init

added_tokens.json

@@ -0,0 +1,116 @@
+{
+  "</img>": 32003,
+  "</p>": 32007,
+  "</roi>": 32005,
+  "<ground_box>": 32011,
+  "<image>": 32008,
+  "<img>": 32002,
+  "<p>": 32006,
+  "<r0>": 32014,
+  "<r10>": 32024,
+  "<r11>": 32025,
+  "<r12>": 32026,
+  "<r13>": 32027,
+  "<r14>": 32028,
+  "<r15>": 32029,
+  "<r16>": 32030,
+  "<r17>": 32031,
+  "<r18>": 32032,
+  "<r19>": 32033,
+  "<r1>": 32015,
+  "<r20>": 32034,
+  "<r21>": 32035,
+  "<r22>": 32036,
+  "<r23>": 32037,
+  "<r24>": 32038,
+  "<r25>": 32039,
+  "<r26>": 32040,
+  "<r27>": 32041,
+  "<r28>": 32042,
+  "<r29>": 32043,
+  "<r2>": 32016,
+  "<r30>": 32044,
+  "<r31>": 32045,
+  "<r32>": 32046,
+  "<r33>": 32047,
+  "<r34>": 32048,
+  "<r35>": 32049,
+  "<r36>": 32050,
+  "<r37>": 32051,
+  "<r38>": 32052,
+  "<r39>": 32053,
+  "<r3>": 32017,
+  "<r40>": 32054,
+  "<r41>": 32055,
+  "<r42>": 32056,
+  "<r43>": 32057,
+  "<r44>": 32058,
+  "<r45>": 32059,
+  "<r46>": 32060,
+  "<r47>": 32061,
+  "<r48>": 32062,
+  "<r49>": 32063,
+  "<r4>": 32018,
+  "<r50>": 32064,
+  "<r51>": 32065,
+  "<r52>": 32066,
+  "<r53>": 32067,
+  "<r54>": 32068,
+  "<r55>": 32069,
+  "<r56>": 32070,
+  "<r57>": 32071,
+  "<r58>": 32072,
+  "<r59>": 32073,
+  "<r5>": 32019,
+  "<r60>": 32074,
+  "<r61>": 32075,
+  "<r62>": 32076,
+  "<r63>": 32077,
+  "<r64>": 32078,
+  "<r65>": 32079,
+  "<r66>": 32080,
+  "<r67>": 32081,
+  "<r68>": 32082,
+  "<r69>": 32083,
+  "<r6>": 32020,
+  "<r70>": 32084,
+  "<r71>": 32085,
+  "<r72>": 32086,
+  "<r73>": 32087,
+  "<r74>": 32088,
+  "<r75>": 32089,
+  "<r76>": 32090,
+  "<r77>": 32091,
+  "<r78>": 32092,
+  "<r79>": 32093,
+  "<r7>": 32021,
+  "<r80>": 32094,
+  "<r81>": 32095,
+  "<r82>": 32096,
+  "<r83>": 32097,
+  "<r84>": 32098,
+  "<r85>": 32099,
+  "<r86>": 32100,
+  "<r87>": 32101,
+  "<r88>": 32102,
+  "<r89>": 32103,
+  "<r8>": 32022,
+  "<r90>": 32104,
+  "<r91>": 32105,
+  "<r92>": 32106,
+  "<r93>": 32107,
+  "<r94>": 32108,
+  "<r95>": 32109,
+  "<r96>": 32110,
+  "<r97>": 32111,
+  "<r98>": 32112,
+  "<r99>": 32113,
+  "<r9>": 32023,
+  "<refer_box>": 32010,
+  "<refer_feat>": 32012,
+  "<region>": 32009,
+  "<roi>": 32004,
+  "<sep>": 32001,
+  "[PAD]": 32000,
+  "[grounding]": 32013
+}

config.json

@@ -0,0 +1,75 @@
+{
+  "architectures": [
+    "GvllmModel"
+  ],
+  "box_score_thres": 0.15,
+  "llm_cfg": {
+    "_name_or_path": "vicuna-7b-v1.5",
+    "architectures": [
+      "LlamaForCausalLM"
+    ],
+    "bos_token_id": 1,
+    "eos_token_id": 2,
+    "hidden_act": "silu",
+    "hidden_size": 4096,
+    "initializer_range": 0.02,
+    "intermediate_size": 11008,
+    "max_position_embeddings": 4096,
+    "model_type": "llama",
+    "num_attention_heads": 32,
+    "num_hidden_layers": 32,
+    "num_key_value_heads": 32,
+    "pad_token_id": 0,
+    "pretraining_tp": 1,
+    "rms_norm_eps": 1e-05,
+    "rope_scaling": null,
+    "tie_word_embeddings": false,
+    "torch_dtype": "float16",
+    "transformers_version": "4.32.0",
+    "use_cache": true,
+    "vocab_size": 32000
+  },
+  "model_type": "gvllm",
+  "num_new_token": 114,
+  "perceiver_cfg": {
+    "architectures": [
+      "CustomDDETRModel"
+    ],
+    "ddetr_cfg": {
+      "auxiliary_loss": true,
+      "class_cost": 2,
+      "cls_loss_coefficient": 2,
+      "id2label": {
+        "0": "LABEL_0"
+      },
+      "label2id": {
+        "LABEL_0": 0
+      },
+      "model_type": "deformable_detr",
+      "num_feature_levels": 1,
+      "two_stage": true,
+      "with_box_refine": true
+    },
+    "model_type": "ddetr",
+    "torch_dtype": "float32",
+    "transformers_version": "4.32.0",
+    "vis_encoder_cfg": {
+      "architectures": [
+        "Dinov2Model"
+      ],
+      "hidden_size": 1024,
+      "image_size": 518,
+      "model_type": "dinov2",
+      "num_attention_heads": 16,
+      "num_hidden_layers": 24,
+      "patch_size": 14,
+      "torch_dtype": "float32"
+    },
+    "vis_encoder_path": "checkpoints/dinov2-large",
+    "vis_output_layer": -2,
+    "zs_weight_path": null
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.32.0",
+  "vocab_size": 32114
+}

generation_config.json

@@ -0,0 +1,10 @@
+{
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": 2,
+  "max_length": 4096,
+  "pad_token_id": 32000,
+  "temperature": 0.9,
+  "top_p": 0.6,
+  "transformers_version": "4.32.0"
+}

preprocessor_config.json

@@ -0,0 +1,27 @@
+{
+  "crop_size": {
+    "height": 224,
+    "width": 224
+  },
+  "do_center_crop": false,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": false,
+  "image_mean": [
+    0.485,
+    0.456,
+    0.406
+  ],
+  "image_processor_type": "BitImageProcessor",
+  "image_std": [
+    0.229,
+    0.224,
+    0.225
+  ],
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 256
+  }
+}

pytorch_model-00001-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b37d25843a5e571d84088a8bf6adc6c4c057fae571ed3a2ede7d9560a259afff
+size 9950049406

pytorch_model-00002-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:255ac8f38ccdcf8d50c6b87265af14326f50f3eb46a8a31af9d63dd0deea04bf
+size 9915768060

pytorch_model-00003-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a2ddd323f60e04825d601a3b6f39cc24fa3d83da0558f623bfbdc9ded39c4559
+size 9643849474

special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "[PAD]",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

tokenizer_config.json

@@ -0,0 +1,37 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "bos_token": {
+    "__type": "AddedToken",
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "clean_up_tokenization_spaces": false,
+  "eos_token": {
+    "__type": "AddedToken",
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "legacy": false,
+  "model_max_length": 2048,
+  "pad_token": null,
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": {
+    "__type": "AddedToken",
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "use_default_system_prompt": true
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3db1fe84a1d4bf5dd1bc7bd709e8d3da22958d8df9ead83e790daf47f0c653f6
+size 4283