| { | |
| "norm_stats": { | |
| "gen72_grasp_stacking_baskets": { | |
| "action": { | |
| "mean": [ | |
| 71.42466735839844, | |
| 84.38833618164062, | |
| -82.9383316040039, | |
| -85.52547454833984, | |
| -4.3938679695129395, | |
| 19.08216094970703, | |
| 0.308868408203125, | |
| 67.15239715576172, | |
| -80.31741333007812, | |
| 84.10924530029297, | |
| 86.23428344726562, | |
| -86.4471206665039, | |
| 3.9559240341186523, | |
| 9.594415664672852, | |
| 4.726510524749756, | |
| 82.21428680419922 | |
| ], | |
| "std": [ | |
| 25.26982879638672, | |
| 15.808586120605469, | |
| 20.259897232055664, | |
| 21.52882957458496, | |
| 20.819337844848633, | |
| 20.531246185302734, | |
| 26.965940475463867, | |
| 45.217315673828125, | |
| 27.104999542236328, | |
| 12.850805282592773, | |
| 14.046908378601074, | |
| 17.79497718811035, | |
| 13.857259750366211, | |
| 17.137197494506836, | |
| 17.71614646911621, | |
| 36.83033752441406 | |
| ], | |
| "max": [ | |
| 124.62999725341797, | |
| 102.0, | |
| 12.920000076293945, | |
| 52.0, | |
| 119.18000030517578, | |
| 89.77999877929688, | |
| 169.0, | |
| 100.0, | |
| 3.9600000381469727, | |
| 102.0, | |
| 123.30999755859375, | |
| 52.0, | |
| 73.56500244140625, | |
| 66.08999633789062, | |
| 86.5250015258789, | |
| 100.0 | |
| ], | |
| "min": [ | |
| 5.84499979019165, | |
| -38.275001525878906, | |
| -169.0, | |
| -97.47000122070312, | |
| -113.19999694824219, | |
| -65.30000305175781, | |
| -139.75, | |
| 0.0, | |
| -121.68499755859375, | |
| 5.84499979019165, | |
| -8.654999732971191, | |
| -98.16999816894531, | |
| -55.459999084472656, | |
| -78.44499969482422, | |
| -92.29000091552734, | |
| 0.0 | |
| ], | |
| "q01": [ | |
| 11.34000015258789, | |
| 35.595001220703125, | |
| -127.60875129699707, | |
| -96.5, | |
| -68.33499908447266, | |
| -37.86875057220459, | |
| -82.83499908447266, | |
| 0.0, | |
| -109.64374732971191, | |
| 33.56999969482422, | |
| 33.582499504089355, | |
| -96.94000244140625, | |
| -22.323750495910645, | |
| -32.78499984741211, | |
| -46.05500030517578, | |
| 0.0 | |
| ], | |
| "q99": [ | |
| 103.84500122070312, | |
| 102.0, | |
| -23.510000228881836, | |
| 24.047500133514404, | |
| 58.91999912261963, | |
| 69.83000183105469, | |
| 63.060001373291016, | |
| 100.0, | |
| -10.866249561309814, | |
| 102.0, | |
| 108.41500091552734, | |
| 2.14000004529953, | |
| 54.625, | |
| 51.900001525878906, | |
| 61.040000915527344, | |
| 100.0 | |
| ], | |
| "mask": [ | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true, | |
| true | |
| ] | |
| }, | |
| "proprio": { | |
| "mean": [ | |
| 71.39665222167969, | |
| 84.42801666259766, | |
| -82.94181823730469, | |
| -85.53479766845703, | |
| -4.377756118774414, | |
| 19.067413330078125, | |
| 0.31575945019721985, | |
| 67.15303802490234, | |
| -80.27192687988281, | |
| 84.11940002441406, | |
| 86.19857788085938, | |
| -86.43647766113281, | |
| 3.944101572036743, | |
| 9.555686950683594, | |
| 4.720884323120117, | |
| 82.21434783935547 | |
| ], | |
| "std": [ | |
| 24.659215927124023, | |
| 15.453649520874023, | |
| 19.847105026245117, | |
| 21.23726463317871, | |
| 20.293472290039062, | |
| 19.495216369628906, | |
| 26.316822052001953, | |
| 45.21790313720703, | |
| 26.583589553833008, | |
| 12.41045093536377, | |
| 13.648248672485352, | |
| 17.389333724975586, | |
| 13.373165130615234, | |
| 16.093711853027344, | |
| 17.135786056518555, | |
| 36.83036804199219 | |
| ], | |
| "max": [ | |
| 124.08000183105469, | |
| 102.05000305175781, | |
| 10.850000381469727, | |
| 52.0099983215332, | |
| 109.37000274658203, | |
| 87.58000183105469, | |
| 168.66000366210938, | |
| 100.0, | |
| 3.8399999141693115, | |
| 102.01000213623047, | |
| 119.8499984741211, | |
| 41.31999969482422, | |
| 68.77999877929688, | |
| 65.19999694824219, | |
| 82.41999816894531, | |
| 100.0 | |
| ], | |
| "min": [ | |
| 6.940000057220459, | |
| -35.029998779296875, | |
| -169.00999450683594, | |
| -97.37999725341797, | |
| -111.45999908447266, | |
| -62.58000183105469, | |
| -136.9199981689453, | |
| 0.0, | |
| -120.04000091552734, | |
| 10.140000343322754, | |
| -3.2699999809265137, | |
| -97.66000366210938, | |
| -54.2599983215332, | |
| -64.5, | |
| -89.81999969482422, | |
| 0.0 | |
| ], | |
| "q01": [ | |
| 11.512500286102295, | |
| 36.560001373291016, | |
| -126.5099983215332, | |
| -96.41000366210938, | |
| -66.58749771118164, | |
| -35.62750053405762, | |
| -81.84500122070312, | |
| 0.0, | |
| -109.31999969482422, | |
| 37.10499858856201, | |
| 34.77500057220459, | |
| -96.91999816894531, | |
| -21.260000228881836, | |
| -29.440000534057617, | |
| -45.08750057220459, | |
| 0.0 | |
| ], | |
| "q99": [ | |
| 102.95499992370605, | |
| 102.01000213623047, | |
| -25.047500610351562, | |
| 23.78499937057495, | |
| 56.52000045776367, | |
| 67.02999877929688, | |
| 61.95750045776367, | |
| 100.0, | |
| -11.315000295639038, | |
| 102.01000213623047, | |
| 107.51000213623047, | |
| 0.6450000107288361, | |
| 53.470001220703125, | |
| 49.709999084472656, | |
| 59.13999938964844, | |
| 100.0 | |
| ] | |
| }, | |
| "num_transitions": 74826, | |
| "num_trajectories": 200 | |
| } | |
| }, | |
| "n_action_bins": 256, | |
| "vision_backbone_id": "dinosiglip-vit-so-224px", | |
| "llm_backbone_id": "llama2-7b-pure", | |
| "arch_specifier": "no-align+fused-gelu-mlp", | |
| "output_projector_states": false, | |
| "use_fused_vision_backbone": true, | |
| "timm_model_ids": [ | |
| "vit_large_patch14_reg4_dinov2.lvd142m", | |
| "vit_so400m_patch14_siglip_224" | |
| ], | |
| "timm_override_act_layers": [ | |
| null, | |
| null | |
| ], | |
| "image_sizes": [ | |
| 224, | |
| 224 | |
| ], | |
| "image_resize_strategy": "resize-naive", | |
| "hf_llm_id": "meta-llama/Llama-2-7b-hf", | |
| "llm_max_length": 2048, | |
| "pad_token_id": 32000, | |
| "pad_to_multiple_of": 64, | |
| "text_config": { | |
| "vocab_size": 32064, | |
| "max_position_embeddings": 2048, | |
| "hidden_size": 4096, | |
| "intermediate_size": 11008, | |
| "num_hidden_layers": 32, | |
| "num_attention_heads": 32, | |
| "num_key_value_heads": 32, | |
| "hidden_act": "silu", | |
| "initializer_range": 0.02, | |
| "rms_norm_eps": 1e-06, | |
| "pretraining_tp": 1, | |
| "use_cache": true, | |
| "rope_theta": 10000.0, | |
| "rope_scaling": null, | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "return_dict": true, | |
| "output_hidden_states": false, | |
| "output_attentions": false, | |
| "torchscript": false, | |
| "torch_dtype": "bfloat16", | |
| "use_bfloat16": false, | |
| "tf_legacy_loss": false, | |
| "pruned_heads": {}, | |
| "tie_word_embeddings": false, | |
| "chunk_size_feed_forward": 0, | |
| "is_encoder_decoder": false, | |
| "is_decoder": false, | |
| "cross_attention_hidden_size": null, | |
| "add_cross_attention": false, | |
| "tie_encoder_decoder": false, | |
| "max_length": 20, | |
| "min_length": 0, | |
| "do_sample": false, | |
| "early_stopping": false, | |
| "num_beams": 1, | |
| "num_beam_groups": 1, | |
| "diversity_penalty": 0.0, | |
| "temperature": 1.0, | |
| "top_k": 50, | |
| "top_p": 1.0, | |
| "typical_p": 1.0, | |
| "repetition_penalty": 1.0, | |
| "length_penalty": 1.0, | |
| "no_repeat_ngram_size": 0, | |
| "encoder_no_repeat_ngram_size": 0, | |
| "bad_words_ids": null, | |
| "num_return_sequences": 1, | |
| "output_scores": false, | |
| "return_dict_in_generate": false, | |
| "forced_bos_token_id": null, | |
| "forced_eos_token_id": null, | |
| "remove_invalid_values": false, | |
| "exponential_decay_length_penalty": null, | |
| "suppress_tokens": null, | |
| "begin_suppress_tokens": null, | |
| "architectures": null, | |
| "finetuning_task": null, | |
| "id2label": { | |
| "0": "LABEL_0", | |
| "1": "LABEL_1" | |
| }, | |
| "label2id": { | |
| "LABEL_0": 0, | |
| "LABEL_1": 1 | |
| }, | |
| "tokenizer_class": null, | |
| "prefix": null, | |
| "bos_token_id": 1, | |
| "pad_token_id": 32000, | |
| "eos_token_id": 2, | |
| "sep_token_id": null, | |
| "decoder_start_token_id": null, | |
| "task_specific_params": null, | |
| "problem_type": null, | |
| "_name_or_path": "", | |
| "model_type": "llama" | |
| }, | |
| "return_dict": true, | |
| "output_hidden_states": false, | |
| "output_attentions": false, | |
| "torchscript": false, | |
| "torch_dtype": "bfloat16", | |
| "use_bfloat16": false, | |
| "tf_legacy_loss": false, | |
| "pruned_heads": {}, | |
| "tie_word_embeddings": true, | |
| "chunk_size_feed_forward": 0, | |
| "is_encoder_decoder": false, | |
| "is_decoder": false, | |
| "cross_attention_hidden_size": null, | |
| "add_cross_attention": false, | |
| "tie_encoder_decoder": false, | |
| "max_length": 20, | |
| "min_length": 0, | |
| "do_sample": false, | |
| "early_stopping": false, | |
| "num_beams": 1, | |
| "num_beam_groups": 1, | |
| "diversity_penalty": 0.0, | |
| "temperature": 1.0, | |
| "top_k": 50, | |
| "top_p": 1.0, | |
| "typical_p": 1.0, | |
| "repetition_penalty": 1.0, | |
| "length_penalty": 1.0, | |
| "no_repeat_ngram_size": 0, | |
| "encoder_no_repeat_ngram_size": 0, | |
| "bad_words_ids": null, | |
| "num_return_sequences": 1, | |
| "output_scores": false, | |
| "return_dict_in_generate": false, | |
| "forced_bos_token_id": null, | |
| "forced_eos_token_id": null, | |
| "remove_invalid_values": false, | |
| "exponential_decay_length_penalty": null, | |
| "suppress_tokens": null, | |
| "begin_suppress_tokens": null, | |
| "architectures": [ | |
| "OpenVLAForActionPrediction" | |
| ], | |
| "finetuning_task": null, | |
| "id2label": { | |
| "0": "LABEL_0", | |
| "1": "LABEL_1" | |
| }, | |
| "label2id": { | |
| "LABEL_0": 0, | |
| "LABEL_1": 1 | |
| }, | |
| "tokenizer_class": null, | |
| "prefix": null, | |
| "bos_token_id": null, | |
| "eos_token_id": null, | |
| "sep_token_id": null, | |
| "decoder_start_token_id": null, | |
| "task_specific_params": null, | |
| "problem_type": null, | |
| "_name_or_path": "/home/guangyu/.cache/huggingface/hub/models--openvla--openvla-7b/snapshots/31f090d05236101ebfc381b61c674dd4746d4ce0", | |
| "transformers_version": "4.40.1", | |
| "auto_map": { | |
| "AutoConfig": "configuration_prismatic.OpenVLAConfig", | |
| "AutoModelForVision2Seq": "modeling_prismatic.OpenVLAForActionPrediction" | |
| }, | |
| "model_type": "openvla" | |
| } |