pid: 672684) of binary: /home/zxn/anaconda3/envs/dreamer/bin/python3.10
+Traceback (most recent call last):
+  File "/home/zxn/anaconda3/envs/dreamer/bin/torchrun", line 6, in <module>
+    sys.exit(main())
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 347, in wrapper
+    return f(*args, **kwargs)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/run.py", line 879, in main
+    run(args)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/run.py", line 870, in run
+    elastic_launch(
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 132, in __call__
+    return launch_agent(self._config, self._entrypoint, list(args))
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 263, in launch_agent
+    raise ChildFailedError(
+torch.distributed.elastic.multiprocessing.errors.ChildFailedError: 
+============================================================
+/home/zxn/Forewarn/vlm/llama-recipes/recipes/quickstart/finetuning/finetuning_wm.py FAILED
+------------------------------------------------------------
+Failures:
+  <NO_OTHER_FAILURES>
+------------------------------------------------------------
+Root Cause (first observed failure):
+[0]:
+  time      : 2026-04-09_10:18:37
+  host      : node0029
+  rank      : 0 (local_rank: 0)
+  exitcode  : 1 (pid: 672684)
+  error_file: <N/A>
+  traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
+============================================================

logs/vlm_train_stack_cups_gpu2_final.log

@@ -0,0 +1,88 @@
+nohup: ignoring input
+ROBOMIMIC WARNING(
+    No private macro file found!
+    It is recommended to use a private macro file
+    To setup, run: python /home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/robomimic/scripts/setup_macros.py
+)
+Warning: unknown parameter wm_ckpt
+Warning: unknown parameter target_modules
+wandb: Using wandb-core as the SDK backend. Please refer to https://wandb.me/wandb-core for more information.
+wandb: Currently logged in as: yukizhang0527 (yukizhang0527-harbin-institute-of-technology). Use `wandb login --relogin` to force relogin
+wandb: - Waiting for wandb.init()...
+wandb: Run data is saved locally in /home/zxn/Forewarn/wandb/run-20260409_104618-abl48zp6
+wandb: Run `wandb offline` to turn off syncing.
+wandb: Syncing run eternal-sponge-63
+wandb: ⭐️ View project at https://wandb.ai/yukizhang0527-harbin-institute-of-technology/llama_recipes
+wandb: 🚀 View run at https://wandb.ai/yukizhang0527-harbin-institute-of-technology/llama_recipes/runs/abl48zp6
+Warning: custom_dataset does not accept parameter: custom_dataset.task_name
+
+--> Model /data/colosseum_dataset/wm_data/mllama_base/Llama-3.2-11B-Vision-Instruct/custom
+
+--> /data/colosseum_dataset/wm_data/mllama_base/Llama-3.2-11B-Vision-Instruct/custom has 9808.885776 Million params
+
+loading world model from ckpt path None
+Traceback (most recent call last):
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 540, in _check_seekable
+    f.seek(f.tell())
+AttributeError: 'NoneType' object has no attribute 'seek'
+
+During handling of the above exception, another exception occurred:
+
+Traceback (most recent call last):
+  File "/home/zxn/Forewarn/vlm/llama-recipes/recipes/quickstart/finetuning/finetuning_wm.py", line 9, in <module>
+    fire.Fire(main)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/fire/core.py", line 135, in Fire
+    component_trace = _Fire(component, args, parsed_flag_args, context, name)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/fire/core.py", line 468, in _Fire
+    component, remaining_args = _CallAndUpdateTrace(
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/fire/core.py", line 684, in _CallAndUpdateTrace
+    component = fn(*varargs, **kwargs)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/llama_recipes/finetuning_wm.py", line 355, in main
+    model.initialize_vision_model()
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/llama_recipes/models/mllama_model.py", line 140, in initialize_vision_model
+    self.wm_model = Dreamer.from_pretrained(path = wm_config.from_ckpt, obs_space = obs_space,
+  File "/home/zxn/Forewarn/model_based_irl_torch/dreamer/dreamer.py", line 62, in from_pretrained
+    ckpt = torch.load(path)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 997, in load
+    with _open_file_like(f, 'rb') as opened_file:
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 449, in _open_file_like
+    return _open_buffer_reader(name_or_buffer)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 434, in __init__
+    _check_seekable(buffer)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 543, in _check_seekable
+    raise_err_msg(["seek", "tell"], e)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 536, in raise_err_msg
+    raise type(e)(msg)
+AttributeError: 'NoneType' object has no attribute 'seek'. You can only torch.load from a file that is seekable. Please pre-load the data into a buffer like io.BytesIO and try to load from it instead.
+wandb: 🚀 View run eternal-sponge-63 at: https://wandb.ai/yukizhang0527-harbin-institute-of-technology/llama_recipes/runs/abl48zp6
+wandb: Find logs at: wandb/run-20260409_104618-abl48zp6/logs
+E0409 10:46:32.241449 139940508129088 torch/distributed/elastic/multiprocessing/api.py:826] failed (exitcode: 1) local_rank: 0 (pid: 688539) of binary: /home/zxn/anaconda3/envs/dreamer/bin/python3.10
+Traceback (most recent call last):
+  File "/home/zxn/anaconda3/envs/dreamer/bin/torchrun", line 6, in <module>
+    sys.exit(main())
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 347, in wrapper
+    return f(*args, **kwargs)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/run.py", line 879, in main
+    run(args)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/run.py", line 870, in run
+    elastic_launch(
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 132, in __call__
+    return launch_agent(self._config, self._entrypoint, list(args))
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 263, in launch_agent
+    raise ChildFailedError(
+torch.distributed.elastic.multiprocessing.errors.ChildFailedError: 
+============================================================
+/home/zxn/Forewarn/vlm/llama-recipes/recipes/quickstart/finetuning/finetuning_wm.py FAILED
+------------------------------------------------------------
+Failures:
+  <NO_OTHER_FAILURES>
+------------------------------------------------------------
+Root Cause (first observed failure):
+[0]:
+  time      : 2026-04-09_10:46:32
+  host      : node0029
+  rank      : 0 (local_rank: 0)
+  exitcode  : 1 (pid: 688539)
+  error_file: <N/A>
+  traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
+============================================================

logs/vlm_train_stack_cups_gpu2_retry.log

@@ -0,0 +1,88 @@
+nohup: ignoring input
+ROBOMIMIC WARNING(
+    No private macro file found!
+    It is recommended to use a private macro file
+    To setup, run: python /home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/robomimic/scripts/setup_macros.py
+)
+Warning: unknown parameter wm_ckpt
+Warning: unknown parameter target_modules
+wandb: Using wandb-core as the SDK backend. Please refer to https://wandb.me/wandb-core for more information.
+wandb: Currently logged in as: yukizhang0527 (yukizhang0527-harbin-institute-of-technology). Use `wandb login --relogin` to force relogin
+wandb: - Waiting for wandb.init()...
+wandb: Run data is saved locally in /home/zxn/Forewarn/wandb/run-20260409_104213-zgx12i43
+wandb: Run `wandb offline` to turn off syncing.
+wandb: Syncing run splendid-gorge-62
+wandb: ⭐️ View project at https://wandb.ai/yukizhang0527-harbin-institute-of-technology/llama_recipes
+wandb: 🚀 View run at https://wandb.ai/yukizhang0527-harbin-institute-of-technology/llama_recipes/runs/zgx12i43
+Warning: custom_dataset does not accept parameter: custom_dataset.task_name
+
+--> Model /data/colosseum_dataset/wm_data/mllama_base/Llama-3.2-11B-Vision-Instruct/custom
+
+--> /data/colosseum_dataset/wm_data/mllama_base/Llama-3.2-11B-Vision-Instruct/custom has 9808.885776 Million params
+
+loading world model from ckpt path None
+Traceback (most recent call last):
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 540, in _check_seekable
+    f.seek(f.tell())
+AttributeError: 'NoneType' object has no attribute 'seek'
+
+During handling of the above exception, another exception occurred:
+
+Traceback (most recent call last):
+  File "/home/zxn/Forewarn/vlm/llama-recipes/recipes/quickstart/finetuning/finetuning_wm.py", line 9, in <module>
+    fire.Fire(main)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/fire/core.py", line 135, in Fire
+    component_trace = _Fire(component, args, parsed_flag_args, context, name)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/fire/core.py", line 468, in _Fire
+    component, remaining_args = _CallAndUpdateTrace(
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/fire/core.py", line 684, in _CallAndUpdateTrace
+    component = fn(*varargs, **kwargs)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/llama_recipes/finetuning_wm.py", line 355, in main
+    model.initialize_vision_model()
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/llama_recipes/models/mllama_model.py", line 140, in initialize_vision_model
+    self.wm_model = Dreamer.from_pretrained(path = wm_config.from_ckpt, obs_space = obs_space,
+  File "/home/zxn/Forewarn/model_based_irl_torch/dreamer/dreamer.py", line 62, in from_pretrained
+    ckpt = torch.load(path)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 997, in load
+    with _open_file_like(f, 'rb') as opened_file:
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 449, in _open_file_like
+    return _open_buffer_reader(name_or_buffer)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 434, in __init__
+    _check_seekable(buffer)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 543, in _check_seekable
+    raise_err_msg(["seek", "tell"], e)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/serialization.py", line 536, in raise_err_msg
+    raise type(e)(msg)
+AttributeError: 'NoneType' object has no attribute 'seek'. You can only torch.load from a file that is seekable. Please pre-load the data into a buffer like io.BytesIO and try to load from it instead.
+wandb: 🚀 View run splendid-gorge-62 at: https://wandb.ai/yukizhang0527-harbin-institute-of-technology/llama_recipes/runs/zgx12i43
+wandb: Find logs at: wandb/run-20260409_104213-zgx12i43/logs
+E0409 10:42:32.646334 140491694446400 torch/distributed/elastic/multiprocessing/api.py:826] failed (exitcode: 1) local_rank: 0 (pid: 686029) of binary: /home/zxn/anaconda3/envs/dreamer/bin/python3.10
+Traceback (most recent call last):
+  File "/home/zxn/anaconda3/envs/dreamer/bin/torchrun", line 6, in <module>
+    sys.exit(main())
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 347, in wrapper
+    return f(*args, **kwargs)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/run.py", line 879, in main
+    run(args)
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/run.py", line 870, in run
+    elastic_launch(
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 132, in __call__
+    return launch_agent(self._config, self._entrypoint, list(args))
+  File "/home/zxn/anaconda3/envs/dreamer/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 263, in launch_agent
+    raise ChildFailedError(
+torch.distributed.elastic.multiprocessing.errors.ChildFailedError: 
+============================================================
+/home/zxn/Forewarn/vlm/llama-recipes/recipes/quickstart/finetuning/finetuning_wm.py FAILED
+------------------------------------------------------------
+Failures:
+  <NO_OTHER_FAILURES>
+------------------------------------------------------------
+Root Cause (first observed failure):
+[0]:
+  time      : 2026-04-09_10:42:32
+  host      : node0029
+  rank      : 0 (local_rank: 0)
+  exitcode  : 1 (pid: 686029)
+  error_file: <N/A>
+  traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html
+============================================================