Upload folder using huggingface_hub

config.json

@@ -0,0 +1,1386 @@
+{
+  "_name_or_path": "/shared_work/markhsp/checkpoints/X-VLA-Pt",
+  "action_mode": "auto",
+  "architectures": [
+    "XVLA"
+  ],
+  "auto_map": {
+    "AutoConfig": "configuration_xvla.XVLAConfig",
+    "AutoModel": "modeling_xvla.XVLA"
+  },
+  "depth": 24,
+  "dim_time": 32,
+  "florence_config": {
+    "_attn_implementation_autoset": true,
+    "bos_token_id": 0,
+    "eos_token_id": 2,
+    "is_encoder_decoder": true,
+    "model_type": "florence2",
+    "pad_token_id": 1,
+    "text_config": {
+      "_attn_implementation_autoset": true,
+      "_name_or_path": "",
+      "activation_dropout": 0.1,
+      "activation_function": "gelu",
+      "add_bias_logits": false,
+      "add_cross_attention": false,
+      "add_final_layer_norm": false,
+      "architectures": null,
+      "attention_dropout": 0.1,
+      "bad_words_ids": null,
+      "begin_suppress_tokens": null,
+      "bos_token_id": 0,
+      "chunk_size_feed_forward": 0,
+      "classif_dropout": 0.1,
+      "classifier_dropout": 0.0,
+      "cross_attention_hidden_size": null,
+      "d_model": 1024,
+      "decoder_attention_heads": 16,
+      "decoder_ffn_dim": 4096,
+      "decoder_layerdrop": 0.0,
+      "decoder_layers": 12,
+      "decoder_start_token_id": 2,
+      "diversity_penalty": 0.0,
+      "do_sample": false,
+      "dropout": 0.1,
+      "early_stopping": true,
+      "encoder_attention_heads": 16,
+      "encoder_ffn_dim": 4096,
+      "encoder_layerdrop": 0.0,
+      "encoder_layers": 12,
+      "encoder_no_repeat_ngram_size": 0,
+      "eos_token_id": 2,
+      "exponential_decay_length_penalty": null,
+      "finetuning_task": null,
+      "forced_bos_token_id": 0,
+      "forced_eos_token_id": 2,
+      "gradient_checkpointing": false,
+      "id2label": {
+        "0": "LABEL_0",
+        "1": "LABEL_1",
+        "2": "LABEL_2"
+      },
+      "init_std": 0.02,
+      "is_decoder": false,
+      "is_encoder_decoder": true,
+      "label2id": {
+        "LABEL_0": 0,
+        "LABEL_1": 1,
+        "LABEL_2": 2
+      },
+      "length_penalty": 1.0,
+      "max_length": 20,
+      "max_position_embeddings": 4096,
+      "min_length": 0,
+      "model_type": "florence2_language",
+      "no_repeat_ngram_size": 3,
+      "normalize_before": false,
+      "num_beam_groups": 1,
+      "num_beams": 3,
+      "num_hidden_layers": 12,
+      "num_return_sequences": 1,
+      "output_attentions": false,
+      "output_hidden_states": false,
+      "output_scores": false,
+      "pad_token_id": 1,
+      "prefix": null,
+      "problem_type": null,
+      "pruned_heads": {},
+      "remove_invalid_values": false,
+      "repetition_penalty": 1.0,
+      "return_dict": true,
+      "return_dict_in_generate": false,
+      "scale_embedding": false,
+      "sep_token_id": null,
+      "suppress_tokens": null,
+      "task_specific_params": null,
+      "temperature": 1.0,
+      "tf_legacy_loss": false,
+      "tie_encoder_decoder": false,
+      "tie_word_embeddings": true,
+      "tokenizer_class": null,
+      "top_k": 50,
+      "top_p": 1.0,
+      "torch_dtype": null,
+      "torchscript": false,
+      "typical_p": 1.0,
+      "use_bfloat16": false,
+      "use_cache": true,
+      "vocab_size": 51289
+    },
+    "torch_dtype": "float32",
+    "vision_config": {
+      "_attn_implementation_autoset": false,
+      "_name_or_path": "",
+      "add_cross_attention": false,
+      "architectures": null,
+      "bad_words_ids": null,
+      "begin_suppress_tokens": null,
+      "bos_token_id": null,
+      "chunk_size_feed_forward": 0,
+      "cross_attention_hidden_size": null,
+      "decoder_start_token_id": null,
+      "depths": [
+        1,
+        1,
+        9,
+        1
+      ],
+      "dim_embed": [
+        256,
+        512,
+        1024,
+        2048
+      ],
+      "diversity_penalty": 0.0,
+      "do_sample": false,
+      "drop_path_rate": 0.1,
+      "early_stopping": false,
+      "enable_checkpoint": false,
+      "encoder_no_repeat_ngram_size": 0,
+      "eos_token_id": null,
+      "exponential_decay_length_penalty": null,
+      "finetuning_task": null,
+      "forced_bos_token_id": null,
+      "forced_eos_token_id": null,
+      "id2label": {
+        "0": "LABEL_0",
+        "1": "LABEL_1"
+      },
+      "image_feature_source": [
+        "spatial_avg_pool",
+        "temporal_avg_pool"
+      ],
+      "image_pos_embed": {
+        "max_pos_embeddings": 50,
+        "type": "learned_abs_2d"
+      },
+      "is_decoder": false,
+      "is_encoder_decoder": false,
+      "label2id": {
+        "LABEL_0": 0,
+        "LABEL_1": 1
+      },
+      "length_penalty": 1.0,
+      "max_length": 20,
+      "min_length": 0,
+      "model_type": "davit",
+      "no_repeat_ngram_size": 0,
+      "num_beam_groups": 1,
+      "num_beams": 1,
+      "num_groups": [
+        8,
+        16,
+        32,
+        64
+      ],
+      "num_heads": [
+        8,
+        16,
+        32,
+        64
+      ],
+      "num_return_sequences": 1,
+      "output_attentions": false,
+      "output_hidden_states": false,
+      "output_scores": false,
+      "pad_token_id": null,
+      "patch_padding": [
+        3,
+        1,
+        1,
+        1
+      ],
+      "patch_prenorm": [
+        false,
+        true,
+        true,
+        true
+      ],
+      "patch_size": [
+        7,
+        3,
+        3,
+        3
+      ],
+      "patch_stride": [
+        4,
+        2,
+        2,
+        2
+      ],
+      "prefix": null,
+      "problem_type": null,
+      "projection_dim": 1024,
+      "pruned_heads": {},
+      "remove_invalid_values": false,
+      "repetition_penalty": 1.0,
+      "return_dict": true,
+      "return_dict_in_generate": false,
+      "sep_token_id": null,
+      "suppress_tokens": null,
+      "task_specific_params": null,
+      "temperature": 1.0,
+      "tf_legacy_loss": false,
+      "tie_encoder_decoder": false,
+      "tie_word_embeddings": true,
+      "tokenizer_class": null,
+      "top_k": 50,
+      "top_p": 1.0,
+      "torch_dtype": null,
+      "torchscript": false,
+      "typical_p": 1.0,
+      "use_bfloat16": false,
+      "visual_temporal_embedding": {
+        "max_temporal_embeddings": 100,
+        "type": "COSINE"
+      },
+      "window_size": 12
+    }
+  },
+  "hidden_size": 1024,
+  "len_soft_prompts": 32,
+  "max_action_dim": 23,
+  "max_len_seq": 512,
+  "mlp_ratio": 4.0,
+  "model_type": "xvla",
+  "num_actions": 30,
+  "num_domains": 30,
+  "num_heads": 16,
+  "num_objects": 172,
+  "num_skills": 34,
+  "num_tasks": 50,
+  "object_class_weights": [
+    0.4501,
+    0.4205,
+    0.3228,
+    0.3057,
+    0.3521,
+    7.3765,
+    0.2465,
+    0.5885,
+    0.4622,
+    0.3009,
+    0.281,
+    0.3189,
+    0.1221,
+    0.4479,
+    0.312,
+    0.4966,
+    0.3079,
+    0.2972,
+    0.5216,
+    0.1139,
+    0.2091,
+    0.2318,
+    0.1754,
+    0.2122,
+    0.3788,
+    0.2522,
+    0.2869,
+    0.2588,
+    0.5604,
+    0.2401,
+    0.4887,
+    0.2479,
+    0.4339,
+    0.2858,
+    0.4104,
+    0.1655,
+    0.5658,
+    0.1832,
+    0.4361,
+    0.3443,
+    0.4715,
+    0.4035,
+    0.3555,
+    0.5542,
+    0.5411,
+    0.3756,
+    0.6119,
+    0.3592,
+    0.3795,
+    0.17,
+    0.6553,
+    1.8098,
+    0.2901,
+    0.2248,
+    0.2027,
+    0.4599,
+    0.5536,
+    0.508,
+    0.2674,
+    0.5678,
+    0.3297,
+    0.4523,
+    0.2186,
+    0.2029,
+    0.4368,
+    0.432,
+    0.0822,
+    0.3718,
+    0.6172,
+    0.3748,
+    0.2685,
+    0.2913,
+    0.4405,
+    0.3391,
+    0.3676,
+    6.0925,
+    2.0803,
+    2.153,
+    2.2968,
+    2.1286,
+    12.6916,
+    0.5856,
+    5.4459,
+    2.6565,
+    13.1846,
+    23.4236,
+    0.478,
+    0.462,
+    0.1174,
+    0.4406,
+    0.2615,
+    0.2668,
+    0.2285,
+    0.3084,
+    0.2523,
+    0.3026,
+    0.2844,
+    0.5057,
+    0.343,
+    0.2129,
+    0.2497,
+    0.3409,
+    0.3574,
+    0.3983,
+    0.1783,
+    0.2182,
+    0.3948,
+    3.4148,
+    0.3134,
+    0.4798,
+    0.3458,
+    0.2877,
+    0.4385,
+    0.2716,
+    0.4203,
+    0.2383,
+    0.2431,
+    0.2523,
+    0.4795,
+    0.4197,
+    0.2102,
+    0.15,
+    0.3516,
+    1.1414,
+    0.1527,
+    0.4403,
+    0.188,
+    0.5024,
+    0.367,
+    0.4584,
+    0.293,
+    0.3649,
+    23.0659,
+    0.1816,
+    0.4641,
+    0.438,
+    0.2461,
+    2.3487,
+    0.3496,
+    0.405,
+    0.3314,
+    0.489,
+    0.4397,
+    0.4996,
+    0.1229,
+    3.6319,
+    0.4439,
+    0.2366,
+    0.2766,
+    0.2168,
+    0.3219,
+    0.3337,
+    0.4465,
+    0.7571,
+    0.257,
+    0.3779,
+    0.1382,
+    0.2154,
+    0.2358,
+    0.4292,
+    0.3343,
+    0.3047,
+    0.8751,
+    0.3987,
+    0.146,
+    0.2943,
+    0.8747,
+    0.1515,
+    0.1481,
+    0.5527,
+    2.8632,
+    0.3923
+  ],
+  "object_classifier_weight": 0.1,
+  "object_prompt_length": 0,
+  "real_action_dim": 23,
+  "skill_class_weights": [
+    0.0859,
+    0.1021,
+    0.2138,
+    0.1674,
+    0.8973,
+    1.059,
+    1.6727,
+    0.9436,
+    0.1984,
+    1.1787,
+    0.2744,
+    0.621,
+    0.6252,
+    1.8634,
+    0.5304,
+    0.5203,
+    0.8339,
+    0.8988,
+    1.4175,
+    2.2557,
+    0.5537,
+    1.1661,
+    3.0415,
+    0.1975,
+    0.3228,
+    1.2419,
+    0.6963,
+    1.2705,
+    0.4855,
+    1.1792,
+    1.4312,
+    2.6619,
+    1.3428,
+    2.0497
+  ],
+  "skill_classifier_weight": 0.1,
+  "skill_prompt_length": 32,
+  "soft_prompt_length": 32,
+  "task_class_weights": [
+    2.173,
+    1.4191,
+    0.8104,
+    0.8292,
+    0.8919,
+    0.9968,
+    1.1146,
+    0.707,
+    1.029,
+    0.8308,
+    1.2263,
+    0.928,
+    1.0689,
+    0.6989,
+    0.8124,
+    0.8448,
+    0.8113,
+    1.0045,
+    0.9249,
+    0.9423,
+    0.8886,
+    0.7005,
+    1.1158,
+    0.6354,
+    0.6881,
+    0.8477,
+    0.5946,
+    0.7707,
+    0.7753,
+    0.6626,
+    1.0688,
+    1.0711,
+    1.0735,
+    0.9245,
+    2.0169,
+    1.6108,
+    0.8854,
+    1.4132,
+    1.2177,
+    1.0697,
+    1.7812,
+    0.8162,
+    1.2199,
+    0.7956,
+    0.9433,
+    1.0236,
+    1.1092,
+    0.8773,
+    0.6393,
+    0.6997
+  ],
+  "task_classifier_weight": 0.1,
+  "task_prompt_length": 0,
+  "task_valid_objects": {
+    "0": [
+      47,
+      131
+    ],
+    "1": [
+      33,
+      62,
+      159
+    ],
+    "2": [
+      19,
+      38,
+      47,
+      62,
+      121,
+      130
+    ],
+    "3": [
+      22,
+      27,
+      56,
+      66,
+      124
+    ],
+    "4": [
+      26,
+      49,
+      90,
+      132,
+      156
+    ],
+    "5": [
+      19,
+      62,
+      68,
+      104,
+      153
+    ],
+    "6": [
+      58,
+      161,
+      168
+    ],
+    "7": [
+      7,
+      12,
+      50,
+      94,
+      151,
+      157
+    ],
+    "8": [
+      5,
+      64,
+      65,
+      132,
+      152,
+      156
+    ],
+    "9": [
+      27,
+      34,
+      43,
+      47,
+      62,
+      70,
+      121,
+      132,
+      140,
+      168,
+      170
+    ],
+    "10": [
+      15,
+      45,
+      46,
+      48,
+      59,
+      114,
+      134,
+      137
+    ],
+    "11": [
+      5,
+      48,
+      124,
+      156
+    ],
+    "12": [
+      18,
+      30,
+      40,
+      42,
+      44,
+      48,
+      66,
+      72,
+      112
+    ],
+    "13": [
+      35,
+      47,
+      53,
+      54,
+      62,
+      132,
+      150,
+      157
+    ],
+    "14": [
+      8,
+      27,
+      35,
+      36,
+      54,
+      62,
+      66,
+      113
+    ],
+    "15": [
+      54,
+      62,
+      126
+    ],
+    "16": [
+      54,
+      62,
+      144
+    ],
+    "17": [
+      9,
+      47,
+      66
+    ],
+    "18": [
+      7,
+      62,
+      88,
+      109,
+      133
+    ],
+    "19": [
+      0,
+      48,
+      55,
+      61,
+      125,
+      135,
+      154
+    ],
+    "20": [
+      13,
+      28,
+      62,
+      97,
+      101,
+      146,
+      165,
+      168
+    ],
+    "21": [
+      7,
+      12,
+      14,
+      50,
+      52,
+      62,
+      145,
+      149,
+      158
+    ],
+    "22": [
+      62,
+      71,
+      86,
+      133
+    ],
+    "23": [
+      14,
+      62,
+      88,
+      140,
+      144
+    ],
+    "24": [
+      3,
+      17,
+      19,
+      23,
+      48,
+      93
+    ],
+    "25": [
+      5,
+      27,
+      66,
+      73,
+      74,
+      124,
+      164
+    ],
+    "26": [
+      19,
+      21,
+      31,
+      47,
+      62,
+      121,
+      147,
+      168
+    ],
+    "27": [
+      16,
+      24,
+      45,
+      62,
+      87,
+      106,
+      132,
+      137,
+      138,
+      139,
+      155,
+      163
+    ],
+    "28": [
+      50,
+      51,
+      62,
+      63,
+      95,
+      102,
+      103,
+      110,
+      117,
+      132,
+      148
+    ],
+    "29": [
+      7,
+      14,
+      50,
+      57,
+      62,
+      63,
+      96,
+      115,
+      117,
+      118,
+      119,
+      132,
+      142
+    ],
+    "30": [
+      47,
+      60,
+      62,
+      98,
+      108,
+      169
+    ],
+    "31": [
+      25,
+      48,
+      132,
+      167
+    ],
+    "32": [
+      6,
+      48,
+      132,
+      167
+    ],
+    "33": [
+      20,
+      141,
+      149,
+      151,
+      167
+    ],
+    "34": [
+      5,
+      62,
+      128,
+      166
+    ],
+    "35": [
+      32,
+      53,
+      62
+    ],
+    "36": [
+      29,
+      62,
+      69,
+      105,
+      132
+    ],
+    "37": [
+      50,
+      62,
+      136,
+      162
+    ],
+    "38": [
+      62,
+      69,
+      92,
+      129,
+      132
+    ],
+    "39": [
+      62,
+      120,
+      161
+    ],
+    "40": [
+      5,
+      100,
+      127
+    ],
+    "41": [
+      30,
+      37,
+      39,
+      48,
+      49,
+      56,
+      66,
+      67,
+      75,
+      89,
+      124,
+      132
+    ],
+    "42": [
+      5,
+      22,
+      49,
+      56,
+      82,
+      116,
+      132,
+      165
+    ],
+    "43": [
+      10,
+      11,
+      48,
+      49,
+      56,
+      66,
+      84,
+      116,
+      132,
+      171
+    ],
+    "44": [
+      2,
+      41,
+      76,
+      77,
+      78,
+      79,
+      85,
+      99
+    ],
+    "45": [
+      48,
+      62,
+      66,
+      91,
+      100
+    ],
+    "46": [
+      30,
+      48,
+      62,
+      66,
+      67,
+      160
+    ],
+    "47": [
+      1,
+      20,
+      48,
+      62,
+      66,
+      124,
+      164
+    ],
+    "48": [
+      19,
+      22,
+      37,
+      48,
+      49,
+      56,
+      66,
+      81,
+      122,
+      143
+    ],
+    "49": [
+      4,
+      5,
+      22,
+      37,
+      42,
+      56,
+      62,
+      66,
+      80,
+      83,
+      107,
+      111,
+      123,
+      132,
+      156,
+      164,
+      165
+    ]
+  },
+  "task_valid_skills": {
+    "0": [
+      0,
+      1,
+      2,
+      19
+    ],
+    "1": [
+      0,
+      1,
+      2,
+      3
+    ],
+    "2": [
+      0,
+      1,
+      3,
+      7,
+      9,
+      24
+    ],
+    "3": [
+      0,
+      1,
+      3,
+      8,
+      10,
+      23
+    ],
+    "4": [
+      0,
+      1,
+      3,
+      4,
+      8,
+      10,
+      11,
+      12,
+      26
+    ],
+    "5": [
+      0,
+      1,
+      24
+    ],
+    "6": [
+      0,
+      1,
+      2,
+      24
+    ],
+    "7": [
+      0,
+      1,
+      2,
+      3,
+      4,
+      23
+    ],
+    "8": [
+      0,
+      1,
+      3,
+      8,
+      10,
+      23,
+      25,
+      26
+    ],
+    "9": [
+      0,
+      1,
+      2,
+      14,
+      23,
+      24
+    ],
+    "10": [
+      0,
+      1,
+      2,
+      3,
+      23,
+      24
+    ],
+    "11": [
+      0,
+      1,
+      2,
+      3,
+      8,
+      10
+    ],
+    "12": [
+      0,
+      1,
+      3,
+      8,
+      10,
+      23,
+      24
+    ],
+    "13": [
+      0,
+      1,
+      3,
+      8,
+      11,
+      12,
+      23
+    ],
+    "14": [
+      0,
+      1,
+      2,
+      3,
+      4,
+      8,
+      10,
+      12,
+      14,
+      30
+    ],
+    "15": [
+      0,
+      1,
+      2,
+      8
+    ],
+    "16": [
+      0,
+      1,
+      2,
+      8
+    ],
+    "17": [
+      0,
+      1,
+      2,
+      4,
+      8,
+      10
+    ],
+    "18": [
+      0,
+      1,
+      2,
+      4,
+      23,
+      24
+    ],
+    "19": [
+      0,
+      1,
+      2,
+      3,
+      11,
+      12,
+      23
+    ],
+    "20": [
+      0,
+      1,
+      2,
+      3,
+      23
+    ],
+    "21": [
+      0,
+      1,
+      2,
+      3,
+      4,
+      23,
+      24
+    ],
+    "22": [
+      0,
+      1,
+      3
+    ],
+    "23": [
+      0,
+      1,
+      2,
+      3,
+      23
+    ],
+    "24": [
+      0,
+      1,
+      2,
+      3,
+      8,
+      10,
+      23,
+      30
+    ],
+    "25": [
+      0,
+      1,
+      2,
+      3,
+      8,
+      10,
+      23
+    ],
+    "26": [
+      0,
+      1,
+      2,
+      3,
+      4,
+      23
+    ],
+    "27": [
+      0,
+      1,
+      2,
+      3,
+      8,
+      23,
+      29
+    ],
+    "28": [
+      0,
+      1,
+      2,
+      4,
+      23,
+      24,
+      26
+    ],
+    "29": [
+      0,
+      1,
+      2,
+      3,
+      4,
+      5,
+      12,
+      23,
+      24,
+      26
+    ],
+    "30": [
+      0,
+      1,
+      3,
+      20,
+      21,
+      22,
+      23
+    ],
+    "31": [
+      0,
+      1,
+      3,
+      8,
+      10,
+      20,
+      26
+    ],
+    "32": [
+      0,
+      1,
+      3,
+      8,
+      10,
+      20,
+      26
+    ],
+    "33": [
+      0,
+      1,
+      3,
+      8,
+      10,
+      20,
+      23
+    ],
+    "34": [
+      0,
+      1,
+      4,
+      18,
+      23
+    ],
+    "35": [
+      0,
+      1,
+      2,
+      6,
+      13,
+      27
+    ],
+    "36": [
+      0,
+      1,
+      2,
+      4,
+      6,
+      17,
+      23,
+      27
+    ],
+    "37": [
+      0,
+      1,
+      2,
+      4,
+      16
+    ],
+    "38": [
+      0,
+      1,
+      20,
+      21,
+      23,
+      28
+    ],
+    "39": [
+      0,
+      1,
+      20,
+      21,
+      28
+    ],
+    "40": [
+      0,
+      1,
+      3,
+      8,
+      10,
+      20
+    ],
+    "41": [
+      0,
+      1,
+      2,
+      8,
+      10,
+      14,
+      15,
+      20,
+      23,
+      24,
+      26
+    ],
+    "42": [
+      0,
+      1,
+      2,
+      3,
+      14,
+      15,
+      23,
+      24,
+      26
+    ],
+    "43": [
+      0,
+      1,
+      2,
+      8,
+      10,
+      15,
+      23,
+      26
+    ],
+    "44": [
+      0,
+      1,
+      2,
+      15,
+      23,
+      33
+    ],
+    "45": [
+      0,
+      1,
+      2,
+      3,
+      8,
+      10,
+      20
+    ],
+    "46": [
+      0,
+      1,
+      2,
+      8,
+      10,
+      14,
+      20,
+      24,
+      26
+    ],
+    "47": [
+      0,
+      1,
+      2,
+      3,
+      8,
+      10
+    ],
+    "48": [
+      0,
+      1,
+      2,
+      3,
+      8,
+      10,
+      14,
+      15,
+      23,
+      24
+    ],
+    "49": [
+      0,
+      1,
+      2,
+      3,
+      8,
+      10,
+      14,
+      15,
+      20,
+      23,
+      24,
+      26,
+      31,
+      32
+    ]
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.49.0",
+  "use_hetero_proj": false,
+  "use_proprio": true
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e04770e92b416b2f92d45fdca1b6f2e72061dc006c372079541ffbb75b9e8ead
+size 3626506856

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd470b6cf5eff22e027ff218190b5255ea28eee8e4176ba90aa94ab3ccaa0b46
+size 7243647846

preprocessor_config.json

@@ -0,0 +1,32 @@
+{
+  "auto_map": {
+    "AutoProcessor": "processing_xvla.XVLAProcessor"
+  },
+  "crop_size": {
+    "height": 224,
+    "width": 224
+  },
+  "do_center_crop": false,
+  "do_convert_rgb": null,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.485,
+    0.456,
+    0.406
+  ],
+  "image_processor_type": "CLIPImageProcessor",
+  "image_std": [
+    0.229,
+    0.224,
+    0.225
+  ],
+  "processor_class": "XVLAProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}

special_tokens_map.json

@@ -0,0 +1,15 @@
+{
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": {
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

state.json

@@ -0,0 +1 @@
+{"global_step": 100000}

tokenizer_config.json

@@ -0,0 +1,59 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "50264": {
+      "content": "<mask>",
+      "lstrip": true,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "mask_token": "<mask>",
+  "model_max_length": 1024,
+  "pad_token": "<pad>",
+  "processor_class": "XVLAProcessor",
+  "sep_token": "</s>",
+  "tokenizer_class": "BartTokenizer",
+  "trim_offsets": true,
+  "unk_token": "<unk>"
+}