{ "norm_stats": { "gen72_grasp_best2": { "action": { "mean": [ 79.11602783203125, 94.15535736083984, -80.4166259765625, -93.67428588867188, -3.3462841510772705, 17.82291603088379, 6.387060165405273, 78.9906997680664, -82.27678680419922, 88.60293579101562, 82.36312103271484, -95.4980239868164, 3.2222084999084473, 15.378264427185059, 1.2210661172866821, 80.26712799072266 ], "std": [ 21.024213790893555, 5.583713531494141, 10.143651008605957, 2.8190293312072754, 6.1579060554504395, 12.958017349243164, 9.206350326538086, 38.70134735107422, 24.093891143798828, 10.70143985748291, 13.960134506225586, 3.1140949726104736, 12.975412368774414, 19.39619255065918, 11.416078567504883, 38.01255416870117 ], "max": [ 112.63500213623047, 102.0, -43.02000045776367, -68.55000305175781, 20.434999465942383, 61.125, 46.845001220703125, 100.0, -18.19499969482422, 102.0, 110.38999938964844, -72.7699966430664, 71.93499755859375, 67.31999969482422, 42.23500061035156, 100.0 ], "min": [ 31.375, 66.01000213623047, -99.53500366210938, -95.75499725341797, -38.540000915527344, -22.940000534057617, -37.529998779296875, 0.0, -115.30999755859375, 14.984999656677246, 18.020000457763672, -97.47000122070312, -22.719999313354492, -23.639999389648438, -41.84000015258789, 0.0 ], "q01": [ 34.29199905395508, 78.30999755859375, -93.43000030517578, -95.44999694824219, -20.040000915527344, -3.5509999990463235, -15.819999694824219, 0.0, -109.94999694824219, 46.845001220703125, 35.15999984741211, -97.29000091552734, -15.1479998588562, -12.739999771118164, -30.791500282287597, 0.0 ], "q99": [ 105.26450119018561, 102.0, -49.8484996795654, -80.68149948120117, 8.350000381469727, 47.290000915527344, 32.16999816894531, 100.0, -30.818500328063934, 102.0, 100.80999755859375, -81.88200302124021, 52.8650016784668, 55.61800079345706, 30.233500289916996, 100.0 ], "mask": [ true, true, true, true, true, true, true, true, true, true, true, true, true, true, true, true ] }, "proprio": { "mean": [ 79.11497497558594, 94.18975830078125, -80.38821411132812, -93.6719970703125, -3.3557398319244385, 17.77706527709961, 6.375845909118652, 78.99101257324219, -82.22480773925781, 88.610107421875, 82.32734680175781, -95.49593353271484, 3.2511658668518066, 15.379179000854492, 1.202197551727295, 80.26780700683594 ], "std": [ 20.358299255371094, 5.266684055328369, 9.759184837341309, 2.6425156593322754, 5.872856140136719, 12.372116088867188, 8.679393768310547, 38.70150375366211, 23.474021911621094, 10.264312744140625, 13.466100692749023, 2.960315227508545, 12.600417137145996, 18.766708374023438, 11.020885467529297, 38.01284408569336 ], "max": [ 108.94000244140625, 102.0199966430664, -45.0099983215332, -72.97000122070312, 14.890000343322754, 55.84000015258789, 46.47999954223633, 100.0, -21.459999084472656, 102.01000213623047, 104.38999938964844, -73.69999694824219, 62.630001068115234, 64.4000015258789, 40.790000915527344, 100.0 ], "min": [ 32.54999923706055, 73.81999969482422, -98.31999969482422, -95.58000183105469, -36.43000030517578, -16.389999389648438, -26.139999389648438, 0.0, -111.0999984741211, 28.559999465942383, 19.8700008392334, -97.37000274658203, -16.200000762939453, -18.959999084472656, -39.83000183105469, 0.0 ], "q01": [ 36.040000915527344, 79.12000274658203, -92.80000305175781, -95.37999725341797, -19.156999778747558, -0.11400000452995164, -12.917000007629394, 0.0, -109.78100128173828, 49.36499938964844, 36.233000946044925, -97.2300033569336, -14.706999969482421, -12.258000183105468, -30.653999710083006, 0.0 ], "q99": [ 102.68300018310553, 102.01000213623047, -51.226001358032214, -81.30999755859375, 8.239999771118164, 45.500998687744165, 31.277000617980963, 100.0, -31.829999351501392, 101.95999908447266, 100.08000183105469, -82.35599975585936, 51.74799919128421, 53.9569995880127, 29.15399971008302, 100.0 ] }, "num_transitions": 16131, "num_trajectories": 49 } }, "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" }