{ "architectures": [ "MolmoForCausalLM" ], "auto_map": { "AutoConfig": "config_molmo.MolmoConfig", "AutoModelForCausalLM": "modeling_molmo.MolmoForCausalLM" }, "model_type": "molmo", "quantization": { "group_size": 64, "bits": 3 }, "text_config": { "model_type": "molmo", "max_position_embeddings": 4096, "d_model": 3584, "n_heads": 28, "n_kv_heads": 4, "n_layers": 28, "mlp_ratio": 4, "max_sequence_length": 1024, "act_output_multiplier": 0.5, "mlp_hidden_size": 37888, "vocab_size": 152064, "embedding_size": 152064, "additional_vocab_size": 128, "attention_dropout": 0.1, "residual_dropout": 0.1, "embedding_dropout": 0.1, "layer_norm_eps": 1e-05, "initializer_range": 0.02, "pad_token_id": -1, "rope": true, "rope_theta": 1000000.0, "weight_tying": false, "rope_full_precision": true, "rope_impl": "interleave" }, "vision_config": { "model_type": "molmo", "num_channels": 3, "image_default_input_size": [ 336, 336 ], "image_patch_size": 14, "image_pos_patch_size": 14, "hidden_size": 18944, "image_emb_dim": 1024, "image_num_heads": 16, "image_num_key_value_heads": 16, "image_num_layers": 23, "image_head_dim": 64, "image_mlp_dim": 4096, "image_mlp_activations": "gelu", "image_dropout_rate": 0.0, "image_num_pos": 577, "image_norm_eps": 1e-05, "attention_dropout": 0.0, "residual_dropout": 0.0, "initializer_range": 0.02, "d_model": 3584, "image_pooling_h": 2, "image_pooling_w": 2, "vit_layers": [ -2, -9 ], "image_pooling_2d": "attention-meanq", "image_padding_embed": "pad_and_partial_pad", "intermediate_size": 588, "skip_vision_non_divisible": true }, "vocab_size": 152064 }