Update tiny models for OmDetTurboForObjectDetection

config.json

@@ -0,0 +1,124 @@
+{
+  "apply_layernorm_after_vision_backbone": false,
+  "architectures": [
+    "OmDetTurboForObjectDetection"
+  ],
+  "backbone": null,
+  "backbone_config": {
+    "attention_probs_dropout_prob": 0.0,
+    "depths": [
+      1,
+      1,
+      1,
+      1
+    ],
+    "drop_path_rate": 0.1,
+    "embed_dim": 8,
+    "encoder_stride": 32,
+    "hidden_act": "gelu",
+    "hidden_dropout_prob": 0.0,
+    "hidden_size": 64,
+    "image_size": 224,
+    "initializer_range": 0.02,
+    "layer_norm_eps": 1e-05,
+    "mlp_ratio": 4.0,
+    "model_type": "swin",
+    "num_channels": 3,
+    "num_heads": [
+      1,
+      1,
+      1,
+      1
+    ],
+    "num_layers": 4,
+    "out_features": [
+      "stage2",
+      "stage3",
+      "stage4"
+    ],
+    "out_indices": [
+      2,
+      3,
+      4
+    ],
+    "patch_size": 4,
+    "qkv_bias": true,
+    "stage_names": [
+      "stem",
+      "stage1",
+      "stage2",
+      "stage3",
+      "stage4"
+    ],
+    "use_absolute_embeddings": false,
+    "window_size": 7
+  },
+  "batch_norm_eps": 1e-05,
+  "cache_size": 100,
+  "class_distance_type": "cosine",
+  "class_embed_dim": 16,
+  "conv_norm_activation": "gelu",
+  "csp_activation": "silu",
+  "d_model": 8,
+  "decoder_activation": "relu",
+  "decoder_dim_feedforward": 32,
+  "decoder_dropout": 0.0,
+  "decoder_hidden_dim": 8,
+  "decoder_num_heads": 2,
+  "decoder_num_layers": 2,
+  "decoder_num_points": 4,
+  "disable_custom_kernels": false,
+  "dtype": "float32",
+  "encoder_attention_heads": 2,
+  "encoder_dim_feedforward": 32,
+  "encoder_dropout": 0.0,
+  "encoder_feedforward_activation": "relu",
+  "encoder_feedforward_dropout": 0.0,
+  "encoder_in_channels": [
+    16,
+    32,
+    64
+  ],
+  "encoder_layers": 2,
+  "encoder_projection_indices": [
+    2
+  ],
+  "eval_size": null,
+  "hidden_expansion": 1,
+  "image_size": 224,
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "layer_norm_eps": 1e-05,
+  "learn_initial_query": false,
+  "model_type": "omdet-turbo",
+  "num_feature_levels": 3,
+  "num_queries": 20,
+  "positional_encoding_temperature": 10000,
+  "task_encoder_hidden_dim": 32,
+  "text_config": {
+    "attention_dropout": 0.0,
+    "bos_token_id": 49406,
+    "eos_token_id": 49407,
+    "hidden_act": "quick_gelu",
+    "hidden_size": 16,
+    "initializer_factor": 1.0,
+    "initializer_range": 0.02,
+    "intermediate_size": 16,
+    "layer_norm_eps": 1e-05,
+    "max_position_embeddings": 200,
+    "model_type": "clip_text_model",
+    "num_attention_heads": 2,
+    "num_hidden_layers": 2,
+    "pad_token_id": 49407,
+    "projection_dim": 512,
+    "vocab_size": 49408
+  },
+  "text_projection_in_dim": 16,
+  "text_projection_out_dim": 16,
+  "transformers_version": "5.16.0.dev0",
+  "vision_features_channels": [
+    8,
+    8,
+    8
+  ]
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f83383913569dbb1e917bdcfb8fc43066a7dccbd7ae15e43f4d1a23fc59490bb
+size 3647208

preprocessor_config.json

@@ -0,0 +1,29 @@
+{
+  "crop_size": {
+    "height": 224,
+    "width": 224
+  },
+  "do_convert_annotations": true,
+  "do_normalize": true,
+  "do_pad": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "format": "coco_detection",
+  "image_mean": [
+    0.485,
+    0.456,
+    0.406
+  ],
+  "image_processor_type": "DetrImageProcessor",
+  "image_std": [
+    0.229,
+    0.224,
+    0.225
+  ],
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}

tokenizer_config.json

@@ -0,0 +1,16 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": "<|startoftext|>",
+  "clean_up_tokenization_spaces": true,
+  "do_lower_case": true,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "is_local": true,
+  "local_files_only": false,
+  "model_max_length": 77,
+  "pad_token": "<|endoftext|>",
+  "processor_class": "OmDetTurboProcessor",
+  "tokenizer_class": "CLIPTokenizer",
+  "unk_token": "<|endoftext|>"
+}