{ "architectures": [ "Glm4vMoeForConditionalGeneration" ], "dtype": "bfloat16", "image_end_token_id": 151340, "image_start_token_id": 151339, "image_token_id": 151363, "model_type": "glm4v_moe", "quantization_config": { "config_groups": { "group_0": { "format": "pack-quantized", "input_activations": null, "output_activations": null, "targets": [ "Linear" ], "weights": { "actorder": null, "block_structure": null, "dynamic": false, "group_size": 32, "num_bits": 4, "observer": "mse", "observer_kwargs": {}, "strategy": "group", "symmetric": true, "type": "int" } } }, "format": "pack-quantized", "global_compression_ratio": null, "ignore": [ "model.visual.blocks.0.attn.qkv", "model.visual.blocks.0.attn.proj", "model.visual.blocks.0.mlp.gate_proj", "model.visual.blocks.0.mlp.up_proj", "model.visual.blocks.0.mlp.down_proj", "model.visual.blocks.1.attn.qkv", "model.visual.blocks.1.attn.proj", "model.visual.blocks.1.mlp.gate_proj", "model.visual.blocks.1.mlp.up_proj", "model.visual.blocks.1.mlp.down_proj", "model.visual.blocks.2.attn.qkv", "model.visual.blocks.2.attn.proj", "model.visual.blocks.2.mlp.gate_proj", "model.visual.blocks.2.mlp.up_proj", "model.visual.blocks.2.mlp.down_proj", "model.visual.blocks.3.attn.qkv", "model.visual.blocks.3.attn.proj", "model.visual.blocks.3.mlp.gate_proj", "model.visual.blocks.3.mlp.up_proj", "model.visual.blocks.3.mlp.down_proj", "model.visual.blocks.4.attn.qkv", "model.visual.blocks.4.attn.proj", "model.visual.blocks.4.mlp.gate_proj", "model.visual.blocks.4.mlp.up_proj", "model.visual.blocks.4.mlp.down_proj", "model.visual.blocks.5.attn.qkv", "model.visual.blocks.5.attn.proj", "model.visual.blocks.5.mlp.gate_proj", "model.visual.blocks.5.mlp.up_proj", "model.visual.blocks.5.mlp.down_proj", "model.visual.blocks.6.attn.qkv", "model.visual.blocks.6.attn.proj", "model.visual.blocks.6.mlp.gate_proj", "model.visual.blocks.6.mlp.up_proj", "model.visual.blocks.6.mlp.down_proj", "model.visual.blocks.7.attn.qkv", "model.visual.blocks.7.attn.proj", "model.visual.blocks.7.mlp.gate_proj", "model.visual.blocks.7.mlp.up_proj", "model.visual.blocks.7.mlp.down_proj", "model.visual.blocks.8.attn.qkv", "model.visual.blocks.8.attn.proj", "model.visual.blocks.8.mlp.gate_proj", "model.visual.blocks.8.mlp.up_proj", "model.visual.blocks.8.mlp.down_proj", "model.visual.blocks.9.attn.qkv", "model.visual.blocks.9.attn.proj", "model.visual.blocks.9.mlp.gate_proj", "model.visual.blocks.9.mlp.up_proj", "model.visual.blocks.9.mlp.down_proj", "model.visual.blocks.10.attn.qkv", "model.visual.blocks.10.attn.proj", "model.visual.blocks.10.mlp.gate_proj", "model.visual.blocks.10.mlp.up_proj", "model.visual.blocks.10.mlp.down_proj", "model.visual.blocks.11.attn.qkv", "model.visual.blocks.11.attn.proj", "model.visual.blocks.11.mlp.gate_proj", "model.visual.blocks.11.mlp.up_proj", "model.visual.blocks.11.mlp.down_proj", "model.visual.blocks.12.attn.qkv", "model.visual.blocks.12.attn.proj", "model.visual.blocks.12.mlp.gate_proj", "model.visual.blocks.12.mlp.up_proj", "model.visual.blocks.12.mlp.down_proj", "model.visual.blocks.13.attn.qkv", "model.visual.blocks.13.attn.proj", "model.visual.blocks.13.mlp.gate_proj", "model.visual.blocks.13.mlp.up_proj", "model.visual.blocks.13.mlp.down_proj", "model.visual.blocks.14.attn.qkv", "model.visual.blocks.14.attn.proj", "model.visual.blocks.14.mlp.gate_proj", "model.visual.blocks.14.mlp.up_proj", "model.visual.blocks.14.mlp.down_proj", "model.visual.blocks.15.attn.qkv", "model.visual.blocks.15.attn.proj", "model.visual.blocks.15.mlp.gate_proj", "model.visual.blocks.15.mlp.up_proj", "model.visual.blocks.15.mlp.down_proj", "model.visual.blocks.16.attn.qkv", "model.visual.blocks.16.attn.proj", "model.visual.blocks.16.mlp.gate_proj", "model.visual.blocks.16.mlp.up_proj", "model.visual.blocks.16.mlp.down_proj", "model.visual.blocks.17.attn.qkv", "model.visual.blocks.17.attn.proj", "model.visual.blocks.17.mlp.gate_proj", "model.visual.blocks.17.mlp.up_proj", "model.visual.blocks.17.mlp.down_proj", "model.visual.blocks.18.attn.qkv", "model.visual.blocks.18.attn.proj", "model.visual.blocks.18.mlp.gate_proj", "model.visual.blocks.18.mlp.up_proj", "model.visual.blocks.18.mlp.down_proj", "model.visual.blocks.19.attn.qkv", "model.visual.blocks.19.attn.proj", "model.visual.blocks.19.mlp.gate_proj", "model.visual.blocks.19.mlp.up_proj", "model.visual.blocks.19.mlp.down_proj", "model.visual.blocks.20.attn.qkv", "model.visual.blocks.20.attn.proj", "model.visual.blocks.20.mlp.gate_proj", "model.visual.blocks.20.mlp.up_proj", "model.visual.blocks.20.mlp.down_proj", "model.visual.blocks.21.attn.qkv", "model.visual.blocks.21.attn.proj", "model.visual.blocks.21.mlp.gate_proj", "model.visual.blocks.21.mlp.up_proj", "model.visual.blocks.21.mlp.down_proj", "model.visual.blocks.22.attn.qkv", "model.visual.blocks.22.attn.proj", "model.visual.blocks.22.mlp.gate_proj", "model.visual.blocks.22.mlp.up_proj", "model.visual.blocks.22.mlp.down_proj", "model.visual.blocks.23.attn.qkv", "model.visual.blocks.23.attn.proj", "model.visual.blocks.23.mlp.gate_proj", "model.visual.blocks.23.mlp.up_proj", "model.visual.blocks.23.mlp.down_proj", "model.visual.merger.proj", "model.visual.merger.gate_proj", "model.visual.merger.up_proj", "model.visual.merger.down_proj", "model.language_model.layers.0.self_attn.q_proj", "model.language_model.layers.0.self_attn.k_proj", "model.language_model.layers.0.self_attn.v_proj", "model.language_model.layers.0.self_attn.o_proj", "model.language_model.layers.0.mlp.gate_proj", "model.language_model.layers.0.mlp.up_proj", "model.language_model.layers.0.mlp.down_proj", "model.language_model.layers.1.mlp.shared_experts.gate_proj", "model.language_model.layers.1.mlp.shared_experts.up_proj", "model.language_model.layers.1.mlp.shared_experts.down_proj", "model.language_model.layers.2.mlp.shared_experts.gate_proj", "model.language_model.layers.2.mlp.shared_experts.up_proj", "model.language_model.layers.2.mlp.shared_experts.down_proj", "model.language_model.layers.3.mlp.shared_experts.gate_proj", "model.language_model.layers.3.mlp.shared_experts.up_proj", "model.language_model.layers.3.mlp.shared_experts.down_proj", "model.language_model.layers.4.mlp.shared_experts.gate_proj", "model.language_model.layers.4.mlp.shared_experts.up_proj", "model.language_model.layers.4.mlp.shared_experts.down_proj", "model.language_model.layers.5.mlp.shared_experts.gate_proj", "model.language_model.layers.5.mlp.shared_experts.up_proj", "model.language_model.layers.5.mlp.shared_experts.down_proj", "model.language_model.layers.6.mlp.shared_experts.gate_proj", "model.language_model.layers.6.mlp.shared_experts.up_proj", "model.language_model.layers.6.mlp.shared_experts.down_proj", "model.language_model.layers.7.mlp.shared_experts.gate_proj", "model.language_model.layers.7.mlp.shared_experts.up_proj", "model.language_model.layers.7.mlp.shared_experts.down_proj", "model.language_model.layers.8.mlp.shared_experts.gate_proj", "model.language_model.layers.8.mlp.shared_experts.up_proj", "model.language_model.layers.8.mlp.shared_experts.down_proj", "model.language_model.layers.9.mlp.shared_experts.gate_proj", "model.language_model.layers.9.mlp.shared_experts.up_proj", "model.language_model.layers.9.mlp.shared_experts.down_proj", "model.language_model.layers.10.mlp.shared_experts.gate_proj", "model.language_model.layers.10.mlp.shared_experts.up_proj", "model.language_model.layers.10.mlp.shared_experts.down_proj", "model.language_model.layers.11.mlp.shared_experts.gate_proj", "model.language_model.layers.11.mlp.shared_experts.up_proj", "model.language_model.layers.11.mlp.shared_experts.down_proj", "model.language_model.layers.12.mlp.shared_experts.gate_proj", "model.language_model.layers.12.mlp.shared_experts.up_proj", "model.language_model.layers.12.mlp.shared_experts.down_proj", "model.language_model.layers.13.mlp.shared_experts.gate_proj", "model.language_model.layers.13.mlp.shared_experts.up_proj", "model.language_model.layers.13.mlp.shared_experts.down_proj", "model.language_model.layers.14.mlp.shared_experts.gate_proj", "model.language_model.layers.14.mlp.shared_experts.up_proj", "model.language_model.layers.14.mlp.shared_experts.down_proj", "model.language_model.layers.15.mlp.shared_experts.gate_proj", "model.language_model.layers.15.mlp.shared_experts.up_proj", "model.language_model.layers.15.mlp.shared_experts.down_proj", "model.language_model.layers.16.mlp.shared_experts.gate_proj", "model.language_model.layers.16.mlp.shared_experts.up_proj", "model.language_model.layers.16.mlp.shared_experts.down_proj", "model.language_model.layers.17.mlp.shared_experts.gate_proj", "model.language_model.layers.17.mlp.shared_experts.up_proj", "model.language_model.layers.17.mlp.shared_experts.down_proj", "model.language_model.layers.18.mlp.shared_experts.gate_proj", "model.language_model.layers.18.mlp.shared_experts.up_proj", "model.language_model.layers.18.mlp.shared_experts.down_proj", "model.language_model.layers.19.mlp.shared_experts.gate_proj", "model.language_model.layers.19.mlp.shared_experts.up_proj", "model.language_model.layers.19.mlp.shared_experts.down_proj", "model.language_model.layers.20.mlp.shared_experts.gate_proj", "model.language_model.layers.20.mlp.shared_experts.up_proj", "model.language_model.layers.20.mlp.shared_experts.down_proj", "model.language_model.layers.21.mlp.shared_experts.gate_proj", "model.language_model.layers.21.mlp.shared_experts.up_proj", "model.language_model.layers.21.mlp.shared_experts.down_proj", "model.language_model.layers.22.mlp.shared_experts.gate_proj", "model.language_model.layers.22.mlp.shared_experts.up_proj", "model.language_model.layers.22.mlp.shared_experts.down_proj", "model.language_model.layers.23.mlp.shared_experts.gate_proj", "model.language_model.layers.23.mlp.shared_experts.up_proj", "model.language_model.layers.23.mlp.shared_experts.down_proj", "model.language_model.layers.24.mlp.shared_experts.gate_proj", "model.language_model.layers.24.mlp.shared_experts.up_proj", "model.language_model.layers.24.mlp.shared_experts.down_proj", "model.language_model.layers.25.mlp.shared_experts.gate_proj", "model.language_model.layers.25.mlp.shared_experts.up_proj", "model.language_model.layers.25.mlp.shared_experts.down_proj", "model.language_model.layers.26.mlp.shared_experts.gate_proj", "model.language_model.layers.26.mlp.shared_experts.up_proj", "model.language_model.layers.26.mlp.shared_experts.down_proj", "model.language_model.layers.27.mlp.shared_experts.gate_proj", "model.language_model.layers.27.mlp.shared_experts.up_proj", "model.language_model.layers.27.mlp.shared_experts.down_proj", "model.language_model.layers.28.mlp.shared_experts.gate_proj", "model.language_model.layers.28.mlp.shared_experts.up_proj", "model.language_model.layers.28.mlp.shared_experts.down_proj", "model.language_model.layers.29.mlp.shared_experts.gate_proj", "model.language_model.layers.29.mlp.shared_experts.up_proj", "model.language_model.layers.29.mlp.shared_experts.down_proj", "model.language_model.layers.30.mlp.shared_experts.gate_proj", "model.language_model.layers.30.mlp.shared_experts.up_proj", "model.language_model.layers.30.mlp.shared_experts.down_proj", "model.language_model.layers.31.mlp.shared_experts.gate_proj", "model.language_model.layers.31.mlp.shared_experts.up_proj", "model.language_model.layers.31.mlp.shared_experts.down_proj", "model.language_model.layers.32.mlp.shared_experts.gate_proj", "model.language_model.layers.32.mlp.shared_experts.up_proj", "model.language_model.layers.32.mlp.shared_experts.down_proj", "model.language_model.layers.33.mlp.shared_experts.gate_proj", "model.language_model.layers.33.mlp.shared_experts.up_proj", "model.language_model.layers.33.mlp.shared_experts.down_proj", "model.language_model.layers.34.mlp.shared_experts.gate_proj", "model.language_model.layers.34.mlp.shared_experts.up_proj", "model.language_model.layers.34.mlp.shared_experts.down_proj", "model.language_model.layers.35.mlp.shared_experts.gate_proj", "model.language_model.layers.35.mlp.shared_experts.up_proj", "model.language_model.layers.35.mlp.shared_experts.down_proj", "model.language_model.layers.36.mlp.shared_experts.gate_proj", "model.language_model.layers.36.mlp.shared_experts.up_proj", "model.language_model.layers.36.mlp.shared_experts.down_proj", "model.language_model.layers.37.mlp.shared_experts.gate_proj", "model.language_model.layers.37.mlp.shared_experts.up_proj", "model.language_model.layers.37.mlp.shared_experts.down_proj", "model.language_model.layers.38.mlp.shared_experts.gate_proj", "model.language_model.layers.38.mlp.shared_experts.up_proj", "model.language_model.layers.38.mlp.shared_experts.down_proj", "model.language_model.layers.39.mlp.shared_experts.gate_proj", "model.language_model.layers.39.mlp.shared_experts.up_proj", "model.language_model.layers.39.mlp.shared_experts.down_proj", "model.language_model.layers.40.mlp.shared_experts.gate_proj", "model.language_model.layers.40.mlp.shared_experts.up_proj", "model.language_model.layers.40.mlp.shared_experts.down_proj", "model.language_model.layers.41.mlp.shared_experts.gate_proj", "model.language_model.layers.41.mlp.shared_experts.up_proj", "model.language_model.layers.41.mlp.shared_experts.down_proj", "model.language_model.layers.42.mlp.shared_experts.gate_proj", "model.language_model.layers.42.mlp.shared_experts.up_proj", "model.language_model.layers.42.mlp.shared_experts.down_proj", "model.language_model.layers.43.mlp.shared_experts.gate_proj", "model.language_model.layers.43.mlp.shared_experts.up_proj", "model.language_model.layers.43.mlp.shared_experts.down_proj", "model.language_model.layers.44.mlp.shared_experts.gate_proj", "model.language_model.layers.44.mlp.shared_experts.up_proj", "model.language_model.layers.44.mlp.shared_experts.down_proj", "model.language_model.layers.45.mlp.shared_experts.gate_proj", "model.language_model.layers.45.mlp.shared_experts.up_proj", "model.language_model.layers.45.mlp.shared_experts.down_proj", "lm_head" ], "kv_cache_scheme": null, "quant_method": "compressed-tensors", "quantization_status": "compressed", "sparsity_config": {}, "transform_config": {}, "version": "0.11.1.a20250828" }, "text_config": { "attention_bias": true, "attention_dropout": 0.0, "dtype": "bfloat16", "eos_token_id": [ 151329, 151336, 151338 ], "first_k_dense_replace": 1, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "image_end_token_id": 151340, "image_start_token_id": 151339, "image_token_id": 151363, "initializer_range": 0.02, "intermediate_size": 10944, "max_position_embeddings": 65536, "model_type": "Glm4vMoe_text", "moe_intermediate_size": 1408, "n_group": 1, "n_routed_experts": 128, "n_shared_experts": 1, "norm_topk_prob": true, "num_attention_heads": 96, "num_experts_per_tok": 8, "num_hidden_layers": 46, "num_key_value_heads": 8, "pad_token_id": 151329, "partial_rotary_factor": 0.5, "rms_norm_eps": 1e-05, "rope_scaling": { "mrope_section": [ 8, 12, 12 ], "rope_type": "default" }, "rope_theta": 10000.0, "routed_scaling_factor": 1.0, "topk_group": 1, "use_cache": true, "use_qk_norm": false, "vocab_size": 151552 }, "tie_word_embeddings": false, "transformers_version": "4.56.0.dev0", "video_end_token_id": 151342, "video_start_token_id": 151341, "video_token_id": 151364, "vision_config": { "attention_bias": false, "attention_dropout": 0.0, "depth": 24, "hidden_act": "silu", "hidden_size": 1536, "image_size": 336, "in_channels": 3, "initializer_range": 0.02, "intermediate_size": 10944, "model_type": "glm4v_moe", "num_heads": 12, "out_hidden_size": 4096, "patch_size": 14, "rms_norm_eps": 1e-05, "spatial_merge_size": 2, "temporal_patch_size": 2 } }