| { |
| "architectures": [ |
| "SharedSpaceDecoderForCausalLM" |
| ], |
| "model_type": "shared_subspace_decoder", |
| "auto_map": { |
| "AutoConfig": "shared_space_config.SharedSpaceDecoderConfig", |
| "AutoModelForCausalLM": "task_heads.SharedSpaceDecoderForCausalLM" |
| }, |
| "attention_backend": "flash_attention_2", |
| "attention_bias": false, |
| "attention_dropout_prob": 0.1, |
| "bos_token_id": 50256, |
| "classifier_dropout": null, |
| "dtype": "float32", |
| "eos_token_id": 50256, |
| "ffn_decompose": false, |
| "ffn_rank": null, |
| "hidden_dropout_prob": 0.1, |
| "hidden_size": 768, |
| "initializer_range": 0.02, |
| "intermediate_size": 2048, |
| "kv_shared_dim": null, |
| "layer_norm_eps": 1e-12, |
| "max_position_embeddings": 1024, |
| "nope_dims": 32, |
| "norm_type": "rmsnorm", |
| "num_attention_heads": 12, |
| "num_dense_layers": 0, |
| "num_hidden_layers": 12, |
| "o_shared_dim": null, |
| "pad_token_id": 50256, |
| "q_shared_dim": null, |
| "qk_private_dim": 64, |
| "rms_norm_eps": 1e-06, |
| "rope_dims": 32, |
| "rope_scaling": { |
| "factor": 2.0, |
| "type": "linear" |
| }, |
| "rope_theta": 10000.0, |
| "transformers_version": "4.56.0", |
| "vo_private_dim": 64, |
| "vocab_rank": null, |
| "vocab_size": 50257, |
| "vocab_subspace": false |
| } |
