| { | |
| "module": "keras_nlp.src.models.phi3.phi3_causal_lm", | |
| "class_name": "Phi3CausalLM", | |
| "config": { | |
| "backbone": { | |
| "module": "keras_nlp.src.models.phi3.phi3_backbone", | |
| "class_name": "Phi3Backbone", | |
| "config": { | |
| "name": "phi3_backbone_1", | |
| "trainable": true, | |
| "vocabulary_size": 32064, | |
| "num_layers": 32, | |
| "num_query_heads": 32, | |
| "hidden_dim": 3072, | |
| "intermediate_dim": 8192, | |
| "num_key_value_heads": 32, | |
| "layer_norm_epsilon": 1e-05, | |
| "dropout": 0.0, | |
| "max_sequence_length": 4096, | |
| "pretraining_sequence_length": 4096, | |
| "rope_max_wavelength": 10000.0, | |
| "rope_scaling_type": null, | |
| "rope_scaling_short_factor": null, | |
| "rope_scaling_long_factor": null | |
| }, | |
| "registered_name": "keras_nlp>Phi3Backbone" | |
| }, | |
| "preprocessor": { | |
| "module": "keras_nlp.src.models.phi3.phi3_causal_lm_preprocessor", | |
| "class_name": "Phi3CausalLMPreprocessor", | |
| "config": { | |
| "name": "phi3_causal_lm_preprocessor", | |
| "trainable": true, | |
| "dtype": { | |
| "module": "keras", | |
| "class_name": "DTypePolicy", | |
| "config": { | |
| "name": "mixed_bfloat16" | |
| }, | |
| "registered_name": null | |
| }, | |
| "tokenizer": { | |
| "module": "keras_nlp.src.models.phi3.phi3_tokenizer", | |
| "class_name": "Phi3Tokenizer", | |
| "config": { | |
| "name": "phi3_tokenizer", | |
| "trainable": true, | |
| "dtype": { | |
| "module": "keras", | |
| "class_name": "DTypePolicy", | |
| "config": { | |
| "name": "int32" | |
| }, | |
| "registered_name": null | |
| }, | |
| "proto": null, | |
| "sequence_length": null | |
| }, | |
| "registered_name": "keras_nlp>Phi3Tokenizer" | |
| }, | |
| "sequence_length": 1024, | |
| "add_start_token": true, | |
| "add_end_token": false | |
| }, | |
| "registered_name": "keras_nlp>Phi3CausalLMPreprocessor" | |
| }, | |
| "name": "phi3_causal_lm" | |
| }, | |
| "registered_name": "keras_nlp>Phi3CausalLM" | |
| } |