{ "architectures": [ "GlmAsrModel" ], "audio_config": { "attention_dropout": 0.0, "hidden_act": "gelu", "hidden_size": 32, "initializer_range": 0.02, "intermediate_size": 32, "max_position_embeddings": 512, "model_type": "glmasr_encoder", "num_attention_heads": 2, "num_hidden_layers": 2, "num_key_value_heads": 2, "num_mel_bins": 80, "partial_rotary_factor": 0.5, "rope_parameters": { "partial_rotary_factor": 0.5, "rope_theta": 10000.0, "rope_type": "default" } }, "audio_token_id": 0, "dtype": "float32", "model_type": "glmasr", "pad_token_id": 59246, "projector_hidden_act": "gelu", "text_config": { "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 1, "eos_token_id": 59246, "head_dim": 8, "hidden_act": "gelu", "hidden_size": 32, "initializer_range": 0.02, "intermediate_size": 32, "max_position_embeddings": 512, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 2, "num_hidden_layers": 2, "num_key_value_heads": 2, "pad_token_id": 59246, "pretraining_tp": 1, "rms_norm_eps": 1e-06, "rope_parameters": { "rope_theta": 10000.0, "rope_type": "default" }, "tie_word_embeddings": false, "use_cache": true, "vocab_size": 59263 }, "tie_word_embeddings": true, "transformers_version": "5.16.0.dev0" }