| { |
| "module": "keras_hub.src.models.t5gemma.t5gemma_seq_2_seq_lm", |
| "class_name": "T5GemmaSeq2SeqLM", |
| "config": { |
| "backbone": { |
| "module": "keras_hub.src.models.t5gemma.t5gemma_backbone", |
| "class_name": "T5GemmaBackbone", |
| "config": { |
| "name": "t5_gemma_backbone", |
| "trainable": true, |
| "dtype": { |
| "module": "keras", |
| "class_name": "DTypePolicy", |
| "config": { |
| "name": "float32" |
| }, |
| "registered_name": null |
| }, |
| "vocabulary_size": 256000, |
| "encoder_hidden_dim": 512, |
| "encoder_intermediate_dim": 1024, |
| "encoder_num_layers": 8, |
| "encoder_num_attention_heads": 8, |
| "encoder_num_key_value_heads": 8, |
| "encoder_layer_types": [ |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "full_attention" |
| ], |
| "encoder_head_dim": 64, |
| "decoder_hidden_dim": 512, |
| "decoder_intermediate_dim": 1024, |
| "decoder_num_layers": 8, |
| "decoder_num_attention_heads": 8, |
| "decoder_num_key_value_heads": 8, |
| "decoder_layer_types": [ |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "full_attention" |
| ], |
| "decoder_head_dim": 64, |
| "dropout_rate": 0.0, |
| "rms_norm_eps": 1e-06, |
| "tie_word_embeddings": true, |
| "query_pre_attn_scalar": 64, |
| "attention_bias": false, |
| "hidden_activation": "gelu_approximate", |
| "initializer_range": 0.02, |
| "attention_dropout": 0.0, |
| "sliding_window": 4096, |
| "cross_attention_hidden_size": 512, |
| "attn_logit_softcapping": 50.0, |
| "final_logit_softcapping": 30.0, |
| "rope_max_wavelength": 10000.0 |
| }, |
| "registered_name": "keras_hub>T5GemmaBackbone" |
| }, |
| "preprocessor": { |
| "module": "keras_hub.src.models.t5gemma.t5gemma_seq_2_seq_lm_preprocessor", |
| "class_name": "T5GemmaSeq2SeqLMPreprocessor", |
| "config": { |
| "name": "t5_gemma_seq2_seq_lm_preprocessor", |
| "trainable": true, |
| "dtype": { |
| "module": "keras", |
| "class_name": "DTypePolicy", |
| "config": { |
| "name": "float32" |
| }, |
| "registered_name": null |
| }, |
| "tokenizer": { |
| "module": "keras_hub.src.models.t5gemma.t5gemma_tokenizer", |
| "class_name": "T5GemmaTokenizer", |
| "config": { |
| "name": "t5_gemma_tokenizer", |
| "trainable": true, |
| "dtype": { |
| "module": "keras", |
| "class_name": "DTypePolicy", |
| "config": { |
| "name": "int32" |
| }, |
| "registered_name": null |
| }, |
| "config_file": "tokenizer.json", |
| "proto": null, |
| "sequence_length": null, |
| "add_bos": false, |
| "add_eos": false |
| }, |
| "registered_name": "keras_hub>T5GemmaTokenizer" |
| }, |
| "config_file": "preprocessor.json", |
| "add_start_token": false, |
| "add_end_token": true |
| }, |
| "registered_name": "keras_hub>T5GemmaSeq2SeqLMPreprocessor" |
| }, |
| "name": "t5_gemma_seq2_seq_lm" |
| }, |
| "registered_name": "keras_hub>T5GemmaSeq2SeqLM" |
| } |
