diff --git a/000500/pretrained_model/config.json b/000500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/000500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/000500/pretrained_model/model.safetensors b/000500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e5022d052cac49d9d22c76f04b6dcc7fd5ab7f59 --- /dev/null +++ b/000500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b0936056fd15678e30f1fb618f49a8df5f6a07494c5cea059ff13dbb33c9a9ae +size 1066517752 diff --git a/000500/pretrained_model/train_config.json b/000500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/000500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/000500/training_state/optimizer_param_groups.json b/000500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..e3e0bec8229eb7bf5358615649deb795d7641e75 --- /dev/null +++ b/000500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 0.0001, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/000500/training_state/optimizer_state.safetensors b/000500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..16c46692a2d706fddf353887ebab21f46ad57908 --- /dev/null +++ b/000500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dff27dfb1aa8e7ae458ff898b287d17d98a9fb6179afae630d59e287fadbadad +size 2133039080 diff --git a/000500/training_state/rng_state.safetensors b/000500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c3f3b18e5e30196d3c84ae922fbbbae5f98d789c --- /dev/null +++ b/000500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:61a69cc74c23cb197fadb2dc7337b56e91d13f40b322ddee76d2305bc7adead5 +size 15708 diff --git a/000500/training_state/scheduler_state.json b/000500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..6757fe472ff57da7ebfc17638a29b3227f0231ab --- /dev/null +++ b/000500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 500, + "_step_count": 501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 0.0001 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/000500/training_state/training_step.json b/000500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..8df625539d7e10ddfbd1e2ac0daf2fdd64ff5c4b --- /dev/null +++ b/000500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 500 +} \ No newline at end of file diff --git a/001000/pretrained_model/config.json b/001000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/001000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/001000/pretrained_model/model.safetensors b/001000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..304353dbc68780ae16f07cf22e73cc671eddd32c --- /dev/null +++ b/001000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:023da28968f1b0f4a445a92e2d0f2defe40148cf134bc936cf40846c0a40565c +size 1066517752 diff --git a/001000/pretrained_model/train_config.json b/001000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/001000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/001000/training_state/optimizer_param_groups.json b/001000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..e156667fd65b6d84ea3072547824c7b1ef1028cd --- /dev/null +++ b/001000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.999376947588288e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/001000/training_state/optimizer_state.safetensors b/001000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ffc34a2a25db4884dfc761caa0e85a5321fcb98d --- /dev/null +++ b/001000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f3ff64691c3e252b929f0762f86c0c608745e4c0480a0d6b514285efce4021a7 +size 2133039080 diff --git a/001000/training_state/rng_state.safetensors b/001000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9130949d6b3e57c310a557ca0cef52dd0d0f910e --- /dev/null +++ b/001000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e330b5a95d13700a91d14ddc3fa7e49d0e0f4a7e7d246680c2d1e0082d6df1fb +size 15708 diff --git a/001000/training_state/scheduler_state.json b/001000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..f6d1e8dab7e5abef67a7e1c679fc476c3fc18c75 --- /dev/null +++ b/001000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 1000, + "_step_count": 1001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.999376947588288e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/001000/training_state/training_step.json b/001000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..d98c94f5b78238bf495ac68b9f9fb446cfac5c07 --- /dev/null +++ b/001000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 1000 +} \ No newline at end of file diff --git a/001500/pretrained_model/config.json b/001500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/001500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/001500/pretrained_model/model.safetensors b/001500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..560500a4aba74a849b51173811eb8a0d9604b466 --- /dev/null +++ b/001500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:da9bf602da9ff0c81eb58c38091c8a1353825b54bef258ab8138f53e38b5597c +size 1066517752 diff --git a/001500/pretrained_model/train_config.json b/001500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/001500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/001500/training_state/optimizer_param_groups.json b/001500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..e45408390671b6ca6a24054fa7d0373a516eb0af --- /dev/null +++ b/001500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.99750794563087e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/001500/training_state/optimizer_state.safetensors b/001500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..241e8fbc265ce622c641544330df6dd7c1fbb19b --- /dev/null +++ b/001500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:301737330900604505fbc158cd277b62fcb756ef7c4af36e6fa787c3cb4498dc +size 2133039080 diff --git a/001500/training_state/rng_state.safetensors b/001500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..075c98026a7f6aa408be09bb6c5e76136c22de02 --- /dev/null +++ b/001500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fcb2b6598cdd31cb8adc9e4d4379a6dc370313cc02ea14c5924130b70f0ddf04 +size 15708 diff --git a/001500/training_state/scheduler_state.json b/001500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..732c5f4589bed68339e45920b9fbd6aecced6d9b --- /dev/null +++ b/001500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 1500, + "_step_count": 1501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.99750794563087e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/001500/training_state/training_step.json b/001500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..332c190c78659d9403699dab54afdd72e41d21b8 --- /dev/null +++ b/001500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 1500 +} \ No newline at end of file diff --git a/002000/pretrained_model/config.json b/002000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/002000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/002000/pretrained_model/model.safetensors b/002000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b108ca6ce0ae6a05d690fea70254cdc59d747994 --- /dev/null +++ b/002000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fae16c26d94e88a57d9573c30b9f772dc3e86e2af8fbd44062e4391271a0c078 +size 1066517752 diff --git a/002000/pretrained_model/train_config.json b/002000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/002000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/002000/training_state/optimizer_param_groups.json b/002000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..53a6ae79bc680b1fdab52b93873815722384c440 --- /dev/null +++ b/002000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.994393459922218e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/002000/training_state/optimizer_state.safetensors b/002000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bb30c19aef6c3e3c9e9f1ff35435943f284e4af1 --- /dev/null +++ b/002000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e6cd97f880efa82d4da8f68efce33351a814e4a2a0efb518e13f800ed49c97a9 +size 2133039080 diff --git a/002000/training_state/rng_state.safetensors b/002000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..878f835293743ab1c7a7d2231d02895bdd92bd03 --- /dev/null +++ b/002000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:58de992f04989b1da3cc4cc6a34ac456ca1d8f4d7d00fbf1a0d2f111d7d7b149 +size 15708 diff --git a/002000/training_state/scheduler_state.json b/002000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..c096396197b150d39ab3ab6ca041df3af8989393 --- /dev/null +++ b/002000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 2000, + "_step_count": 2001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.994393459922218e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/002000/training_state/training_step.json b/002000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..39090bbb986edb821e1602990d19357dcdb5d2ae --- /dev/null +++ b/002000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 2000 +} \ No newline at end of file diff --git a/002500/pretrained_model/config.json b/002500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/002500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/002500/pretrained_model/model.safetensors b/002500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..718c9f356a29ae9545112892ab11ce8a2f24df1e --- /dev/null +++ b/002500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:77359b7a18551ebbb536ae89cfa572d28ad5f478f05e4e2fbbb12e932160d3fc +size 1066517752 diff --git a/002500/pretrained_model/train_config.json b/002500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/002500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/002500/training_state/optimizer_param_groups.json b/002500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..23002a43106b75d9640f163e7db56a1276baadc6 --- /dev/null +++ b/002500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.990034266657467e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/002500/training_state/optimizer_state.safetensors b/002500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3a2c57eb3f9c9e18fc6473521062f19e42779c1c --- /dev/null +++ b/002500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8f2a0122195d4a62b6f11f4204c7776169e9eeb2cb56a7742477fb9575265b62 +size 2133039080 diff --git a/002500/training_state/rng_state.safetensors b/002500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..beb457541d34eba244d5210eabe3c800c91cbded --- /dev/null +++ b/002500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:21745d5f336f656945e02a374f7b3ad442e33117e3d6fe4bfef2d07fff4bef14 +size 15708 diff --git a/002500/training_state/scheduler_state.json b/002500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..32ed3be8fb6d29cbc3f9fa4e8a8545ad7697a7a2 --- /dev/null +++ b/002500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 2500, + "_step_count": 2501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.990034266657467e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/002500/training_state/training_step.json b/002500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..d265f47a09c2910099ed59e197b57b34675d1ae0 --- /dev/null +++ b/002500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 2500 +} \ No newline at end of file diff --git a/003000/pretrained_model/config.json b/003000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/003000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/003000/pretrained_model/model.safetensors b/003000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ee032eb2edad17544dae9cb4bd8b2d348117361d --- /dev/null +++ b/003000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:804ffab9f60c6423c3b8ac49c02e194d4e8d859128140871d45af46d8ef87015 +size 1066517752 diff --git a/003000/pretrained_model/train_config.json b/003000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/003000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/003000/training_state/optimizer_param_groups.json b/003000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..27182c1202dfb4d016865d905813c9dabb15de72 --- /dev/null +++ b/003000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.984431452238967e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/003000/training_state/optimizer_state.safetensors b/003000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..7e6a5acb57e81bc1bf5cb1fafd3ab4619c6b6dc2 --- /dev/null +++ b/003000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9287e3a4b6aa7ea654dc5cf8ced678e17a203d70cf1abede9444665817da4189 +size 2133039080 diff --git a/003000/training_state/rng_state.safetensors b/003000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5ca3c5b6be520ad99378278444e27667a1035277 --- /dev/null +++ b/003000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7fc92c0c7222f74e91b033c3c750712e61f9176923b0005074847f779ae3d91f +size 15708 diff --git a/003000/training_state/scheduler_state.json b/003000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..a6d87d0102d5509f89bde79e6f225ecef6a05e94 --- /dev/null +++ b/003000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 3000, + "_step_count": 3001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.984431452238967e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/003000/training_state/training_step.json b/003000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..c1a44127b7dfea653fd776d529fa83c55d32081c --- /dev/null +++ b/003000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 3000 +} \ No newline at end of file diff --git a/003500/pretrained_model/config.json b/003500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/003500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/003500/pretrained_model/model.safetensors b/003500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..eb65ef9f349a4c1ff32379dfb1863ff9bacf245e --- /dev/null +++ b/003500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9f928cfc843629d630abbc183cc450c382c9b036b1199d7b5527b8dd871df247 +size 1066517752 diff --git a/003500/pretrained_model/train_config.json b/003500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/003500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/003500/training_state/optimizer_param_groups.json b/003500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..6190e1e09096fdf1065766f5f2df624de2fb6377 --- /dev/null +++ b/003500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.977586413005531e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/003500/training_state/optimizer_state.safetensors b/003500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e9e6de123bfbdf82e6f35a2a46fd3d17de7a3817 --- /dev/null +++ b/003500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:974b6b1a40276f19c889149b01751288e8730940bc6b9b0f6a3bfff1c83fcd7e +size 2133039080 diff --git a/003500/training_state/rng_state.safetensors b/003500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fd2eb88929868bc2a26c2fb193b7ff75d908c6e6 --- /dev/null +++ b/003500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9578c3233237b3d46e4a699b43d35a1d2f59cbc21b5fae1dec44af6c453ce8ce +size 15708 diff --git a/003500/training_state/scheduler_state.json b/003500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..e0e429f01881468cfac7abd8506d6ca6c967a48a --- /dev/null +++ b/003500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 3500, + "_step_count": 3501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.977586413005531e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/003500/training_state/training_step.json b/003500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..ae7ae4f2823ad80bcbd73e17fec39ab540bebe83 --- /dev/null +++ b/003500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 3500 +} \ No newline at end of file diff --git a/004000/pretrained_model/config.json b/004000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/004000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/004000/pretrained_model/model.safetensors b/004000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ced018c9867e7b356e89b58517a8035441067d89 --- /dev/null +++ b/004000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:11938c076af85bb4b3063da0619712250b3ba8e3db6935e9df8d358e4aedc546 +size 1066517752 diff --git a/004000/pretrained_model/train_config.json b/004000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/004000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/004000/training_state/optimizer_param_groups.json b/004000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..00940512b2b29633f6e3c989c336a08403975119 --- /dev/null +++ b/004000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.96950085488444e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/004000/training_state/optimizer_state.safetensors b/004000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..86915101d9ebd0d3792d5857016362e3634cbc12 --- /dev/null +++ b/004000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f05941446414ee9b5508dc181c1a050f72c5eeb92de28295519aa676ab6c8fc1 +size 2133039080 diff --git a/004000/training_state/rng_state.safetensors b/004000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0dcad78cda52b2011f986703365fc91fb55abb5c --- /dev/null +++ b/004000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a07497a610dacb0fad7e09487888662da5f09834c0f052ce784b9e686fa71746 +size 15708 diff --git a/004000/training_state/scheduler_state.json b/004000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..86075d95a07026ff5b7e195b2c3125bc68843a82 --- /dev/null +++ b/004000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 4000, + "_step_count": 4001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.96950085488444e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/004000/training_state/training_step.json b/004000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..75d870521c191f77dd9eaa4d83486eab6e768f69 --- /dev/null +++ b/004000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 4000 +} \ No newline at end of file diff --git a/004500/pretrained_model/config.json b/004500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/004500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/004500/pretrained_model/model.safetensors b/004500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b87775932ce3b79783a9f3beb836a777c43455e7 --- /dev/null +++ b/004500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5cd612b86fcccb4fc54663a0f32eca93c0d60d1b3386fb6ed8277a7043d68603 +size 1066517752 diff --git a/004500/pretrained_model/train_config.json b/004500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/004500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/004500/training_state/optimizer_param_groups.json b/004500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..1d3270e584cdeda8ef8d1648fa1125a62327fe26 --- /dev/null +++ b/004500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.960176792966289e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/004500/training_state/optimizer_state.safetensors b/004500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8b4f6f56d425109a9874c25a8e64d8c20e37447a --- /dev/null +++ b/004500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0cc6bf4f720d528ba02a378d1a991a5c19854f9cd0ff74fa59b00ed0c235afc9 +size 2133039080 diff --git a/004500/training_state/rng_state.safetensors b/004500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0d62e1c0e651422ac061aa8048cf7ce1f20100b4 --- /dev/null +++ b/004500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f7cfc69c6dc48af2a1a920ca53e6922375c08025fe6412deadb25eec9b9a35ea +size 15708 diff --git a/004500/training_state/scheduler_state.json b/004500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..07cc4733862bc43e96595c2c0b46c78b98acc949 --- /dev/null +++ b/004500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 4500, + "_step_count": 4501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.960176792966289e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/004500/training_state/training_step.json b/004500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..97f44a46f3389e808bf52d71d209b5cd5d38d52d --- /dev/null +++ b/004500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 4500 +} \ No newline at end of file diff --git a/005000/pretrained_model/config.json b/005000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/005000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/005000/pretrained_model/model.safetensors b/005000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..282fbfe8799f306bf2b6e07dafbb659e38860d4d --- /dev/null +++ b/005000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a36eac13da563478b4258fdaf1b8ff907761966b7327404c703ebd970dc58f6c +size 1066517752 diff --git a/005000/pretrained_model/train_config.json b/005000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/005000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/005000/training_state/optimizer_param_groups.json b/005000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..062d62eb531f0bc4b515f75ad7fb7df02995cd4e --- /dev/null +++ b/005000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.949616551002787e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/005000/training_state/optimizer_state.safetensors b/005000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..45c57525005bd1559853acb450a58d469c1afeb4 --- /dev/null +++ b/005000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:05220f06a0ed18e32e4a8b13726ff047c23d856c90f36732605419498f614aed +size 2133039080 diff --git a/005000/training_state/rng_state.safetensors b/005000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2ec108f1070dd7a2dd214c2cf019b02cdc6988ce --- /dev/null +++ b/005000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:88dbb65924b9aa816d9b557c191a220154f6b4213e97cedc051efdd824596cdf +size 15708 diff --git a/005000/training_state/scheduler_state.json b/005000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..76741cc63d3f487ac15a5d13d3710d3c6c97a3b8 --- /dev/null +++ b/005000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 5000, + "_step_count": 5001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.949616551002787e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/005000/training_state/training_step.json b/005000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..592449d3fc8b35c30c604eb1dabe60537e8224a0 --- /dev/null +++ b/005000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 5000 +} \ No newline at end of file diff --git a/005500/pretrained_model/config.json b/005500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/005500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/005500/pretrained_model/model.safetensors b/005500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..94393317dc2ebdc97a334af7d1d957370ca37c54 --- /dev/null +++ b/005500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a1af3d2f215fe3115325257ce00129db79e86353aafc7af8299782397691f798 +size 1066517752 diff --git a/005500/pretrained_model/train_config.json b/005500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/005500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/005500/training_state/optimizer_param_groups.json b/005500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..47db660a3024abf548d70849b9dcde45d0fd8406 --- /dev/null +++ b/005500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.93782276082762e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/005500/training_state/optimizer_state.safetensors b/005500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6c68820dceb2b9d94d2eb59492741a99600096c6 --- /dev/null +++ b/005500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d8c00fcf73b042b4217340301ca01ad1cae0623e92a5c0bb5c86d1dcf80f19e +size 2133039080 diff --git a/005500/training_state/rng_state.safetensors b/005500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..528a5b264809a60032401d02bc468d43c4a2f21c --- /dev/null +++ b/005500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:37abb1ccb75ce0acba816d35015d55e85f6458e9813f817664065ebb9b2fb6a2 +size 15708 diff --git a/005500/training_state/scheduler_state.json b/005500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..90db0a6e3955d3db002ac6d89d8b275cb32126ad --- /dev/null +++ b/005500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 5500, + "_step_count": 5501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.93782276082762e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/005500/training_state/training_step.json b/005500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..834c98398467b7a3f63786150a5f8159aa6fccfc --- /dev/null +++ b/005500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 5500 +} \ No newline at end of file diff --git a/006000/pretrained_model/config.json b/006000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/006000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/006000/pretrained_model/model.safetensors b/006000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..32f4b06a71646ef54a5254fd40e1d71720254525 --- /dev/null +++ b/006000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5f643d8878bbd4a5cf3eaff5d55c3087cec81463121b03db26d3304bb7c236ee +size 1066517752 diff --git a/006000/pretrained_model/train_config.json b/006000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/006000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/006000/training_state/optimizer_param_groups.json b/006000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..be6edaee1933c6ce9d25939d23e3811888e58fae --- /dev/null +++ b/006000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.924798361700553e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/006000/training_state/optimizer_state.safetensors b/006000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..821f7076d2a58a5144262fb6ad7b3ea23b49d4b3 --- /dev/null +++ b/006000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5bf2bbc108fd46498947d02193ef2ac20fb33951b2875c094b0587a84c02e1e2 +size 2133039080 diff --git a/006000/training_state/rng_state.safetensors b/006000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b21a2c4988ef2d2dabc65bddb6a6aa71ebc8b74d --- /dev/null +++ b/006000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6ca1441bd1a00d37c59a99c8fa4e98b596eabeb14ef87c07ac220bbfac741252 +size 15708 diff --git a/006000/training_state/scheduler_state.json b/006000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..f75e57ceb3503b478a658b577792bb4a772f562b --- /dev/null +++ b/006000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 6000, + "_step_count": 6001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.924798361700553e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/006000/training_state/training_step.json b/006000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..e267ac589be64705f8674638b9f5099c886778da --- /dev/null +++ b/006000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 6000 +} \ No newline at end of file diff --git a/006500/pretrained_model/config.json b/006500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/006500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/006500/pretrained_model/model.safetensors b/006500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..137b587ccdd749027f1fedbd8db7ccb1e80d9103 --- /dev/null +++ b/006500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c63598d440a6cb3387a410384457295c617d5297cd4f8dd7ba8aefc5714f8b5f +size 1066517752 diff --git a/006500/pretrained_model/train_config.json b/006500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/006500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/006500/training_state/optimizer_param_groups.json b/006500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..db5604ae113247434ab8505bb40c947c5d63ca36 --- /dev/null +++ b/006500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.910546599574902e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/006500/training_state/optimizer_state.safetensors b/006500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cfce5c9d4245879c1084609ce7be8fdbadb8b0a5 --- /dev/null +++ b/006500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:87ef570b504d976acc90aa730fce36e6fe2c326c001d1b15a6cc4fe0c06cd475 +size 2133039080 diff --git a/006500/training_state/rng_state.safetensors b/006500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a618e42b1bbed58e0b4f8608a16b73305ff2c607 --- /dev/null +++ b/006500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:044597bd13848df665d653cc33a8bdfd918d7b5a7628ac14c55b1655409fc74d +size 15708 diff --git a/006500/training_state/scheduler_state.json b/006500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..10c30c35f841f806121f3e1ab3911293048e2563 --- /dev/null +++ b/006500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 6500, + "_step_count": 6501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.910546599574902e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/006500/training_state/training_step.json b/006500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..c72d982a5874b38499a7364727b4f7259dfac7ed --- /dev/null +++ b/006500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 6500 +} \ No newline at end of file diff --git a/007000/pretrained_model/config.json b/007000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/007000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/007000/pretrained_model/model.safetensors b/007000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f0b70a426f6ece4624a9290ca931915515c75f2c --- /dev/null +++ b/007000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9d1cd24a4e1333e3d9d36a9a084dd76c43c4bf125a99232e6e28658497295ca8 +size 1066517752 diff --git a/007000/pretrained_model/train_config.json b/007000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/007000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/007000/training_state/optimizer_param_groups.json b/007000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..99eee615ca5be8186c33b90efb5cc3dd445a9049 --- /dev/null +++ b/007000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.895071026288574e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/007000/training_state/optimizer_state.safetensors b/007000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..52106409a8c71d2b53cf861651a0395898618c2b --- /dev/null +++ b/007000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9429187829ba094c23540546b5d883d2869dbdfcdaf0d3197dcdcc6409b5b594 +size 2133039080 diff --git a/007000/training_state/rng_state.safetensors b/007000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cff2a93fc6a386f446448d2b85dda4548cd4a2fc --- /dev/null +++ b/007000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d8dca620c4b4c80cfe7639e69b09cc2a21b745166eb5f4cf001a38e28f3fb6da +size 15708 diff --git a/007000/training_state/scheduler_state.json b/007000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..00fabef49b141efabc2e21bacc896f880bc23abe --- /dev/null +++ b/007000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 7000, + "_step_count": 7001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.895071026288574e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/007000/training_state/training_step.json b/007000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..1f4c730c4d88ece076ae9d4e01ed7fbe60b0415d --- /dev/null +++ b/007000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 7000 +} \ No newline at end of file diff --git a/007500/pretrained_model/config.json b/007500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/007500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/007500/pretrained_model/model.safetensors b/007500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8ba612f93a0d3d1431fc01557187a606e26ca9cb --- /dev/null +++ b/007500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0dd45649130ca2fabd761bb53b9cda0532febcbf846fa0eab21edc31282af6f6 +size 1066517752 diff --git a/007500/pretrained_model/train_config.json b/007500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/007500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/007500/training_state/optimizer_param_groups.json b/007500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..cce0da881eeacf40dbd346588b5006e64ee478ea --- /dev/null +++ b/007500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.87837549867887e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/007500/training_state/optimizer_state.safetensors b/007500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..23a6c38fd67940fc80d82fff1d206e77160c4ea1 --- /dev/null +++ b/007500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:34f7158b0e4a082796a8d91a4b8f4d22ca53ac5cbe73bd6c8e968183ac105037 +size 2133039080 diff --git a/007500/training_state/rng_state.safetensors b/007500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fc538c9f9406791d73e7ca55aeb713e49320b4b3 --- /dev/null +++ b/007500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4bf77151c3462e3a1ecb31e063ec62fc9e3ff2c783029579d164c9e0977d050e +size 15708 diff --git a/007500/training_state/scheduler_state.json b/007500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..bd8094ceb22bc1c019c923e2bb8d4f9462f45e4b --- /dev/null +++ b/007500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 7500, + "_step_count": 7501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.87837549867887e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/007500/training_state/training_step.json b/007500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..119265efe00a465e1aea6ca140089eb7aa8795c0 --- /dev/null +++ b/007500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 7500 +} \ No newline at end of file diff --git a/008000/pretrained_model/config.json b/008000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/008000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/008000/pretrained_model/model.safetensors b/008000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d77104b92b8444af37fabb42cce8757255e96507 --- /dev/null +++ b/008000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fff478d416848b2c5abdabb0c9bcce7c8767af5e98950ff30bd6366e54e59640 +size 1066517752 diff --git a/008000/pretrained_model/train_config.json b/008000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/008000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/008000/training_state/optimizer_param_groups.json b/008000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..5384dc35d40ab6c03636434115f42df3763fb2b2 --- /dev/null +++ b/008000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.860464177621284e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/008000/training_state/optimizer_state.safetensors b/008000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ab53152fc976e0d2f694b100efdc05be5f2bb2a3 --- /dev/null +++ b/008000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4f187cd5f5af2f0c08393864eba6095cae30cacbeed66d6cbe06905f5115e277 +size 2133039080 diff --git a/008000/training_state/rng_state.safetensors b/008000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2fff2bc96b3f7cd156d181b5b6de40d2bd917d67 --- /dev/null +++ b/008000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:64c149e03740684eb18e0b8beec64018e44fdf6801b5d43172b580c0634a8cdb +size 15708 diff --git a/008000/training_state/scheduler_state.json b/008000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3906d6e1aba7157af26139c6656da2ff86b2bff6 --- /dev/null +++ b/008000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 8000, + "_step_count": 8001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.860464177621284e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/008000/training_state/training_step.json b/008000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..74d8cba01cab8506617b2cbae6f268fe80fbfa79 --- /dev/null +++ b/008000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 8000 +} \ No newline at end of file diff --git a/008500/pretrained_model/config.json b/008500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/008500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/008500/pretrained_model/model.safetensors b/008500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..66066bea929578721037bb2eb03269f2322fc16e --- /dev/null +++ b/008500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fd192dff3f782ab25cba56d1e9ecf1da8184b6f1f22957ed396fb2b05497e165 +size 1066517752 diff --git a/008500/pretrained_model/train_config.json b/008500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/008500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/008500/training_state/optimizer_param_groups.json b/008500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..171f4e1bb100bc8e95aa08efa3e44be177e07693 --- /dev/null +++ b/008500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.841341526992536e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/008500/training_state/optimizer_state.safetensors b/008500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5c9008541c599282c2e710ff40ce4f58c2eb7622 --- /dev/null +++ b/008500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ac4135814b1d93cb33f5f389b10c890b84f62b28f29a31de5d5f5d174253f189 +size 2133039080 diff --git a/008500/training_state/rng_state.safetensors b/008500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..387db4f1ef1347fc245e55281636e17195e60978 --- /dev/null +++ b/008500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7c8edd86575a48e1831145b69656e27798445fdff749b096381cc8dc8f2512d5 +size 15708 diff --git a/008500/training_state/scheduler_state.json b/008500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..d621b042826b2dcd6e96947b6c4e05f446fe1b4f --- /dev/null +++ b/008500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 8500, + "_step_count": 8501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.841341526992536e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/008500/training_state/training_step.json b/008500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..b346f4430cefeb66f3951c6c881e44d7ece511cb --- /dev/null +++ b/008500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 8500 +} \ No newline at end of file diff --git a/009000/pretrained_model/config.json b/009000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/009000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/009000/pretrained_model/model.safetensors b/009000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4f106edb6e4c84799a57d88e5776aae568f35bd2 --- /dev/null +++ b/009000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fb87b4da148744c6448ffd948132ecd45375be061736628c0ea50ed01ba70c38 +size 1066517752 diff --git a/009000/pretrained_model/train_config.json b/009000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/009000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/009000/training_state/optimizer_param_groups.json b/009000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..5b288b445ed034768421cf262278ca89f942896a --- /dev/null +++ b/009000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.821012312558058e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/009000/training_state/optimizer_state.safetensors b/009000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cc0b2c9eda6d6fc712c3c1bd497920912e1c4a90 --- /dev/null +++ b/009000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a3d7efffb9eaebb90018098922de05bc654c534adfe761747f8416cda00952db +size 2133039080 diff --git a/009000/training_state/rng_state.safetensors b/009000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..549bae23e8ae40431d8e9cfb766f9952396dd325 --- /dev/null +++ b/009000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:758a80037ace31887944f921ae47bd656ace24691a9f10e1422b697883823b12 +size 15708 diff --git a/009000/training_state/scheduler_state.json b/009000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..ce0d15f45683dc1b38696f9e37aab9b40543e770 --- /dev/null +++ b/009000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 9000, + "_step_count": 9001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.821012312558058e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/009000/training_state/training_step.json b/009000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..7247596175281e9d3f295d2b6936944a71b64d34 --- /dev/null +++ b/009000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 9000 +} \ No newline at end of file diff --git a/009500/pretrained_model/config.json b/009500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/009500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/009500/pretrained_model/model.safetensors b/009500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0d1575f94a3b48d9d4a944766ff0f21f737840e4 --- /dev/null +++ b/009500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d67315f325348c9f500b12b0158647ffbc9478aed7121b651451c7d8fc482287 +size 1066517752 diff --git a/009500/pretrained_model/train_config.json b/009500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/009500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/009500/training_state/optimizer_param_groups.json b/009500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..f67e17976db90b8a48232e50aa1f2d6705fc48af --- /dev/null +++ b/009500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.799481600784286e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/009500/training_state/optimizer_state.safetensors b/009500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8337c062ce4287af967f6e3a925a775010fe4d01 --- /dev/null +++ b/009500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d7f3202866d0549d9b43ee591a5934fea3f244f9a7979c84b3438adb1db8eb80 +size 2133039080 diff --git a/009500/training_state/rng_state.safetensors b/009500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..338c22fd555ba2b9b751f68362069d76517be281 --- /dev/null +++ b/009500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:74d38da3b550d99a6742e7d6d117f9b700642f5d10cd093a0b69f464ff0a20a5 +size 15708 diff --git a/009500/training_state/scheduler_state.json b/009500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3875c4a871404ede23e8dc92002e15a9fe0ce42d --- /dev/null +++ b/009500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 9500, + "_step_count": 9501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.799481600784286e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/009500/training_state/training_step.json b/009500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..ac6e52c486e35ee2dd3f06f836d8903da4e02082 --- /dev/null +++ b/009500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 9500 +} \ No newline at end of file diff --git a/010000/pretrained_model/config.json b/010000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/010000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/010000/pretrained_model/model.safetensors b/010000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..33bf3accfadd799f4f63c1a733ab1cb7425946b5 --- /dev/null +++ b/010000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8efc41be9ab99e44de333031dab84c2c6c5e948c4f864ec9ab24e0d9cba75804 +size 1066517752 diff --git a/010000/pretrained_model/train_config.json b/010000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/010000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/010000/training_state/optimizer_param_groups.json b/010000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..8d79cff91f230e22f4bcac8784367cd818d224d9 --- /dev/null +++ b/010000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.776754757575975e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/010000/training_state/optimizer_state.safetensors b/010000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5c5d798dc913c3d86e32c9937daa8ad395f0be0d --- /dev/null +++ b/010000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:afaf055891f0b001b20c3fb6221bcd87bec8a4d0089f1ed7f9b73fa102dd59be +size 2133039080 diff --git a/010000/training_state/rng_state.safetensors b/010000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c5edeaeae30552583126332a12b2e85aa60aa587 --- /dev/null +++ b/010000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5e9ede37727c489b529def99c636f0c7c3d81061701073e4e987ecbae3525f18 +size 15708 diff --git a/010000/training_state/scheduler_state.json b/010000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..bf830e662415d2591f66bfac4461364019cb03fe --- /dev/null +++ b/010000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 10000, + "_step_count": 10001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.776754757575975e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/010000/training_state/training_step.json b/010000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..7cb7c0986e9e7461ca851ce71e95d235ae3d2732 --- /dev/null +++ b/010000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 10000 +} \ No newline at end of file diff --git a/010500/pretrained_model/config.json b/010500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/010500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/010500/pretrained_model/model.safetensors b/010500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e133878ae8be05f2c9cdde71af83b1153984b676 --- /dev/null +++ b/010500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:24695a55dd4ab6e9d528cea125e01cdbc926d168cf4cace1bf0bfeccc2105b49 +size 1066517752 diff --git a/010500/pretrained_model/train_config.json b/010500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/010500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/010500/training_state/optimizer_param_groups.json b/010500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..1e8110964d43be6afbadcc44bf88b37081974dce --- /dev/null +++ b/010500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.752837446938915e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/010500/training_state/optimizer_state.safetensors b/010500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f5717383c4eeb646f2f425ca5ebbfb067440fbf4 --- /dev/null +++ b/010500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:55f4c2a64156a05a2bfdd37201f23023709bb2410db2d7518daf9c56829b438c +size 2133039080 diff --git a/010500/training_state/rng_state.safetensors b/010500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..17d39d55ae5b70eeea64fa8f0d3d04a0c701fb13 --- /dev/null +++ b/010500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:848d13dc9d8e13b722b5e753bbd5ebaf5b627870ee35da766fa27516a4f4a6b1 +size 15708 diff --git a/010500/training_state/scheduler_state.json b/010500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..54590813be125ce0fafbfc5421f78e5100e746c9 --- /dev/null +++ b/010500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 10500, + "_step_count": 10501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.752837446938915e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/010500/training_state/training_step.json b/010500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..eb4a10eb093bdfa12075c38d2ae912999f69c6a3 --- /dev/null +++ b/010500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 10500 +} \ No newline at end of file diff --git a/011000/pretrained_model/config.json b/011000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/011000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/011000/pretrained_model/model.safetensors b/011000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..acc52b31557e92f6d09dd41782cff96fa19ad755 --- /dev/null +++ b/011000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b7a12004a2f73a92ff5f12bc91647fcd6b28aa0ef012184d73bd6378e6ef38ec +size 1066517752 diff --git a/011000/pretrained_model/train_config.json b/011000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/011000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/011000/training_state/optimizer_param_groups.json b/011000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..a0dd35bead27dfaae1e0ae2d686d976c6280e1b2 --- /dev/null +++ b/011000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.727735629568336e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/011000/training_state/optimizer_state.safetensors b/011000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2653e922628edf441cec73f20974fb033953c125 --- /dev/null +++ b/011000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:48787adfe5ee1572bb457a7fabe5bf616eabdbe5d0ff095c22ddc8e9a5b3ecd4 +size 2133039080 diff --git a/011000/training_state/rng_state.safetensors b/011000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9281115c7c5a71138798fee92f6b94044c903f16 --- /dev/null +++ b/011000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1987347da4ed57ecb6f27e6f55873b341933c6db79b79d633dc35c8b7db63e71 +size 15708 diff --git a/011000/training_state/scheduler_state.json b/011000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..c823ab8c2d1a00a5b137e834dfe1d8ef236938d8 --- /dev/null +++ b/011000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 11000, + "_step_count": 11001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.727735629568336e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/011000/training_state/training_step.json b/011000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..2389cf08c9a887d74ab54bbe783d6a4eccbbee72 --- /dev/null +++ b/011000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 11000 +} \ No newline at end of file diff --git a/011500/pretrained_model/config.json b/011500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/011500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/011500/pretrained_model/model.safetensors b/011500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5d336b6c5c20d453b7a31055720bed2bd62df7b2 --- /dev/null +++ b/011500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0cc51376debb58fa8a1b0d0087e16db51536c8ca036ebe9ddbfe7bdf9aacc72e +size 1066517752 diff --git a/011500/pretrained_model/train_config.json b/011500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/011500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/011500/training_state/optimizer_param_groups.json b/011500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..341bb5ccd3987ecd9575ce064b6790efd5ee09ef --- /dev/null +++ b/011500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.701455561363379e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/011500/training_state/optimizer_state.safetensors b/011500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b74add4b468155b67d872a4461d75a64410e5f26 --- /dev/null +++ b/011500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:54f9d7b4ccb37fcb304bfd780342487eb7f3561449646faf5e74abb51d1cec3f +size 2133039080 diff --git a/011500/training_state/rng_state.safetensors b/011500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6f4322cfd07756f7433b22bf191b4d6aa672cfd5 --- /dev/null +++ b/011500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a26863825fc1bc4d056c88e8dfd9c7bd9cc122d245a5166ad799442f5405768b +size 15708 diff --git a/011500/training_state/scheduler_state.json b/011500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..c4cfb89a92f3f7b913a65d857bb297fd7d8edd64 --- /dev/null +++ b/011500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 11500, + "_step_count": 11501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.701455561363379e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/011500/training_state/training_step.json b/011500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..91d8f370dd0c5bded6a73393f573586b75a2e90f --- /dev/null +++ b/011500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 11500 +} \ No newline at end of file diff --git a/012000/pretrained_model/config.json b/012000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/012000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/012000/pretrained_model/model.safetensors b/012000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..00394b1e13a5b3558d19b2509b264e9b41298e53 --- /dev/null +++ b/012000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:367b32c485f3cf0bea223e9b07429e1646ae92d6e1f7bf4e717a78d719cb6c60 +size 1066517752 diff --git a/012000/pretrained_model/train_config.json b/012000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/012000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/012000/training_state/optimizer_param_groups.json b/012000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..db5f6d661b54faf9b000ae446b52742f4acaa5d7 --- /dev/null +++ b/012000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.674003791867991e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/012000/training_state/optimizer_state.safetensors b/012000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a74bee80e730c36df77d2290946494089b3a77c1 --- /dev/null +++ b/012000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f80733e50dac843018d484809c5fbfd37f87b4f176a8f05eaec8521db5981cba +size 2133039080 diff --git a/012000/training_state/rng_state.safetensors b/012000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3613d39d83cdf5fb6efd8886de9e01b8c824e014 --- /dev/null +++ b/012000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dc97db1f9f5b0ee96c179c3bb69711c45481d772820a884eabf6f0a828ad6789 +size 15708 diff --git a/012000/training_state/scheduler_state.json b/012000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..6c846727a5d2c60994425dab551ad2281d08e1da --- /dev/null +++ b/012000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 12000, + "_step_count": 12001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.674003791867991e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/012000/training_state/training_step.json b/012000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..c4fb27ac819b81943e6545c7c18510bdfb8eae1b --- /dev/null +++ b/012000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 12000 +} \ No newline at end of file diff --git a/012500/pretrained_model/config.json b/012500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/012500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/012500/pretrained_model/model.safetensors b/012500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8f79d93987e4c2d2b731e315cc3ff59f24a8bf90 --- /dev/null +++ b/012500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0de7683cb126e29fed989c4750589073db163cf439777eccdfe373bfe66c3e59 +size 1066517752 diff --git a/012500/pretrained_model/train_config.json b/012500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/012500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/012500/training_state/optimizer_param_groups.json b/012500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..841f09771cc2040bfac2cb82ffd17cb856f7b143 --- /dev/null +++ b/012500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.645387162638652e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/012500/training_state/optimizer_state.safetensors b/012500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5622ca7babbcb1fb461f3cc85f574d3165f70764 --- /dev/null +++ b/012500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:df6e612abccae2b5e1244fbcac9b2361c893cc47dfd6340f45d0c09da9fe3379 +size 2133039080 diff --git a/012500/training_state/rng_state.safetensors b/012500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a93b9ace97680a8350017e18f7e547c5174c35eb --- /dev/null +++ b/012500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9423d625e077de220b25b67360deaa83ae0bcebc587032a57c7bf754075f0f81 +size 15708 diff --git a/012500/training_state/scheduler_state.json b/012500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..1998979a6ec8b6b428cb79f6d7218125362613eb --- /dev/null +++ b/012500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 12500, + "_step_count": 12501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.645387162638652e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/012500/training_state/training_step.json b/012500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..6d00d1fb30e9fa29506662caa05a99b86254c107 --- /dev/null +++ b/012500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 12500 +} \ No newline at end of file diff --git a/013000/pretrained_model/config.json b/013000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/013000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/013000/pretrained_model/model.safetensors b/013000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..be307020fe98e50b2776b73c465c87d0d6259fcd --- /dev/null +++ b/013000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b53ddf8cbfe50a31fa83668c53f849cbf8ca55e60c71f042bb3fec87b0164824 +size 1066517752 diff --git a/013000/pretrained_model/train_config.json b/013000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/013000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/013000/training_state/optimizer_param_groups.json b/013000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..1372609d432e98e0c714c821a77fbd8ce2ebac3b --- /dev/null +++ b/013000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.615612805539305e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/013000/training_state/optimizer_state.safetensors b/013000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..38d66482536609898e188dcbae59eabcde3a46c0 --- /dev/null +++ b/013000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bbf2d044debca322139d4633f688778d872aa921988134605ebe39df7a74af78 +size 2133039080 diff --git a/013000/training_state/rng_state.safetensors b/013000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..70f3856ff9fb98db5b2ccc6b3a78a7c912323567 --- /dev/null +++ b/013000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:13922c98a3c9d7e929f6565c2a9b211621a3dee31624f9558637da831c40f5f3 +size 15708 diff --git a/013000/training_state/scheduler_state.json b/013000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..decbfb5839352b88fa5e1ae141fed7913e39aa9b --- /dev/null +++ b/013000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 13000, + "_step_count": 13001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.615612805539305e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/013000/training_state/training_step.json b/013000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..6f5915e7fb38c851f5b48d58c1f66fe9c9d147ed --- /dev/null +++ b/013000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 13000 +} \ No newline at end of file diff --git a/013500/pretrained_model/config.json b/013500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/013500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/013500/pretrained_model/model.safetensors b/013500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..24e47ed3f28a9cfa851243b3fb6ede8ad362dc4c --- /dev/null +++ b/013500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:790888f5af0051b16f46ef7020d427634f3010dc7714bfc25619f4051a8576ad +size 1066517752 diff --git a/013500/pretrained_model/train_config.json b/013500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/013500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/013500/training_state/optimizer_param_groups.json b/013500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..d43632dc756191b989b99d6e9d5ba62cc7e1797d --- /dev/null +++ b/013500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.584688140963944e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/013500/training_state/optimizer_state.safetensors b/013500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1dc2afacdb6c390a4cf522c1b0b4717049d82aef --- /dev/null +++ b/013500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d4ed772545c50c8ae2a68f85a6e8d261f60fb275b68693ea0a3ba8070c78109f +size 2133039080 diff --git a/013500/training_state/rng_state.safetensors b/013500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c8cd1e55f6e9cebaec631d3d0c0fe5e80193d0d3 --- /dev/null +++ b/013500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8179b6e9e96a67c58b7599c0a3759079de7852b96b4509ddfdee303a1a5288d6 +size 15708 diff --git a/013500/training_state/scheduler_state.json b/013500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..d97a288f07474502f9facb7caee95692521efcd6 --- /dev/null +++ b/013500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 13500, + "_step_count": 13501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.584688140963944e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/013500/training_state/training_step.json b/013500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..cd668526ba807a1d676ac5735b1716d2339fe727 --- /dev/null +++ b/013500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 13500 +} \ No newline at end of file diff --git a/014000/pretrained_model/config.json b/014000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/014000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/014000/pretrained_model/model.safetensors b/014000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..65d19e0f9271eafa562204a06bd71766b78b048d --- /dev/null +++ b/014000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c4da856403f12e8f6293f04a88948db1a1fde6faf48122e3c7a81e49ae782f77 +size 1066517752 diff --git a/014000/pretrained_model/train_config.json b/014000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/014000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/014000/training_state/optimizer_param_groups.json b/014000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..f39399034560af1d6d8d513c60d398b46df1ba11 --- /dev/null +++ b/014000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.552620875987311e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/014000/training_state/optimizer_state.safetensors b/014000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5df211995e11bba318cdf70c1edebe8bbd896e61 --- /dev/null +++ b/014000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ec531cf0c0ae1935ae548f1bcb4adead2c03cb95b2e4f8ef9f5598c0b2b970a0 +size 2133039080 diff --git a/014000/training_state/rng_state.safetensors b/014000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3642d441309791b49ad1b2abb39ca09d31896f06 --- /dev/null +++ b/014000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fcc6a4df78467cd4d1f59c6489d7570f71ec2e0f7976a6b854da140002083a75 +size 15708 diff --git a/014000/training_state/scheduler_state.json b/014000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..cfba715b297f37ecb9e7e927238ae6fff3bbe2ec --- /dev/null +++ b/014000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 14000, + "_step_count": 14001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.552620875987311e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/014000/training_state/training_step.json b/014000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..560b82990f5c806d547ed82a5432ba60aec285e1 --- /dev/null +++ b/014000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 14000 +} \ No newline at end of file diff --git a/014500/pretrained_model/config.json b/014500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/014500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/014500/pretrained_model/model.safetensors b/014500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..890d7ce69fb7755be3fa092081b884f12d04511d --- /dev/null +++ b/014500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6d032f04f6ac6085f90ad1b1878bd34ed33a01a2be61f3a657ff53744883cdd4 +size 1066517752 diff --git a/014500/pretrained_model/train_config.json b/014500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/014500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/014500/training_state/optimizer_param_groups.json b/014500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..98f095f4e97e280905630fab4097953d8e70604e --- /dev/null +++ b/014500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.51941900244412e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/014500/training_state/optimizer_state.safetensors b/014500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8863db34bb8f0a077acf914d78ad92f1e12a2f80 --- /dev/null +++ b/014500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:215956c1f796ff6b91591bc068ba01770603fbb035fd490a51c6aff64379bc40 +size 2133039080 diff --git a/014500/training_state/rng_state.safetensors b/014500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ed87be73970e793dbcff934e8d374bad77eb6121 --- /dev/null +++ b/014500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:83b48b82ecccae9e365c930eb06979b1869932f413c6363abca83f710c7d08d6 +size 15708 diff --git a/014500/training_state/scheduler_state.json b/014500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..8976354c1da8360be8aef8ffc179d91e1f06f072 --- /dev/null +++ b/014500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 14500, + "_step_count": 14501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.51941900244412e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/014500/training_state/training_step.json b/014500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..ce647c73044ce1a2aa2cc2aaaab26913b3b11ef6 --- /dev/null +++ b/014500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 14500 +} \ No newline at end of file diff --git a/015000/pretrained_model/config.json b/015000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/015000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/015000/pretrained_model/model.safetensors b/015000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fd7d5ce3a70507040d42168aebde08fc00e31773 --- /dev/null +++ b/015000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:957e59fc493af48154ea6b36adf84c45828c85fbe729c2f5a5395ca57de9f54d +size 1066517752 diff --git a/015000/pretrained_model/train_config.json b/015000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/015000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/015000/training_state/optimizer_param_groups.json b/015000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..7876f91897760635b2af670d32f81314ccf1a633 --- /dev/null +++ b/015000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.485090794937319e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/015000/training_state/optimizer_state.safetensors b/015000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..aa518776076f995bc6a3531b3a4fd398e615707b --- /dev/null +++ b/015000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c6b028d0af97b99307070e297960282b4599fd41ae0c75b9f8be2b52e66429c1 +size 2133039080 diff --git a/015000/training_state/rng_state.safetensors b/015000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..53b4b2200cab0f300eb07f4cbba67cb2d14ff92f --- /dev/null +++ b/015000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6d08076bf59981dcc56c6070c5dc0f24afd0a54647518e515befdf71475b819f +size 15708 diff --git a/015000/training_state/scheduler_state.json b/015000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..ac6c728f7025b433fd73558eecdec5603f6d5d61 --- /dev/null +++ b/015000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 15000, + "_step_count": 15001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.485090794937319e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/015000/training_state/training_step.json b/015000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..5cec056c8ba7f3c2e865a0f73ae59975a0503067 --- /dev/null +++ b/015000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 15000 +} \ No newline at end of file diff --git a/015500/pretrained_model/config.json b/015500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/015500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/015500/pretrained_model/model.safetensors b/015500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e7f6b930a01309ca9de047961a91b9e87350b426 --- /dev/null +++ b/015500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2e772e09f3640abbe1c609fdb65acf7a7c9646c90c03eb098275a94f76566b51 +size 1066517752 diff --git a/015500/pretrained_model/train_config.json b/015500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/015500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/015500/training_state/optimizer_param_groups.json b/015500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..71e1c3c3f0f3a366d161bf643274a61e29ac9dcf --- /dev/null +++ b/015500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.449644808775902e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/015500/training_state/optimizer_state.safetensors b/015500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..215db22cefe083314f4859958e7a6a46a2168363 --- /dev/null +++ b/015500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:769e366ee4be7c59d7405131b8423e192f01dbdf7992d63cc1973b28b79e613d +size 2133039080 diff --git a/015500/training_state/rng_state.safetensors b/015500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..055a9680742c3514911136ace2690f74bd47282a --- /dev/null +++ b/015500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:968d04fb370ae31443654e33d87a91d011f337ae3ceb2f3708d0dd3f381614c3 +size 15708 diff --git a/015500/training_state/scheduler_state.json b/015500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..fd4c8affdc670326191e6ee0f1376fcc1fe59026 --- /dev/null +++ b/015500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 15500, + "_step_count": 15501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.449644808775902e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/015500/training_state/training_step.json b/015500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..ec37c3b14625d15d017edc273a747a2941dc48ed --- /dev/null +++ b/015500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 15500 +} \ No newline at end of file diff --git a/016000/pretrained_model/config.json b/016000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/016000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/016000/pretrained_model/model.safetensors b/016000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..61b8fd7f3d51794e4d3210e5fa2d089004129cab --- /dev/null +++ b/016000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca203c10b3cdb0b3d09761863bf04d986692aaff7c336d1db570f3e48e46b3d5 +size 1066517752 diff --git a/016000/pretrained_model/train_config.json b/016000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/016000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/016000/training_state/optimizer_param_groups.json b/016000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..f2b2a0734824f87f07ae77d6d8a3155eab39c5ea --- /dev/null +++ b/016000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.413089877842736e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/016000/training_state/optimizer_state.safetensors b/016000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0c1df8c6396690b2d783a2e1526124e356192f68 --- /dev/null +++ b/016000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:94b35f66f5cc890e70ca9c3beff170561ec782ed53dba9f198642745db72e00d +size 2133039080 diff --git a/016000/training_state/rng_state.safetensors b/016000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..82eefff043fb2f67a5b7df04e97c7c7af1e8e215 --- /dev/null +++ b/016000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:be2d58ff1595931882337f96996bdbef6e2e7322415b46fdbe914879ecd847ef +size 15708 diff --git a/016000/training_state/scheduler_state.json b/016000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3e295cd35e3ef093226d1d955630518c930e1966 --- /dev/null +++ b/016000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 16000, + "_step_count": 16001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.413089877842736e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/016000/training_state/training_step.json b/016000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..798020d90d4c39414dfab88708ecefd52f2e9ac5 --- /dev/null +++ b/016000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 16000 +} \ No newline at end of file diff --git a/016500/pretrained_model/config.json b/016500/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/016500/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/016500/pretrained_model/model.safetensors b/016500/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a22f34305d86265cd595bb440dc9f16de4c5c75a --- /dev/null +++ b/016500/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d3b9aee59345bea9d093869488f750ce9a18aaf3a7fa233e1f95dd9ee60994cd +size 1066517752 diff --git a/016500/pretrained_model/train_config.json b/016500/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/016500/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/016500/training_state/optimizer_param_groups.json b/016500/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..c5280339385d22e72401bd39889c3f2f7ad34c9f --- /dev/null +++ b/016500/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.375435112392969e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/016500/training_state/optimizer_state.safetensors b/016500/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3f1d0b61022acfbd4fa68d1979434e965604c4fa --- /dev/null +++ b/016500/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:71990517cad7b3f09ab07f1e19bc29b1ce5eee323763d1946333f861a7b31d1b +size 2133039080 diff --git a/016500/training_state/rng_state.safetensors b/016500/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d237ae94e9bcabaca925f728e39dbd7980a9f3a6 --- /dev/null +++ b/016500/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:17a317dfade39d117f1709da79abd8b631e930c6576bfda123d28c5064358b06 +size 15708 diff --git a/016500/training_state/scheduler_state.json b/016500/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..15cbe32639c63ffe5c1c0dd9daecaa1101debc9d --- /dev/null +++ b/016500/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 16500, + "_step_count": 16501, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.375435112392969e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/016500/training_state/training_step.json b/016500/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..ed476b8c5e25fc727c214da28ce2a9671d6b073b --- /dev/null +++ b/016500/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 16500 +} \ No newline at end of file diff --git a/017000/pretrained_model/config.json b/017000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e6a6646a622f6bf86ae4104d21fb0dbf333dc1c4 --- /dev/null +++ b/017000/pretrained_model/config.json @@ -0,0 +1,84 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 +} \ No newline at end of file diff --git a/017000/pretrained_model/model.safetensors b/017000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d817ee0b52715076e8a2f302623256bab1b14701 --- /dev/null +++ b/017000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ccc82a25f4d6e760b5d656cb44bc44a741382dc9a85336b1e61660c373b04c6e +size 1066517752 diff --git a/017000/pretrained_model/train_config.json b/017000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..3343227d38f8c4bb6fb7bc049f707757f5d610c5 --- /dev/null +++ b/017000/pretrained_model/train_config.json @@ -0,0 +1,194 @@ +{ + "dataset": { + "repo_id": "so100_Dataset_1000_V0.1", + "root": "so100_Dataset_1000_V0.1", + "episodes": null, + "image_transforms": { + "enable": false, + "max_num_transforms": 3, + "random_order": false, + "tfs": { + "brightness": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "brightness": [ + 0.8, + 1.2 + ] + } + }, + "contrast": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "contrast": [ + 0.8, + 1.2 + ] + } + }, + "saturation": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "saturation": [ + 0.5, + 1.5 + ] + } + }, + "hue": { + "weight": 1.0, + "type": "ColorJitter", + "kwargs": { + "hue": [ + -0.05, + 0.05 + ] + } + }, + "sharpness": { + "weight": 1.0, + "type": "SharpnessJitter", + "kwargs": { + "sharpness": [ + 0.5, + 1.5 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec" + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.Camera_1": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + }, + "observation.images.Camera_2": { + "type": "VISUAL", + "shape": [ + 3, + 480, + 640 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "horizon": 16, + "n_action_steps": 8, + "drop_n_last_frames": 7, + "vision_backbone": "resnet18", + "crop_shape": [ + 84, + 84 + ], + "crop_is_random": true, + "pretrained_backbone_weights": null, + "use_group_norm": true, + "spatial_softmax_num_keypoints": 32, + "use_separate_rgb_encoder_per_camera": false, + "down_dims": [ + 512, + 1024, + 2048 + ], + "kernel_size": 5, + "n_groups": 8, + "diffusion_step_embed_dim": 128, + "use_film_scale_modulation": true, + "noise_scheduler_type": "DDPM", + "num_train_timesteps": 100, + "beta_schedule": "squaredcos_cap_v2", + "beta_start": 0.0001, + "beta_end": 0.02, + "prediction_type": "epsilon", + "clip_sample": true, + "clip_sample_range": 1.0, + "num_inference_steps": null, + "do_mask_loss_for_padding": false, + "optimizer_lr": 0.0001, + "optimizer_betas": [ + 0.95, + 0.999 + ], + "optimizer_eps": 1e-08, + "optimizer_weight_decay": 1e-06, + "scheduler_name": "cosine", + "scheduler_warmup_steps": 500 + }, + "output_dir": "outputs/train/my_diffusion_fast_2", + "job_name": "my_diffusion_training_fast", + "resume": false, + "seed": 1000, + "num_workers": 6, + "batch_size": 256, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "save_checkpoint": true, + "save_freq": 500, + "use_policy_training_preset": true, + "optimizer": { + "type": "adam", + "lr": 0.0001, + "weight_decay": 1e-06, + "grad_clip_norm": 10.0, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08 + }, + "scheduler": { + "type": "diffuser", + "num_warmup_steps": 500, + "name": "cosine" + }, + "eval": { + "n_episodes": 50, + "batch_size": 50, + "use_async_envs": false + }, + "wandb": { + "enable": true, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + } +} \ No newline at end of file diff --git a/017000/training_state/optimizer_param_groups.json b/017000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..f4a0f0c022110028f47f9d82f446872b5f31199d --- /dev/null +++ b/017000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.336689896783573e-05, + "betas": [ + 0.95, + 0.999 + ], + "eps": 1e-08, + "weight_decay": 1e-06, + "amsgrad": false, + "maximize": false, + "foreach": null, + "capturable": false, + "differentiable": false, + "fused": null, + "decoupled_weight_decay": false, + "initial_lr": 0.0001, + "params": [ + 0, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 83, + 84, + 85, + 86, + 87, + 88, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 102, + 103, + 104, + 105, + 106, + 107, + 108, + 109, + 110, + 111, + 112, + 113, + 114, + 115, + 116, + 117, + 118, + 119, + 120, + 121, + 122, + 123, + 124, + 125, + 126, + 127, + 128, + 129, + 130, + 131, + 132, + 133, + 134, + 135, + 136, + 137, + 138, + 139, + 140, + 141, + 142, + 143, + 144, + 145, + 146, + 147, + 148, + 149, + 150, + 151, + 152, + 153, + 154, + 155, + 156, + 157, + 158, + 159, + 160, + 161, + 162, + 163, + 164, + 165, + 166, + 167, + 168, + 169, + 170, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 179, + 180, + 181, + 182, + 183, + 184, + 185, + 186, + 187, + 188, + 189, + 190, + 191, + 192, + 193, + 194, + 195, + 196, + 197, + 198, + 199, + 200, + 201, + 202, + 203, + 204, + 205, + 206, + 207, + 208, + 209, + 210, + 211 + ] + } +] \ No newline at end of file diff --git a/017000/training_state/optimizer_state.safetensors b/017000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..749df701ee5f9d10b0b2c5b686f11a7085e37263 --- /dev/null +++ b/017000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3d2ee7ef56ec7640c0b2d63772f5c680f354c57dfc686323b003f4d6042506e0 +size 2133039080 diff --git a/017000/training_state/rng_state.safetensors b/017000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..db8b52ff2645d96ee73c981d919c2042feccb1a1 --- /dev/null +++ b/017000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6a5051919816a71d747916470e65d8cbf562b7306598bc790095a4c35700e240 +size 15708 diff --git a/017000/training_state/scheduler_state.json b/017000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..6613e9a83cb510dc44def4effce16d623dd10e8a --- /dev/null +++ b/017000/training_state/scheduler_state.json @@ -0,0 +1,15 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 17000, + "_step_count": 17001, + "_is_initial": false, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.336689896783573e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/017000/training_state/training_step.json b/017000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..c397907165bb49af5c8ed9f0802e1ca19fad31fb --- /dev/null +++ b/017000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 17000 +} \ No newline at end of file