{ "_attn_implementation_autoset": true, "architectures": [ "LlamaForCausalLM" ], "attention_bias": true, "attention_dropout": 0.0, "auto_map": { "AutoConfig": "config.NextStepConfig", "AutoModel": "models/nextstep_model.NextStep" }, "base_image_grid_size": 64, "boi": 151667, "bos_token_id": 151643, "create_kwargs": { "snr_type": "lognorm" }, "dtype": "float32", "eoi": 151668, "eos_token_id": 151643, "fm_head_batch_mul": 4, "fm_head_dim": 1536, "fm_head_layers": 12, "genloss_batch_mul": 4, "genloss_depth": 12, "genloss_net_arch": "mlp", "genloss_num_sampling_steps": "100", "genloss_type": "transport", "genloss_width": 1536, "head_dim": 128, "hidden_act": "silu", "hidden_size": 5120, "im_loss_weight": 1.0, "image_decoder_arch": "Trans_E", "image_encoder_name": null, "image_feature_layer": -2, "image_loss_weight": 1.0, "image_placeholder_id": 151669, "image_size": 64, "initializer_range": 0.02, "intermediate_size": 13824, "latent_channels": 16, "latent_patch_size": 2, "latent_size": 32, "lm_loss_weight": 0.01, "max_position_embeddings": 131072, "max_window_layers": 48, "mlp_bias": false, "model_type": "nextstep", "noise_strength": 0.0, "num_attention_heads": 40, "num_channels": 16, "num_hidden_layers": 48, "num_key_value_heads": 8, "o_attention_bias": false, "pad_token_id_added": 151665, "patch_size": 2, "pretraining_tp": 1, "quantization_config": { "autoround_version": "0.13.0", "batch_size": 1, "bits": 4, "block_name_to_quantize": "layers,image_head.net.res_blocks", "data_type": "int", "gradient_accumulate_steps": 8, "group_size": 128, "packing_format": "auto_round:auto_gptq", "quant_method": "auto-round", "sym": true }, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": 131072, "tie_word_embeddings": false, "transformers_version": "4.57.6", "use_2d_rope": false, "use_cache": true, "use_gen_pos_embed": false, "use_mlp_before_lm_head": false, "use_sliding_window": false, "use_token_length_weight": false, "vae_name_or_path": "vae/", "vocab_size": 152064 }