| { |
| "_name_or_path": "google/long-t5-tglobal-base", |
| "architectures": [ |
| "EffLongT5ForConditionalGeneration" |
| ], |
| "copy_skipped_hidden_states": false, |
| "d_ff": 2048, |
| "d_kv": 64, |
| "d_model": 768, |
| "decoder_start_token_id": 0, |
| "dense_act_fn": "gelu_new", |
| "distill_layer_alpha": null, |
| "do_layer_transformation": false, |
| "dropout_rate": 0.1, |
| "encoder_attention_type": "transient-global", |
| "eos_token_id": 1, |
| "exit_conf_threshold": 1.0, |
| "exit_conf_type": null, |
| "exit_position_temp": null, |
| "feed_forward_proj": "gated-gelu", |
| "global_block_size": 16, |
| "initializer_factor": 1.0, |
| "intermediate_loss_fn": "weighted_ce", |
| "is_encoder_decoder": true, |
| "is_gated_act": true, |
| "layer_norm_epsilon": 1e-06, |
| "local_radius": 127, |
| "max_answer_length": 512, |
| "model_type": "longt5", |
| "n_positions": 4096, |
| "num_decoder_layers": 12, |
| "num_heads": 12, |
| "num_layers": 12, |
| "output_hidden_states_decoder": true, |
| "output_past": true, |
| "pad_token_id": 0, |
| "parallel_causal_mask": true, |
| "parallel_gen_token": true, |
| "relative_attention_max_distance": 128, |
| "relative_attention_num_buckets": 32, |
| "rollback_conf_threshold": null, |
| "shallow2deep_conf_threshold": 1.0, |
| "shallow2deep_conf_type": null, |
| "shallow_exit_layer": null, |
| "static_exit_layer": null, |
| "tie_word_embeddings": false, |
| "tome_combining_method": null, |
| "tome_matching_algo": null, |
| "tome_metric": null, |
| "tome_partition_method": null, |
| "tome_prop_attn": true, |
| "tome_threshold": null, |
| "tome_trace_source": false, |
| "torch_dtype": "float32", |
| "train_meta_cm_head": false, |
| "transformers_version": "4.29.2", |
| "use_cache": true, |
| "use_early_exit": false, |
| "use_shallow_deep": false, |
| "use_token_merging": false, |
| "vocab_size": 32128 |
| } |
|
|
