{ "architectures": [ "GlmOcrForConditionalGeneration" ], "model_type": "glm_ocr", "text_config": { "model_type": "glm_ocr_text", "pad_token_id": 59246, "vocab_size": 59392, "eos_token_id": [ 59246, 59253 ], "attention_bias": false, "attention_dropout": 0.0, "head_dim": 128, "hidden_act": "silu", "hidden_size": 1536, "initializer_range": 0.02, "intermediate_size": 4608, "max_position_embeddings": 131072, "num_attention_heads": 16, "num_hidden_layers": 16, "num_nextn_predict_layers": 1, "num_key_value_heads": 8, "rms_norm_eps": 1e-05, "dtype": "bfloat16", "rope_parameters": { "rope_type": "default", "mrope_section": [ 16, 24, 24 ], "partial_rotary_factor": 1.0, "rope_theta": 10000 }, "tie_word_embeddings": false, "use_cache": true }, "vision_config": { "model_type": "glm_ocr_vision", "hidden_size": 1024, "depth": 24, "num_heads": 16, "attention_bias": true, "intermediate_size": 4096, "hidden_act": "silu", "hidden_dropout_prob": 0.0, "initializer_range": 0.02, "image_size": 336, "patch_size": 14, "out_hidden_size": 1536, "rms_norm_eps": 1e-05, "spatial_merge_size": 2, "temporal_patch_size": 2 }, "image_start_token_id": 59256, "image_end_token_id": 59257, "video_start_token_id": 59258, "video_end_token_id": 59259, "image_token_id": 59280, "video_token_id": 59281, "transformers_version": "5.0.1dev0" }