提交

kongdeqiang

2026-03-13 28ccfbfc51068a663a80764e14074df5edf2b5ba

 funasr/models/rwkv_bat/rwkv_encoder.py

@@ -1,17 +1,20 @@
"""RWKV encoder definition for Transducer models."""

import math
from typing import Dict, List, Optional, Tuple
#!/usr/bin/env python3
# -*- encoding: utf-8 -*-
# Copyright FunASR (https://github.com/alibaba-damo-academy/FunASR). All Rights Reserved.
#  MIT License  (https://opensource.org/licenses/MIT)

import torch
from typing import Dict, List, Optional, Tuple

from funasr.models.encoder.abs_encoder import AbsEncoder
from funasr.register import tables
from funasr.models.rwkv_bat.rwkv import RWKV
from funasr.models.transformer.layer_norm import LayerNorm
from funasr.models.rwkv_bat.rwkv_subsampling import RWKVConvInput
from funasr.models.transformer.utils.nets_utils import make_source_mask
from funasr.models.rwkv_bat.rwkv_subsampling import RWKVConvInput

class RWKVEncoder(AbsEncoder):

@tables.register("encoder_classes", "RWKVEncoder")
class RWKVEncoder(torch.nn.Module):
    """RWKV encoder module.

    Based on https://arxiv.org/pdf/2305.13048.pdf.
@@ -41,16 +44,17 @@
        att_dropout_rate: float = 0.0,
        ffn_dropout_rate: float = 0.0,
        dropout_rate: float = 0.0,
        subsampling_factor: int =4,
        subsampling_factor: int = 4,
        time_reduction_factor: int = 1,
        kernel: int = 3,
        **kwargs,
    ) -> None:
        """Construct a RWKVEncoder object."""
        super().__init__()

        self.embed = RWKVConvInput(
            input_size,
            [output_size//4, output_size//2, output_size],
            [output_size // 4, output_size // 2, output_size],
            subsampling_factor,
            conv_kernel_size=kernel,
            output_size=output_size,
@@ -60,7 +64,7 @@

        linear_size = output_size * 4 if linear_size is None else linear_size
        attention_size = output_size if attention_size is None else attention_size
        

        self.rwkv_blocks = torch.nn.ModuleList(
            [
                RWKV(
@@ -118,12 +122,12 @@
                x, _ = block(x)
        else:
            x = self.rwkv_infer(x)
            

        x = self.final_norm(x)

        if self.time_reduction_factor > 1:
            x = x[:,::self.time_reduction_factor,:]
            olens = torch.floor_divide(olens-1, self.time_reduction_factor) + 1
            x = x[:, :: self.time_reduction_factor, :]
            olens = torch.floor_divide(olens - 1, self.time_reduction_factor) + 1

        return x, olens, None

@@ -131,9 +135,7 @@

        batch_size = xs_pad.shape[0]

        hidden_sizes = [
            self._output_size for i in range(5)
        ]
        hidden_sizes = [self._output_size for i in range(5)]

        state = [
            torch.zeros(
@@ -148,7 +150,7 @@

        xs_out = []
        for t in range(xs_pad.shape[1]):
            x_t = xs_pad[:,t,:]
            x_t = xs_pad[:, t, :]
            for idx, block in enumerate(self.rwkv_blocks):
                x_t, state = block(x_t, state=state)
            xs_out.append(x_t)