| { | |
| "adapter_reduction": 16, | |
| "architectures": [ | |
| "DistillationWrapper" | |
| ], | |
| "attention_probs_dropout_prob": 0.1, | |
| "classifier_dropout": null, | |
| "dtype": "float32", | |
| "embedding_size": 128, | |
| "expert_intermediate_size": 2624, | |
| "group_depth": 4, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.1, | |
| "hidden_size": 1024, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 2624, | |
| "layer_norm_eps": 1e-06, | |
| "load_balancing_loss_coef": 0.2, | |
| "lora_alpha": 32, | |
| "lora_rank": 16, | |
| "max_position_embeddings": 8192, | |
| "model_type": "ModernALBERT", | |
| "num_attention_heads": 16, | |
| "num_expert_modules": 3, | |
| "num_experts": 8, | |
| "num_hidden_layers": 16, | |
| "output_hidden_states": true, | |
| "pad_token_id": 0, | |
| "router_jitter_noise": 0.01, | |
| "top_k": 2, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.51.3", | |
| "use_adapter": true, | |
| "use_cache": true, | |
| "use_moa": true, | |
| "vocab_size": 50368 | |
| } | |