{ "module": "keras_hub.src.models.blip2.blip2_causal_lm", "class_name": "BLIP2CausalLM", "config": { "backbone": { "module": "keras_hub.src.models.blip2.blip2_backbone", "class_name": "BLIP2Backbone", "config": { "name": "blip2_backbone", "trainable": true, "dtype": { "module": "keras", "class_name": "DTypePolicy", "config": { "name": "float32" }, "registered_name": null }, "vision_encoder": { "module": "keras_hub.src.models.blip2.blip2_vision_encoder", "class_name": "BLIP2VisionEncoder", "config": { "name": "blip2_vision_encoder", "trainable": true, "dtype": { "module": "keras", "class_name": "DTypePolicy", "config": { "name": "float32" }, "registered_name": null }, "image_size": [ 224, 224 ], "patch_size": [ 14, 14 ], "num_layers": 39, "num_heads": 16, "hidden_dim": 1408, "intermediate_dim": 6144, "use_patch_bias": true, "use_class_token": true, "use_mha_bias": true, "use_mlp_bias": true, "dropout_rate": 0.0, "layer_norm_epsilon": 1e-06, "initializer_range": 1e-10 }, "registered_name": "keras_hub>BLIP2VisionEncoder" }, "qformer": { "module": "keras_hub.src.models.blip2.blip2_qformer", "class_name": "BLIP2QFormer", "config": { "name": "blip2q_former", "trainable": true, "dtype": { "module": "keras", "class_name": "DTypePolicy", "config": { "name": "float32" }, "registered_name": null }, "num_query_tokens": 32, "num_layers": 12, "num_heads": 12, "hidden_dim": 768, "intermediate_dim": 3072, "vision_dim": 1408, "cross_attention_frequency": 2, "dropout": 0.1, "layer_norm_epsilon": 1e-12 }, "registered_name": "keras_hub>BLIP2QFormer" }, "language_model": { "module": "keras_hub.src.models.blip2.blip2_custom_opt", "class_name": "BLIP2CustomOPT", "config": { "name": "blip2_custom_opt", "trainable": true, "dtype": { "module": "keras", "class_name": "DTypePolicy", "config": { "name": "float32" }, "registered_name": null }, "vocabulary_size": 50304, "num_layers": 32, "num_heads": 32, "hidden_dim": 2560, "intermediate_dim": 10240, "num_query_tokens": 32, "dropout": 0.1, "max_sequence_length": 2048, "qformer_hidden_dim": 768, "initializer_range": 0.02, "layer_norm_epsilon": 1e-05, "language_projection": { "module": "keras.layers", "class_name": "EinsumDense", "config": { "name": "language_projection", "trainable": true, "dtype": { "module": "keras", "class_name": "DTypePolicy", "config": { "name": "float32" }, "registered_name": null }, "output_shape": [ null, 2560 ], "equation": "btd,df->btf", "activation": "linear", "bias_axes": "f", "kernel_initializer": { "module": "keras.initializers", "class_name": "GlorotUniform", "config": { "seed": null, "input_axes": null, "output_axes": null }, "registered_name": null }, "bias_initializer": { "module": "keras.initializers", "class_name": "Zeros", "config": {}, "registered_name": null }, "kernel_regularizer": null, "bias_regularizer": null, "activity_regularizer": null, "kernel_constraint": null, "bias_constraint": null, "quantization_config": null }, "registered_name": null } }, "registered_name": "keras_hub>BLIP2CustomOPT" } }, "registered_name": "keras_hub>BLIP2Backbone" }, "preprocessor": { "module": "keras_hub.src.models.blip2.blip2_causal_lm_preprocessor", "class_name": "BLIP2CausalLMPreprocessor", "config": { "name": "blip2_causal_lm_preprocessor", "trainable": true, "dtype": { "module": "keras", "class_name": "DTypePolicy", "config": { "name": "float32" }, "registered_name": null }, "tokenizer": { "module": "keras_hub.src.models.blip2.blip2_opt_tokenizer", "class_name": "BLIP2OPTTokenizer", "config": { "name": "blip2opt_tokenizer", "trainable": true, "dtype": { "module": "keras", "class_name": "DTypePolicy", "config": { "name": "int32" }, "registered_name": null }, "config_file": "tokenizer.json", "sequence_length": null, "add_prefix_space": false, "unsplittable_tokens": [ "", "", "" ] }, "registered_name": "keras_hub>BLIP2OPTTokenizer" }, "image_converter": { "module": "keras_hub.src.models.blip2.blip2_image_converter", "class_name": "BLIP2ImageConverter", "config": { "name": "blip2_image_converter", "trainable": true, "dtype": { "module": "keras", "class_name": "DTypePolicy", "config": { "name": "float32" }, "registered_name": null }, "image_size": [ 224, 224 ], "interpolation": "bicubic", "antialias": true, "crop_to_aspect_ratio": false, "pad_to_aspect_ratio": false, "bounding_box_format": "yxyx" }, "registered_name": "keras_hub>BLIP2ImageConverter" }, "config_file": "preprocessor.json", "sequence_length": 512, "add_start_token": true, "add_end_token": true }, "registered_name": "keras_hub>BLIP2CausalLMPreprocessor" }, "name": "blip2_causal_lm" }, "registered_name": "keras_hub>BLIP2CausalLM" }