Update mllama_audio_model.py

mllama_audio_model.py

@@ -2,7 +2,7 @@ from typing import Optional, Tuple, Union
 import torch
 from torch import nn
 from transformers.modeling_outputs import BaseModelOutput
-from transformers import Wav2Vec2Model, Wav2Vec2Config, MllamaPreTrainedModel
+from transformers import Wav2Vec2Model, Wav2Vec2Config, MllamaPreTrainedModel, Wav2Vec2BertConfig
 from transformers.models.wav2vec2_bert.modeling_wav2vec2_bert import Wav2Vec2BertAdapterLayer
 from .configuration_llama3 import Llama3Config
 
@@ -14,7 +14,7 @@ class AudioAdapter(nn.Module):
             self.proj = nn.Linear(config.hidden_size, config.output_hidden_size)
         else:
             self.proj = None
-        self.layers = nn.ModuleList(Wav2Vec2BertAdapterLayer(config) for _ in range(config.num_adapter_layers))
+        self.layers = nn.ModuleList(Wav2Vec2BertAdapterLayer(Wav2Vec2BertConfig(adapter_kernel_size=config.adapter_kernel_size, adapter_stride=config.adapter_stride)) for _ in range(config.num_adapter_layers))
 
         self.kernel_size = config.adapter_kernel_size
         self.stride = config.adapter_stride