| { |
| "transformers_version": "5.9.0", |
| "architectures": null, |
| "output_hidden_states": false, |
| "return_dict": true, |
| "dtype": null, |
| "chunk_size_feed_forward": 0, |
| "is_encoder_decoder": false, |
| "id2label": { |
| "0": "LABEL_0", |
| "1": "LABEL_1" |
| }, |
| "label2id": { |
| "LABEL_0": 0, |
| "LABEL_1": 1 |
| }, |
| "problem_type": null, |
| "vocab_size": 1024, |
| "hidden_size": 160, |
| "intermediate_size": 640, |
| "num_hidden_layers": 6, |
| "num_attention_heads": 5, |
| "num_key_value_heads": 1, |
| "head_dim": 32, |
| "hidden_activation": "gelu_pytorch_tanh", |
| "max_position_embeddings": 1024, |
| "initializer_range": 0.02, |
| "rms_norm_eps": 1e-06, |
| "use_cache": true, |
| "pad_token_id": 2, |
| "eos_token_id": 1, |
| "bos_token_id": 0, |
| "tie_word_embeddings": true, |
| "rope_parameters": { |
| "sliding_attention": { |
| "rope_type": "default", |
| "rope_theta": 10000.0 |
| }, |
| "full_attention": { |
| "rope_type": "proportional", |
| "partial_rotary_factor": 0.25, |
| "rope_theta": 1000000.0 |
| } |
| }, |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "sliding_window": 128, |
| "layer_types": [ |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention" |
| ], |
| "final_logit_softcapping": null, |
| "use_bidirectional_attention": null, |
| "vocab_size_per_layer_input": 1024, |
| "hidden_size_per_layer_input": 24, |
| "num_global_key_value_heads": 1, |
| "global_head_dim": 32, |
| "attention_k_eq_v": false, |
| "num_kv_shared_layers": 0, |
| "enable_moe_block": false, |
| "use_double_wide_mlp": false, |
| "num_experts": null, |
| "top_k_experts": null, |
| "moe_intermediate_size": null, |
| "_name_or_path": "", |
| "model_type": "gemma4_text", |
| "output_attentions": false |
| } |
