| { |
| "alpha_entropy": 0.1, |
| "alpha_sparsity": 0.1, |
| "architectures": [ |
| "PULSAR" |
| ], |
| "attention_probs_dropout_prob": 0.1, |
| "cell_state_size": 256, |
| "classifier_dropout": null, |
| "cls_transform": false, |
| "contrastive_loss_weight": 1.0, |
| "coordinated_masking_program_path": null, |
| "decoder_intermediate_size": 4096, |
| "decoder_loop_num": 1, |
| "decoder_num_attention_heads": 8, |
| "decoder_num_hidden_layers": 6, |
| "dtype": "float32", |
| "element_mask_ratio": 0.85, |
| "element_mask_strategy": "correlated_program", |
| "encoder_intermediate_size": 4096, |
| "encoder_num_attention_heads": 8, |
| "encoder_num_hidden_layers": 8, |
| "encoder_ratio": 0.15, |
| "expansion_factor": 4, |
| "feed_cls_only": true, |
| "frozen": false, |
| "hidden_act": "gelu", |
| "hidden_dropout_prob": 0.1, |
| "hidden_size": 512, |
| "initializer_range": 0.02, |
| "input_size": 1280, |
| "layer_norm_eps": 1e-12, |
| "mask_hidden_size": 512, |
| "model_name": null, |
| "model_type": "PULSAR", |
| "noise_norm": false, |
| "num_masks": 10, |
| "pca_model_path": null, |
| "recon_loss_weight": 0.1, |
| "resample_decoding": false, |
| "seq_length": 1024, |
| "transformers_version": "4.57.0", |
| "use_cache": true, |
| "use_cell_state": false, |
| "use_decoder": true |
| } |
|
|
