| { |
| "architectures": [ |
| "OlmoHybridForCausalLM" |
| ], |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "bos_token_id": 1, |
| "dtype": "float32", |
| "eos_token_id": 100257, |
| "hidden_act": "silu", |
| "hidden_size": 32, |
| "initializer_range": 0.02, |
| "intermediate_size": 32, |
| "is_decoder": true, |
| "layer_types": [ |
| "linear_attention", |
| "full_attention" |
| ], |
| "linear_a_log_max": 16.0, |
| "linear_a_log_min": 0.0, |
| "linear_allow_neg_eigval": false, |
| "linear_conv_kernel_dim": 4, |
| "linear_dt_init_floor": 0.0001, |
| "linear_dt_max": 0.1, |
| "linear_dt_min": 0.001, |
| "linear_key_head_dim": 8, |
| "linear_num_key_heads": 4, |
| "linear_num_value_heads": 4, |
| "linear_value_head_dim": 8, |
| "max_position_embeddings": 512, |
| "model_type": "olmo_hybrid", |
| "num_attention_heads": 2, |
| "num_hidden_layers": 2, |
| "num_key_value_heads": 2, |
| "pad_token_id": 100277, |
| "rms_norm_eps": 1e-06, |
| "rope_parameters": { |
| "rope_theta": 10000.0, |
| "rope_type": "default" |
| }, |
| "tie_word_embeddings": false, |
| "transformers_version": "5.16.0.dev0", |
| "use_cache": true, |
| "vocab_size": 100278 |
| } |
|
|
