| { | |
| "_name_or_path": "/home/azureuser/phi4/qwen_works/Speech-to-Text-Training/MODELS/Qwen-Nahin-3-6/checkpoint-201000", | |
| "architectures": [ | |
| "Qwen2MMForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "audio_processor": { | |
| "config": { | |
| "activation": "swish", | |
| "activation_checkpointing": { | |
| "interval": 1, | |
| "module": "transformer", | |
| "offload": false | |
| }, | |
| "attention_dim": 1024, | |
| "attention_heads": 16, | |
| "batch_norm": false, | |
| "bias_in_glu": true, | |
| "causal": true, | |
| "chunk_size": -1, | |
| "cnn_layer_norm": true, | |
| "conv_activation": "swish", | |
| "conv_glu_type": "swish", | |
| "depthwise_multiplier": 1, | |
| "depthwise_seperable_out_channel": 1024, | |
| "dropout_rate": 0.0, | |
| "encoder_embedding_config": { | |
| "input_size": 80 | |
| }, | |
| "ext_pw_kernel_size": 1, | |
| "ext_pw_out_channel": 1024, | |
| "input_layer": "nemo_conv", | |
| "input_size": 80, | |
| "kernel_size": 3, | |
| "left_chunk": 18, | |
| "linear_units": 1536, | |
| "nemo_conv_settings": { | |
| "conv_channels": 1024 | |
| }, | |
| "num_blocks": 24, | |
| "relative_attention_bias_args": { | |
| "t5_bias_max_distance": 500, | |
| "type": "t5" | |
| }, | |
| "time_reduction": 8 | |
| }, | |
| "name": "cascades" | |
| }, | |
| "auto_map": { | |
| "AutoConfig": "configuration_qwen2mm.Qwen2MMConfig", | |
| "AutoModelForCausalLM": "modeling_phi4mm.Qwen2MMForCausalLM", | |
| "AutoTokenizer": "./" | |
| }, | |
| "bos_token_id": 151644, | |
| "embd_layer": { | |
| "audio_embd_layer": { | |
| "compression_rate": 8, | |
| "downsample_rate": 1, | |
| "embedding_cls": "audio", | |
| "enable_gradient_checkpointing": true, | |
| "projection_cls": "mlp", | |
| "use_conv_downsample": false, | |
| "use_qformer": false | |
| }, | |
| "embedding_cls": "image_audio", | |
| "image_embd_layer": { | |
| "crop_size": 448, | |
| "embedding_cls": "tune_image", | |
| "enable_gradient_checkpointing": true, | |
| "hd_transform_order": "sub_glb", | |
| "image_token_compression_cls": "avg_pool_2d", | |
| "projection_cls": "mlp", | |
| "use_hd_transform": true, | |
| "with_learnable_separator": true | |
| } | |
| }, | |
| "eos_token_id": 151645, | |
| "hidden_act": "silu", | |
| "hidden_size": 896, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 4864, | |
| "max_position_embeddings": 131072, | |
| "max_window_layers": 24, | |
| "model_type": "qwen2-mm", | |
| "num_attention_heads": 14, | |
| "num_hidden_layers": 24, | |
| "num_key_value_heads": 2, | |
| "pad_token_id": 151643, | |
| "rms_norm_eps": 1e-06, | |
| "rope_scaling": null, | |
| "rope_theta": 1000000.0, | |
| "sliding_window": null, | |
| "speech_lora": { | |
| "dp": 0.01, | |
| "layer": "((layers.*self_attn\\.(qkv|o)_proj)|(layers.*mlp\\.(gate_up|down)_proj))", | |
| "lora_alpha": 640, | |
| "r": 320 | |
| }, | |
| "tie_word_embeddings": true, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.48.3", | |
| "use_cache": false, | |
| "use_sliding_window": false, | |
| "vision_lora": { | |
| "dp": 0.0, | |
| "layer": "layers.*((self_attn\\.(qkv_proj|o_proj))|(mlp\\.(gate_up|down)_proj))", | |
| "lora_alpha": 512, | |
| "r": 256 | |
| }, | |
| "vocab_size": 194498 | |
| } | |