blip2_opt_2.7b / config.json
laxmareddyp's picture
Upload folder using huggingface_hub
996df21 verified
Raw
History Blame Contribute Delete
5.67 kB
{
"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"
}