models/clap_encoder.py

import torch
import torch.nn as nn
from models.CLAP.open_clip import create_model
from transformers import RobertaTokenizer
# removed: import random, torchaudio, get_audio_features — only used by _get_audio_embed (audio modality), not needed for text inference


class CLAP_Encoder(nn.Module):
    def __init__(
        self,
        pretrained_path='checkpoint/music_speech_audioset_epoch_15_esc_89.98.pt',
        amodel = "HTSAT-base",
    ):
        super().__init__()
        self.device = "cpu"
        self.precision = "fp32"
        self.amodel = amodel  # or 'PANN-14'
        self.tmodel = "roberta"  # the best text encoder in our training
        self.enable_fusion = False  # False if you do not want to use the fusion model
        self.fusion_type = "aff_2d"
        self.pretrained = pretrained_path
        # removed: self.sampling_rate — only used by _get_audio_embed
        self.tokenize = RobertaTokenizer.from_pretrained("roberta-base")
        
        self.model, self.model_cfg = create_model(
            self.amodel,
            self.tmodel,
            self.pretrained,
            precision=self.precision,
            device=self.device,
            enable_fusion=self.enable_fusion,
            fusion_type=self.fusion_type,
        )

        for p in self.model.parameters():
            p.requires_grad = False

        self.model.eval()
        self.encoder_type = 'CLAP'

    # removed: batch_to_list, _get_audio_embed — audio modality not used in inference

    def _get_text_embed(self, batch):
        double_batch = False
        if len(batch) == 1:
            batch = batch * 2
            double_batch = True
        with torch.no_grad():
            # the 'fusion' truncate mode can be changed to 'rand_trunc' if run in unfusion mode
            text_data = self.tokenizer(batch)
            embed = self.model.get_text_embedding(text_data)
        if double_batch:
            embed = embed[0].unsqueeze(0)
        
        return embed.detach()


    def get_query_embed(self, modality, audio=None, text=None, use_text_ratio=0.5, device=None):
        # removed: audio and hybird modality branches — only text modality used in inference
        if modality == 'text':
            embed = self._get_text_embed(text)
        else:
            raise NotImplementedError("Please check flag 'training_modality'.")

        return embed.float()

    def tokenizer(self, text):
        result = self.tokenize(
            text,
            padding="max_length",
            truncation=True,
            max_length=512,
            return_tensors="pt",
        )
        return {k: v.squeeze(0) for k, v in result.items()}