diff --git a/005000/pretrained_model/config.json b/005000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/005000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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..dbe81de54643043ccccddffb123bd16c91bd1246 --- /dev/null +++ b/005000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:817237465bf3cb7dbad4693ca15ae09632c23d80bec2c5ede4da5199398dfc1d +size 1066516384 diff --git a/005000/pretrained_model/policy_postprocessor.json b/005000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/005000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/005000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/005000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/005000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/005000/pretrained_model/policy_preprocessor.json b/005000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/005000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/005000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/005000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/005000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/005000/pretrained_model/train_config.json b/005000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/005000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": 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..414efd94b1bc35fecc55b5b2efcf1a30e49f77cf --- /dev/null +++ b/005000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:86d934eb1e17a63c1312acb2f622bfbc9f7531e7ccdad30bccf5ef582a361534 +size 2133039080 diff --git a/005000/training_state/rng_state.safetensors b/005000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0f03fcb5e988f1b7a7d919dbd51b2aa5da40e05a --- /dev/null +++ b/005000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bda4fb89977c2260e22ba78e00ef93598a4c98a5fa113c1f5f1d501b648ce899 +size 15708 diff --git a/005000/training_state/scheduler_state.json b/005000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..f1c3b0b139110f01aa199130d6afbdc24f619069 --- /dev/null +++ b/005000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 5000, + "_step_count": 5001, + "_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/010000/pretrained_model/config.json b/010000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/010000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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..76f8fb535aa6c8ae095b4b0df9f76bf5f3c2137a --- /dev/null +++ b/010000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3c99984cc6cc213b29ea3c1157105c34ec06a71a1bd95e4a94dece7151a7298d +size 1066516384 diff --git a/010000/pretrained_model/policy_postprocessor.json b/010000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/010000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/010000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/010000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/010000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/010000/pretrained_model/policy_preprocessor.json b/010000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/010000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/010000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/010000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/010000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/010000/pretrained_model/train_config.json b/010000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/010000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": 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..edc61726688d0965dd24d57c10a85b7415b00c68 --- /dev/null +++ b/010000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:559a54d8879c3d0502acf41bd92cd16373defefb75a4331f1db1162213cbe2f5 +size 2133039080 diff --git a/010000/training_state/rng_state.safetensors b/010000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..43d0e16a17f1e6be3e779f48bdb1a63b1ab21be2 --- /dev/null +++ b/010000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:171e77093e616576f10466ebbde9f96c5b25a4ac68cd15e3c0ffa6bf75bb267c +size 15708 diff --git a/010000/training_state/scheduler_state.json b/010000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..f80bf0fe89bb9581da28e69a2bba97b9e5cb1a3c --- /dev/null +++ b/010000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 10000, + "_step_count": 10001, + "_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/015000/pretrained_model/config.json b/015000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/015000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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..45c888a46cd6cb58613912121f8c010337780f4e --- /dev/null +++ b/015000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:991fb403d6fc7aa0dc0c638ee853e50764daea5f54dfd24033dc33ee9c1dcbc1 +size 1066516384 diff --git a/015000/pretrained_model/policy_postprocessor.json b/015000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/015000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/015000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/015000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/015000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/015000/pretrained_model/policy_preprocessor.json b/015000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/015000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/015000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/015000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/015000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/015000/pretrained_model/train_config.json b/015000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/015000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": 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..b87f3a0fa8fa0936822168218cf585f82a61952b --- /dev/null +++ b/015000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9cf86c8b757db291af3d0f17894ca30f25d01df5b1c0db0dc24e339adc71186f +size 2133039080 diff --git a/015000/training_state/rng_state.safetensors b/015000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..34fa7d62497ca84a0dcf501852eb3c99f77919a4 --- /dev/null +++ b/015000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a7d75055cf50f4c0e7feef4b160fad30292df64549d0142820763b79e145cd50 +size 15708 diff --git a/015000/training_state/scheduler_state.json b/015000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..5db26b9d170fc46482e984d8bbc3829312f8438f --- /dev/null +++ b/015000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 15000, + "_step_count": 15001, + "_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/020000/pretrained_model/config.json b/020000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/020000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/020000/pretrained_model/model.safetensors b/020000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..49365238c78af473a494be57a242214a24b9d046 --- /dev/null +++ b/020000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8a50eab1e5670691f8adf414715cc214cd1587d2600bf7c40aaa8879462644a1 +size 1066516384 diff --git a/020000/pretrained_model/policy_postprocessor.json b/020000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/020000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/020000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/020000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/020000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/020000/pretrained_model/policy_preprocessor.json b/020000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/020000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/020000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/020000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/020000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/020000/pretrained_model/train_config.json b/020000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/020000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/020000/training_state/optimizer_param_groups.json b/020000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..bf39c5c897a050ff8265bb1308f13cf05d6399da --- /dev/null +++ b/020000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.081878607071996e-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/020000/training_state/optimizer_state.safetensors b/020000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5db8b6c6a7c56bb1e32a6710ac35a6fb96e91473 --- /dev/null +++ b/020000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:34cc486176d93239897ab9f02f2ee187b8155eee63907715fcdc1b2781ad6d12 +size 2133039080 diff --git a/020000/training_state/rng_state.safetensors b/020000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5088eeb4e0130e0b4be6a03f20e75b854f96c6d5 --- /dev/null +++ b/020000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aab2156407d06e57f26325af5c4e8b9b163e3129b1364c3a19a5436fbe34f261 +size 15708 diff --git a/020000/training_state/scheduler_state.json b/020000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..7079c5eb5d1f64d9d6c22b336677c1354414261d --- /dev/null +++ b/020000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 20000, + "_step_count": 20001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.081878607071996e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/020000/training_state/training_step.json b/020000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..dc9bb47026c5d5237ca6fc5dbff6020dd122ea05 --- /dev/null +++ b/020000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 20000 +} \ No newline at end of file diff --git a/025000/pretrained_model/config.json b/025000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/025000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/025000/pretrained_model/model.safetensors b/025000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9afb0c0b63d57af8b49edb8e3abc55ead99ddada --- /dev/null +++ b/025000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c71e7e00397f34a9bdf1076c98751a2e1d24bb0568ff9b93a5f537c0477e206e +size 1066516384 diff --git a/025000/pretrained_model/policy_postprocessor.json b/025000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/025000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/025000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/025000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/025000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/025000/pretrained_model/policy_preprocessor.json b/025000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/025000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/025000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/025000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/025000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/025000/pretrained_model/train_config.json b/025000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/025000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/025000/training_state/optimizer_param_groups.json b/025000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..3a7924bcaa6df0558822a2bf3b46c1ae40225ece --- /dev/null +++ b/025000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 8.577146442236857e-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/025000/training_state/optimizer_state.safetensors b/025000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4fee55a42cc3c8e23fccf9578b6db6d097f38ceb --- /dev/null +++ b/025000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e773acac37090486e9889b0419d9c9de994b95887f6cc3ca5be90708c0e34234 +size 2133039080 diff --git a/025000/training_state/rng_state.safetensors b/025000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cadf87fc5eb25082b769535ec8d9139b1c113558 --- /dev/null +++ b/025000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:21a3b189955b57739fb1d6b57ddf678b1b9295249c030fa60ea60f9ae42c8920 +size 15708 diff --git a/025000/training_state/scheduler_state.json b/025000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..f6690f6c7c24f86903385ccf2455bc34bf3489b6 --- /dev/null +++ b/025000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 25000, + "_step_count": 25001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 8.577146442236857e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/025000/training_state/training_step.json b/025000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..ac6903ae05e7bbaec1e2da710fdbe91d8dc3f37d --- /dev/null +++ b/025000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 25000 +} \ No newline at end of file diff --git a/030000/pretrained_model/config.json b/030000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/030000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/030000/pretrained_model/model.safetensors b/030000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..63e5817b6b17b2f2ff3710dc4cd3f05771a3ba92 --- /dev/null +++ b/030000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9d1ca7129a691b8d6c838e6f4425ec19ce1c50694c051f392714cba0a20de9f0 +size 1066516384 diff --git a/030000/pretrained_model/policy_postprocessor.json b/030000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/030000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/030000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/030000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/030000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/030000/pretrained_model/policy_preprocessor.json b/030000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/030000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/030000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/030000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/030000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/030000/pretrained_model/train_config.json b/030000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/030000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/030000/training_state/optimizer_param_groups.json b/030000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..abce58b2a5bcc2a7a13b9ab9a4b68548873cae50 --- /dev/null +++ b/030000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 7.983447441444281e-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/030000/training_state/optimizer_state.safetensors b/030000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bff30087998a77de6ad95f78b486c1874fc77a52 --- /dev/null +++ b/030000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b0b9bed9141ba4503a2f937dde615172ae7d69c7d8a8c15076f46fda52436124 +size 2133039080 diff --git a/030000/training_state/rng_state.safetensors b/030000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..97ad787ba2760c43b872ac1e40d1a3fb55b83879 --- /dev/null +++ b/030000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:68180ead099373c1e869646c2ac3073def7a324868aae2c6d610988458f2d28e +size 15708 diff --git a/030000/training_state/scheduler_state.json b/030000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..b8a69fc2e96d19197fe551b89febf6089816bcbd --- /dev/null +++ b/030000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 30000, + "_step_count": 30001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 7.983447441444281e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/030000/training_state/training_step.json b/030000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..f4945f660f45b332883dccfccf18d8b8815d916a --- /dev/null +++ b/030000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 30000 +} \ No newline at end of file diff --git a/035000/pretrained_model/config.json b/035000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/035000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/035000/pretrained_model/model.safetensors b/035000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c5a06b83ab0eb93af60e8f8fae964e0852c4370f --- /dev/null +++ b/035000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8dae2191afdb7a2d8ce1ef084f785d2bd51b18f22cbe13908415bbc78ba57e35 +size 1066516384 diff --git a/035000/pretrained_model/policy_postprocessor.json b/035000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/035000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/035000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/035000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/035000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/035000/pretrained_model/policy_preprocessor.json b/035000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/035000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/035000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/035000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/035000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/035000/pretrained_model/train_config.json b/035000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/035000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/035000/training_state/optimizer_param_groups.json b/035000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..bb15a39908f51722f53c672f34c370ab164344a8 --- /dev/null +++ b/035000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 7.31554743060174e-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/035000/training_state/optimizer_state.safetensors b/035000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c53d429d050515ab082f3c7523367f725ab0e141 --- /dev/null +++ b/035000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:01b537f5f772a650fd30d58db13c670cd40142734badc98da87cae6119bd4614 +size 2133039080 diff --git a/035000/training_state/rng_state.safetensors b/035000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1cd7832261f7d4d76ff32c61bfdd74d5018848fb --- /dev/null +++ b/035000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6a085f9db8737ca1bab1ad240f708aa741701bd700dc9274c05c5ecb468e15e5 +size 15708 diff --git a/035000/training_state/scheduler_state.json b/035000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..e30ec9e17a9efa494b332a7ea3be0e08b2b66876 --- /dev/null +++ b/035000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 35000, + "_step_count": 35001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 7.31554743060174e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/035000/training_state/training_step.json b/035000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..60080824ceb9805d29c885917dc77e0277e52eb2 --- /dev/null +++ b/035000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 35000 +} \ No newline at end of file diff --git a/040000/pretrained_model/config.json b/040000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/040000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/040000/pretrained_model/model.safetensors b/040000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..638e2089f1e4c73b32457af69af5184b101741f4 --- /dev/null +++ b/040000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5273df4da902dcd0a5f780aa721dc27ad5f1327805790586782a22787bddc5c3 +size 1066516384 diff --git a/040000/pretrained_model/policy_postprocessor.json b/040000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/040000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/040000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/040000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/040000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/040000/pretrained_model/policy_preprocessor.json b/040000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/040000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/040000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/040000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/040000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/040000/pretrained_model/train_config.json b/040000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/040000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/040000/training_state/optimizer_param_groups.json b/040000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..da769670afd79c7409bd31e4f84f2a93bad74c99 --- /dev/null +++ b/040000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 6.590057681196191e-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/040000/training_state/optimizer_state.safetensors b/040000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bf4ec98749d4b1ea47728023babba5a782778a66 --- /dev/null +++ b/040000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0a208a1a63add453cb43cf90590cf1203e94905f72965db6d4c5c5b966a0494a +size 2133039080 diff --git a/040000/training_state/rng_state.safetensors b/040000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cde8516d2490cc41f1289c6877bb740847958bad --- /dev/null +++ b/040000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:af672952424d6702aa2acc25a663f3c0e99d552bb2f77b6afbeae25a1025f7f6 +size 15708 diff --git a/040000/training_state/scheduler_state.json b/040000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3d5ef807651649e49fd130ba01189f1e3e79f035 --- /dev/null +++ b/040000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 40000, + "_step_count": 40001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 6.590057681196191e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/040000/training_state/training_step.json b/040000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..fe40d8ff4312c7e5a8fba9bcc932a43a1384ba77 --- /dev/null +++ b/040000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 40000 +} \ No newline at end of file diff --git a/045000/pretrained_model/config.json b/045000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/045000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/045000/pretrained_model/model.safetensors b/045000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f389cdf053f7a74636bcddc28cc228f0d1f072a9 --- /dev/null +++ b/045000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:20c6ddcd39bc2a5a1f713258e84f921696939d322ef92d18a68e1deb84379217 +size 1066516384 diff --git a/045000/pretrained_model/policy_postprocessor.json b/045000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/045000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/045000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/045000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/045000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/045000/pretrained_model/policy_preprocessor.json b/045000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/045000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/045000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/045000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/045000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/045000/pretrained_model/train_config.json b/045000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/045000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/045000/training_state/optimizer_param_groups.json b/045000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..24901a41a524e7f063a10da848d9641e55acefda --- /dev/null +++ b/045000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 5.8250217730939973e-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/045000/training_state/optimizer_state.safetensors b/045000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ad799b7da9b4691fa8b0be81b109a5029b2395a0 --- /dev/null +++ b/045000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:55140525a2e01a4f182e7b87d0a9ffb06adf68a3d18c2883a0ece41c17f9e394 +size 2133039080 diff --git a/045000/training_state/rng_state.safetensors b/045000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c4f9dd3ee94a6e779dd8eaa13b8b771a65af7ad7 --- /dev/null +++ b/045000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c1b36b2a98a373be0b6ebb4b2a89924b56825f36837f99c3949183e74714a9bf +size 15708 diff --git a/045000/training_state/scheduler_state.json b/045000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..fbdda42900826f76ed1d6d15f0e17f5c0c2ddc25 --- /dev/null +++ b/045000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 45000, + "_step_count": 45001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 5.8250217730939973e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/045000/training_state/training_step.json b/045000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..472b301a48ca897c9f2b7da46adc0312cf9a91d9 --- /dev/null +++ b/045000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 45000 +} \ No newline at end of file diff --git a/050000/pretrained_model/config.json b/050000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/050000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/050000/pretrained_model/model.safetensors b/050000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ae9047f1267e4eca965ee97d8c3596b1c3cb19f2 --- /dev/null +++ b/050000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a9d2aa18be5542ffe46e03b2c0e45e250a10d7b3dcdcbaa139e2667ba5bd0a9 +size 1066516384 diff --git a/050000/pretrained_model/policy_postprocessor.json b/050000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/050000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/050000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/050000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/050000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/050000/pretrained_model/policy_preprocessor.json b/050000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/050000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/050000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/050000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/050000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/050000/pretrained_model/train_config.json b/050000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/050000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/050000/training_state/optimizer_param_groups.json b/050000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..7da1770ac06dfc0528a5a8df5006be72e0f2b226 --- /dev/null +++ b/050000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 5.039466834548568e-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/050000/training_state/optimizer_state.safetensors b/050000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5616270d1df42464f9e3ce7c8bdcc25205aeea6b --- /dev/null +++ b/050000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9dec4d1666de6fa4c83cf6b222e1ea15b222314f61bc8ae97213c11ad015261a +size 2133039080 diff --git a/050000/training_state/rng_state.safetensors b/050000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..051b6c4c641f245a7b81e4eae755dca1bc60f699 --- /dev/null +++ b/050000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d12980a1de821c8b2a70902b14a445ae770a42fdccaef87a33ba0cfaa1fa330a +size 15708 diff --git a/050000/training_state/scheduler_state.json b/050000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..45384c87f46fc5e39e00b7134a617ce880502e71 --- /dev/null +++ b/050000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 50000, + "_step_count": 50001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 5.039466834548568e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/050000/training_state/training_step.json b/050000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..b3c03563f17d6d98552117ccf00705b693aa3fbd --- /dev/null +++ b/050000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 50000 +} \ No newline at end of file diff --git a/055000/pretrained_model/config.json b/055000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/055000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/055000/pretrained_model/model.safetensors b/055000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6d9a1617d3d688582468cac6eaa26b67d0cf8ffa --- /dev/null +++ b/055000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:58466ac7761c038d0661cde85fc73bd7cb6b6c0549424a92896d82deff1fd058 +size 1066516384 diff --git a/055000/pretrained_model/policy_postprocessor.json b/055000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/055000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/055000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/055000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/055000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/055000/pretrained_model/policy_preprocessor.json b/055000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/055000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/055000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/055000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/055000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/055000/pretrained_model/train_config.json b/055000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/055000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/055000/training_state/optimizer_param_groups.json b/055000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..4e98243db520cb68fa181e81b80b622f0e731837 --- /dev/null +++ b/055000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 4.2529303204786953e-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/055000/training_state/optimizer_state.safetensors b/055000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..43caaf305bb014d500e755b78fb37d99bb722a5a --- /dev/null +++ b/055000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d313310873247ca6933d71d6c72eb1d32cf3bded41d1bd2ed3efb0363d3916f0 +size 2133039080 diff --git a/055000/training_state/rng_state.safetensors b/055000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fbf04425bdf447e9433d99ccc6ad9266e07eac80 --- /dev/null +++ b/055000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:55c4ebf3353d7e1c05da2855d34823cb678cdcc40d9539f4473af75612cc449d +size 15708 diff --git a/055000/training_state/scheduler_state.json b/055000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..2afce6ab503bcd7cf6925dcb232b033a56376dc2 --- /dev/null +++ b/055000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 55000, + "_step_count": 55001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 4.2529303204786953e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/055000/training_state/training_step.json b/055000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..57c2477e950fdfdedf6a402eec3c3f0fe2d039ed --- /dev/null +++ b/055000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 55000 +} \ No newline at end of file diff --git a/060000/pretrained_model/config.json b/060000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/060000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/060000/pretrained_model/model.safetensors b/060000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3c23ae4764fa5232d4b21a54781f606fc631c17b --- /dev/null +++ b/060000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c1560bc17f9c0d0c5c7b5ebd51fbaacc0ea8766a907c2f6f1f2343d0e2ee5eff +size 1066516384 diff --git a/060000/pretrained_model/policy_postprocessor.json b/060000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/060000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/060000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/060000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/060000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/060000/pretrained_model/policy_preprocessor.json b/060000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/060000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/060000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/060000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/060000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/060000/pretrained_model/train_config.json b/060000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/060000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/060000/training_state/optimizer_param_groups.json b/060000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..044c2e63a486a01bdda61b8fdf8b369988a133ac --- /dev/null +++ b/060000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 3.484974098465636e-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/060000/training_state/optimizer_state.safetensors b/060000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..546430eeb466091cc65d1972addfca21cc255e75 --- /dev/null +++ b/060000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7c6d7bfb505efc2f6d65e2730c7fcae3811cade87e3efdc6ab0dad550eb9bd59 +size 2133039080 diff --git a/060000/training_state/rng_state.safetensors b/060000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a3561431eae3b4ddaeb3f60f287754249cc3d9c9 --- /dev/null +++ b/060000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7ada000508f8e08a8540328197b386d781b2f1f79d16e0825c809a4c41ce7a27 +size 15708 diff --git a/060000/training_state/scheduler_state.json b/060000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..668b51a9f5d74f350a4aa0a1b35f5d4f4498a280 --- /dev/null +++ b/060000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 60000, + "_step_count": 60001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 3.484974098465636e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/060000/training_state/training_step.json b/060000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..d360b6037f1c8c2c85d38e951160eb02ace507dd --- /dev/null +++ b/060000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 60000 +} \ No newline at end of file diff --git a/065000/pretrained_model/config.json b/065000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/065000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/065000/pretrained_model/model.safetensors b/065000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..55e8a2a68c5ee027a55bc857c540dfa2ab291eb6 --- /dev/null +++ b/065000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7eec5dc62a5a0959ebde24f3a45b414bb6cfdbc481a73fa194ffae3334678408 +size 1066516384 diff --git a/065000/pretrained_model/policy_postprocessor.json b/065000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/065000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/065000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/065000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/065000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/065000/pretrained_model/policy_preprocessor.json b/065000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/065000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/065000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/065000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/065000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/065000/pretrained_model/train_config.json b/065000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/065000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/065000/training_state/optimizer_param_groups.json b/065000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..379a5effa99a9b00300a71e14dea0a3096837a1d --- /dev/null +++ b/065000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 2.754697927585399e-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/065000/training_state/optimizer_state.safetensors b/065000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a778fe56e8ccb0c818627d3bb3df687d4435c5db --- /dev/null +++ b/065000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f4de80e7cf23de188554db971e566c8f2c4e26a994f0e618e3223527b1f804f8 +size 2133039080 diff --git a/065000/training_state/rng_state.safetensors b/065000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3870290913dd4c02dd5ffac6bbe0094ff896f558 --- /dev/null +++ b/065000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:826a3ea3a3367aebeefb973fc0b977a99549deba536128677eddc199699c49c2 +size 15708 diff --git a/065000/training_state/scheduler_state.json b/065000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..e176e22388fcfa1c44384aac2429743fb879e9d9 --- /dev/null +++ b/065000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 65000, + "_step_count": 65001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 2.754697927585399e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/065000/training_state/training_step.json b/065000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..29c11af0e7f84ef97479db014fb5c2ce76abe095 --- /dev/null +++ b/065000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 65000 +} \ No newline at end of file diff --git a/070000/pretrained_model/config.json b/070000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/070000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/070000/pretrained_model/model.safetensors b/070000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9dea318ce9fabf5e4c36cd1be5100a0f9112540e --- /dev/null +++ b/070000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3233d154a84d4e021c6dab80f5d9a95619e601af69d4f3e0f5f48d7caacb9ef2 +size 1066516384 diff --git a/070000/pretrained_model/policy_postprocessor.json b/070000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/070000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/070000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/070000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/070000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/070000/pretrained_model/policy_preprocessor.json b/070000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/070000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/070000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/070000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/070000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/070000/pretrained_model/train_config.json b/070000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/070000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/070000/training_state/optimizer_param_groups.json b/070000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..ade33faaef851172aa00c798bfe222cfa196ad4c --- /dev/null +++ b/070000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 2.0802644302934683e-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/070000/training_state/optimizer_state.safetensors b/070000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..044c57d1e8488ccb8b4ebe54e17e63b5a6d06dea --- /dev/null +++ b/070000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:105890716da7531ee6d9403db00051bbc89589ec1ae70e5f18ae7a2e871bc790 +size 2133039080 diff --git a/070000/training_state/rng_state.safetensors b/070000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..62f4d2a8bce7d327c83bf07285b9b632956a9796 --- /dev/null +++ b/070000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:69b9ca1c046893ffa7a1b03255eb5dc0f68a5d97657e3f73b65d4233762a80bb +size 15708 diff --git a/070000/training_state/scheduler_state.json b/070000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..610b49fe41ad7377908f3bdf2c19add4dc4ed585 --- /dev/null +++ b/070000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 70000, + "_step_count": 70001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 2.0802644302934683e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/070000/training_state/training_step.json b/070000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..1ec0c1eb0f525e34fa3a57743f750b2b34b5b77b --- /dev/null +++ b/070000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 70000 +} \ No newline at end of file diff --git a/075000/pretrained_model/config.json b/075000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/075000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/075000/pretrained_model/model.safetensors b/075000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..1fd0e206404080737170286a778a4053226f395b --- /dev/null +++ b/075000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1cfae28b2a67e4a2102f33b81bc0b027d446862c732c07376a03954a4d4f8c2e +size 1066516384 diff --git a/075000/pretrained_model/policy_postprocessor.json b/075000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/075000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/075000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/075000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/075000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/075000/pretrained_model/policy_preprocessor.json b/075000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/075000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/075000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/075000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/075000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/075000/pretrained_model/train_config.json b/075000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/075000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/075000/training_state/optimizer_param_groups.json b/075000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..8ba6661c158f562c329a9a3b19689bac2a29d05c --- /dev/null +++ b/075000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 1.4784473717366387e-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/075000/training_state/optimizer_state.safetensors b/075000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c16a80fea9dfc342429df98a3e13fa8b1fc98b1d --- /dev/null +++ b/075000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8452c99cb562029f56ccf3882ca5e0d4789e0fe9025e0ca2039ecb64cec3b873 +size 2133039080 diff --git a/075000/training_state/rng_state.safetensors b/075000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ff6cde67f592e2df1ae185bea52fd92b99e0ee59 --- /dev/null +++ b/075000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dcb6d9d22fd79d83354fcf356c98c5abd17d53333cdd6d91b919ada8738fa9eb +size 15708 diff --git a/075000/training_state/scheduler_state.json b/075000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..4937a8474c4f299d49436e31f559014aadf09363 --- /dev/null +++ b/075000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 75000, + "_step_count": 75001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 1.4784473717366387e-05 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/075000/training_state/training_step.json b/075000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..19674b7211fb0b7620c8755c555877b160e5a745 --- /dev/null +++ b/075000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 75000 +} \ No newline at end of file diff --git a/080000/pretrained_model/config.json b/080000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/080000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/080000/pretrained_model/model.safetensors b/080000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4a2f8853543bb7e0d8759874fe9b00eb05c6b055 --- /dev/null +++ b/080000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2e4b133f16465421ae43ec7dcc5ab0875ce98ec72168998366b54ed1a98a58ef +size 1066516384 diff --git a/080000/pretrained_model/policy_postprocessor.json b/080000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/080000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/080000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/080000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/080000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/080000/pretrained_model/policy_preprocessor.json b/080000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/080000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/080000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/080000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/080000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/080000/pretrained_model/train_config.json b/080000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/080000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/080000/training_state/optimizer_param_groups.json b/080000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..a3e37a0becbe350cfc73f7ac1886b0322872e19b --- /dev/null +++ b/080000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 9.642144811900739e-06, + "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/080000/training_state/optimizer_state.safetensors b/080000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b149948c9a8f196d4ceac54a64abc0c509d2d00d --- /dev/null +++ b/080000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f6543f6b0106baea498ef018e9317d7d27ee4c81aed155d00ec0324118069da2 +size 2133039080 diff --git a/080000/training_state/rng_state.safetensors b/080000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c99d9291cbf8416028420b256d13abf177f08ea9 --- /dev/null +++ b/080000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:63c950a0e40418db23d8829550d3dd09a9af5e0db6d3014c54fcbf4d0239e39e +size 15708 diff --git a/080000/training_state/scheduler_state.json b/080000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..0d7972eada3dacc85eaf6bf64c05d9e14b9cb216 --- /dev/null +++ b/080000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 80000, + "_step_count": 80001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 9.642144811900739e-06 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/080000/training_state/training_step.json b/080000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..384eace4ecc2a6cba352aa7cf27f04405b7319c3 --- /dev/null +++ b/080000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 80000 +} \ No newline at end of file diff --git a/085000/pretrained_model/config.json b/085000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/085000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/085000/pretrained_model/model.safetensors b/085000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..20f8b23e9b9f411de900380244b328a842a7ed31 --- /dev/null +++ b/085000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a69d2c5c4cf1033b57d78b34b7af25c7bcc94079eaef9642a899e33d02d1c579 +size 1066516384 diff --git a/085000/pretrained_model/policy_postprocessor.json b/085000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/085000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/085000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/085000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/085000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/085000/pretrained_model/policy_preprocessor.json b/085000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/085000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/085000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/085000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/085000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/085000/pretrained_model/train_config.json b/085000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/085000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/085000/training_state/optimizer_param_groups.json b/085000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..5572f2cad7c1fa2e72f5d76b232a3f7757ff0b4b --- /dev/null +++ b/085000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 5.5035519122409895e-06, + "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/085000/training_state/optimizer_state.safetensors b/085000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..755da7287399ada1598dbcf82a754094cd6768e2 --- /dev/null +++ b/085000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e2da656105add95dd36d3a308c55721ae5e2612e3d638adf9b063c89187a1a88 +size 2133039080 diff --git a/085000/training_state/rng_state.safetensors b/085000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..0d096a5e9f178bb8e8d199f1dd4b0a110dca71fd --- /dev/null +++ b/085000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e6a891a298b3b97d5ebdd00477a95ce38ead043c93509ef5fb48608ca6ae6a6c +size 15708 diff --git a/085000/training_state/scheduler_state.json b/085000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..0a01bb6a0ddf36ceadc41a150b4ae5f1701a135d --- /dev/null +++ b/085000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 85000, + "_step_count": 85001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 5.5035519122409895e-06 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/085000/training_state/training_step.json b/085000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..bfef6bbed897f0b32861a6990674bf65f9ae03d8 --- /dev/null +++ b/085000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 85000 +} \ No newline at end of file diff --git a/090000/pretrained_model/config.json b/090000/pretrained_model/config.json new file mode 100644 index 0000000000000000000000000000000000000000..da063fed06f25460cbda9415456abc9157f94d06 --- /dev/null +++ b/090000/pretrained_model/config.json @@ -0,0 +1,91 @@ +{ + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/090000/pretrained_model/model.safetensors b/090000/pretrained_model/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bb0e96dacf73af441df153678bcc98d25a1573fe --- /dev/null +++ b/090000/pretrained_model/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c39705e7b1212e4adeeb4680982de03eecfdf58a2f0c789ed3bd93218e95cab7 +size 1066516384 diff --git a/090000/pretrained_model/policy_postprocessor.json b/090000/pretrained_model/policy_postprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..f2c5e6cea7850bb999ab193f96021fa0129ce4db --- /dev/null +++ b/090000/pretrained_model/policy_postprocessor.json @@ -0,0 +1,32 @@ +{ + "name": "policy_postprocessor", + "steps": [ + { + "registry_name": "unnormalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors" + }, + { + "registry_name": "device_processor", + "config": { + "device": "cpu", + "float_dtype": null + } + } + ] +} \ No newline at end of file diff --git a/090000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors b/090000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/090000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/090000/pretrained_model/policy_preprocessor.json b/090000/pretrained_model/policy_preprocessor.json new file mode 100644 index 0000000000000000000000000000000000000000..d29f8373b5bda19d8b1e81578673580ec486534a --- /dev/null +++ b/090000/pretrained_model/policy_preprocessor.json @@ -0,0 +1,64 @@ +{ + "name": "policy_preprocessor", + "steps": [ + { + "registry_name": "rename_observations_processor", + "config": { + "rename_map": {} + } + }, + { + "registry_name": "to_batch_processor", + "config": {} + }, + { + "registry_name": "device_processor", + "config": { + "device": "cuda", + "float_dtype": null + } + }, + { + "registry_name": "normalizer_processor", + "config": { + "eps": 1e-08, + "features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "norm_map": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + } + }, + "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors" + } + ] +} \ No newline at end of file diff --git a/090000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors b/090000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3dceee44d9554b6680b8d0ddbdbd5a154cd402c7 --- /dev/null +++ b/090000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c1a60199fb565914d8569b5357388e11461bda530fd0a854b1b6908fa9a60d3 +size 7584 diff --git a/090000/pretrained_model/train_config.json b/090000/pretrained_model/train_config.json new file mode 100644 index 0000000000000000000000000000000000000000..92cf6453e898631c4b6e2f1f2b4eeefa25dca2aa --- /dev/null +++ b/090000/pretrained_model/train_config.json @@ -0,0 +1,225 @@ +{ + "dataset": { + "repo_id": "TANAY779/record_actV2_200", + "root": null, + "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 + ] + } + }, + "affine": { + "weight": 1.0, + "type": "RandomAffine", + "kwargs": { + "degrees": [ + -5.0, + 5.0 + ], + "translate": [ + 0.05, + 0.05 + ] + } + } + } + }, + "revision": null, + "use_imagenet_stats": true, + "video_backend": "torchcodec", + "streaming": false + }, + "env": null, + "policy": { + "type": "diffusion", + "n_obs_steps": 2, + "input_features": { + "observation.state": { + "type": "STATE", + "shape": [ + 6 + ] + }, + "observation.images.overhead": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + }, + "observation.images.wrist": { + "type": "VISUAL", + "shape": [ + 3, + 960, + 1280 + ] + } + }, + "output_features": { + "action": { + "type": "ACTION", + "shape": [ + 6 + ] + } + }, + "device": "cuda", + "use_amp": true, + "use_peft": false, + "push_to_hub": true, + "repo_id": "TANAY779/diffusion_so101_v1", + "private": null, + "tags": null, + "license": null, + "pretrained_path": null, + "horizon": 32, + "n_action_steps": 8, + "normalization_mapping": { + "VISUAL": "MEAN_STD", + "STATE": "MIN_MAX", + "ACTION": "MIN_MAX" + }, + "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": 50, + "beta_schedule": "linear", + "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/diffusion_so101_v1", + "job_name": "diffusion", + "resume": false, + "seed": 1000, + "num_workers": 2, + "batch_size": 8, + "steps": 100000, + "eval_freq": 20000, + "log_freq": 200, + "tolerance_s": 0.0001, + "save_checkpoint": true, + "save_freq": 5000, + "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": false, + "disable_artifact": false, + "project": "lerobot", + "entity": null, + "notes": null, + "run_id": null, + "mode": null + }, + "peft": null, + "use_rabc": false, + "rabc_progress_path": null, + "rabc_kappa": 0.01, + "rabc_epsilon": 1e-06, + "rabc_head_mode": "sparse", + "rename_map": {}, + "checkpoint_path": null +} \ No newline at end of file diff --git a/090000/training_state/optimizer_param_groups.json b/090000/training_state/optimizer_param_groups.json new file mode 100644 index 0000000000000000000000000000000000000000..3ee70da36fd40f402a033a3015e6b6282bc8491a --- /dev/null +++ b/090000/training_state/optimizer_param_groups.json @@ -0,0 +1,233 @@ +[ + { + "lr": 2.4716255306108605e-06, + "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/090000/training_state/optimizer_state.safetensors b/090000/training_state/optimizer_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2b3723dbaa448c11af9a4b14ea6e8db474131847 --- /dev/null +++ b/090000/training_state/optimizer_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:55fdfd5ff977fa2f8f109796e0fc2d8b81cd724a81e165245bf6161b9935b7d4 +size 2133039080 diff --git a/090000/training_state/rng_state.safetensors b/090000/training_state/rng_state.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..48429416fb5d522097052cdf8633ba3054926699 --- /dev/null +++ b/090000/training_state/rng_state.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c561365cb7cfd10c38e7cf34450700969c308f50c41b973f3d7bd393f9711568 +size 15708 diff --git a/090000/training_state/scheduler_state.json b/090000/training_state/scheduler_state.json new file mode 100644 index 0000000000000000000000000000000000000000..d423c04a1b76a8738eba07c5c97cba6bc20c7cf4 --- /dev/null +++ b/090000/training_state/scheduler_state.json @@ -0,0 +1,14 @@ +{ + "base_lrs": [ + 0.0001 + ], + "last_epoch": 90000, + "_step_count": 90001, + "_get_lr_called_within_step": false, + "_last_lr": [ + 2.4716255306108605e-06 + ], + "lr_lambdas": [ + null + ] +} \ No newline at end of file diff --git a/090000/training_state/training_step.json b/090000/training_state/training_step.json new file mode 100644 index 0000000000000000000000000000000000000000..cec965b83b7f817ac99559792f84cbb69a721bfd --- /dev/null +++ b/090000/training_state/training_step.json @@ -0,0 +1,3 @@ +{ + "step": 90000 +} \ No newline at end of file