seamless_communication/ggml/test_unity_cpp.py

# Copyright (c) Meta Platforms, Inc. and affiliates
# All rights reserved.
#
# This source code is licensed under the license found in the
# MIT_LICENSE file in the root directory of this source tree.

import ctypes
import functools
import shutil
from ctypes import c_void_p
from pathlib import Path
from typing import Any, Iterator, Tuple

import fairseq2.nn
import fairseq2.nn.transformer
import numpy as np
import pytest
import requests  # type: ignore
import torch
import torchaudio  # type: ignore
from ctypes_utils import NULLPTR, Ptr
from fairseq2.data.audio import WaveformToFbankConverter
from fairseq2.models.wav2vec2.feature_extractor import Wav2Vec2FbankFeatureExtractor
from ggml_convert import convert_model, read_layer_config

import ggml
from ggml import NativeObj
from seamless_communication.inference.generator import SequenceGeneratorOptions
from seamless_communication.inference.translator import Modality, Translator

Ctx = ggml.ggml_context_p

UNITY_MODELS = Path(__file__).parent / "examples/unity/models"
FAIRSEQ2_CPP = Path(__file__).parent / "examples/unity/fairseq2.cpp"
UNITY_FLASH_ATTN = "\n# define UNITY_FLASH_ATTN 0\n" not in FAIRSEQ2_CPP.read_text()

DATA = Path(__file__).parent / "test_data"
LOCAL_AUDIO_SAMPLE_PATH = DATA / "LJ037-0171_sr16k.wav"
TEST_AUDIO_SAMPLE_URL = (
    "https://dl.fbaipublicfiles.com/seamless/tests/LJ037-0171_sr16k.wav"
)


MB = 1024 * 1024


@pytest.fixture(name="ctx")
def _ctx() -> Iterator[Ctx]:
    """Allocate a new context with 1024 MB of memory"""
    try:
        mem_size = 16 * MB
        memory = torch.zeros(mem_size, dtype=torch.uint8)
        ctx = ggml.ggml_init(
            params=ggml.ggml_init_params(
                mem_size=mem_size,
                mem_buffer=ctypes.c_void_p(memory.data_ptr()),
                no_alloc=True,
            )
        )

        # Create 'dot' folder for temporary dump of ggml graphs
        (Path(__file__).parent / "dot").mkdir(exist_ok=True)

        with torch.inference_mode():
            yield ctx
    finally:
        ggml.ggml_free(ctx)


@functools.lru_cache()
def _load_g_model_once() -> NativeObj:
    model_file = Path(__file__).parent / "seamlessM4T_medium.ggml"
    if not model_file.exists():
        convert_model("seamlessM4T_medium", model_file)
    return ggml.load_fairseq2_ggml_file(model_file)


@pytest.fixture()
def g_model(ctx: Ctx) -> c_void_p:
    model = _load_g_model_once()
    ggml.lib.fairseq2_model_set_inference_ctx(model.ptr, ctx)
    return model.ptr


@functools.lru_cache(maxsize=1)
def load_translator() -> Translator:
    return Translator("seamlessM4T_medium", None, device=torch.device("cpu"))


def load_pt_model() -> Any:
    return load_translator().model


def download_sample_audio() -> Any:
    Path(DATA).mkdir(exist_ok=True)
    response = requests.get(TEST_AUDIO_SAMPLE_URL, stream=True)
    with open(DATA / "LJ037-0171_sr16k.wav", "wb") as file:
        for chunk in response.iter_content(chunk_size=1024):
            if chunk:
                file.write(chunk)


def test_convert_linear(tmp_path: Path) -> None:
    module = fairseq2.nn.Linear(16, 24, True)

    layer_config = read_layer_config(module)
    assert layer_config == {"input_dim": 16, "output_dim": 24}

    module_file = Path("module.ggml")
    convert_model(module, module_file)
    g_module = ggml.load_fairseq2_ggml_file(module_file)

    for k, v in layer_config.items():
        assert (
            ggml.fairseq2_model_layer_config_int(g_module.ptr, bytes(k, "ascii")) == v
        )


def test_causal_attention_mask(ctx: Ctx):
    x = torch.zeros((1, 10, 32))
    generator = fairseq2.nn.transformer.CausalAttentionMaskFactory()
    mask_exp = generator(x, x).materialize().numpy()

    gx = ggml.from_numpy(ctx, x)
    gmask = ggml.causal_attention_mask(ctx, gx)
    ggml.build_and_compute(ctx, gmask)
    mask = ggml.to_numpy(gmask)

    assert mask_exp.shape == (10, 10)
    assert mask.shape == (10, 10)
    assert np.all(mask == mask_exp)

    x = x[:, :8, :]
    mask_exp = generator(x, x).materialize().numpy()
    gx = ggml.from_numpy(ctx, x)
    gmask = ggml.causal_attention_mask(ctx, gx)
    ggml.build_and_compute(ctx, gmask)
    mask = ggml.to_numpy(gmask)

    assert mask_exp.shape == (8, 8)
    assert mask.shape == (8, 8)
    assert np.all(mask == mask_exp)


def test_LayerNorm_forward(ctx: Ctx, g_model: c_void_p) -> None:
    x = torch.empty((2, 21, 1024))
    torch.nn.init.uniform_(x, -1, 1)

    pt_model = load_pt_model()
    y_exp = pt_model.text_encoder.layers[0].ffn_layer_norm(x).numpy()
    gx = ggml.from_numpy(ctx, x)
    gy = ggml.forward("LayerNorm", g_model, "text_encoder.layers.0.ffn_layer_norm", gx)
    ggml.build_and_compute(ctx, gy)

    y = ggml.to_numpy(gy)
    assert np.allclose(y_exp, y, atol=1e-5)


def test_Linear_forward(ctx: Ctx, g_model: c_void_p) -> None:
    x = torch.empty((2, 21, 1024))
    torch.nn.init.uniform_(x, -1, 1)

    pt_model = load_pt_model()
    y_exp = pt_model.text_encoder.layers[0].ffn.inner_proj(x).numpy()
    gx = ggml.from_numpy(ctx, x)
    gy = ggml.forward("Linear", g_model, "text_encoder.layers.0.ffn.inner_proj", gx)
    ggml.build_and_compute(ctx, gy, dump="dot/test_Linear_forward.dot")

    y = ggml.to_numpy(gy)
    assert np.allclose(y_exp, y, atol=1e-5)


def test_FeedForwardNetwork_forward(ctx: Ctx, g_model: c_void_p) -> None:
    x = torch.empty((2, 21, 1024))  # (bs, seq_len, model_dim)
    torch.nn.init.uniform_(x, -1 / 32, 1 / 32)

    # Test FFN without LayerNorm
    pt_model = load_pt_model()
    y_exp = pt_model.text_encoder.layers[0].ffn(x).numpy()
    gx = ggml.from_numpy(ctx, x)
    gy = ggml.forward(
        "StandardFeedForwardNetwork", g_model, "text_encoder.layers.0.ffn", gx
    )
    ggml.build_and_compute(ctx, gy)

    y = ggml.to_numpy(gy)
    assert np.allclose(y_exp, y, atol=1e-5)


@pytest.mark.parametrize("lengths", [(11, 21), (21, 13)])
def test_MultiheadAttention_forward(
    ctx: Ctx, g_model: c_void_p, lengths: Tuple[int, int]
) -> None:
    x = torch.empty((2, 21, 1024))
    torch.random.manual_seed(0)
    torch.nn.init.uniform_(x, -1, 1)

    # Note: we use different lengths for queries and keys,
    # this tests the implementation in decoding context too.
    # Note2: ggml_flash_attn requires that we have more keys than queries
    # qlen, klen = (11, 21) if flash_attn else (21, 13)
    qlen, klen = lengths
    xq = x[:, :qlen]
    xk = x[:, :klen]
    if qlen > klen and UNITY_FLASH_ATTN:
        pytest.skip(reason="flash_attn requires qlen > klen")

    gxq = ggml.from_numpy(ctx, xq.contiguous())
    ggml.ggml_set_name(gxq, b"xq")
    gxk = ggml.from_numpy(ctx, xk.contiguous())
    ggml.ggml_set_name(gxk, b"xk")
    ggml.ggml_set_no_alloc(ctx, True)
    gy = ggml.forward(
        "MultiheadAttention",
        g_model,
        "text_encoder.layers.0.self_attn",
        gxq,
        gxk,
        gxk,
        NULLPTR,  # TODO: tests with causal attention masks
    )
    gf = ggml.build_and_compute(ctx, gy, dump="dot/test_MultiheadAttention_forward")
    y = ggml.to_numpy(gy)
    nodes = ggml.nodes(gf)
    node_buffers = set(t.contents.data for t in nodes.values())

    pt_model = load_pt_model()
    self_attn = pt_model.text_encoder.layers[0].self_attn

    # If buffers are overlapping, reading node contents, can be misleading.
    overlap = len(node_buffers) < len(nodes)
    if not overlap:
        q_exp = self_attn._project_q(xq, None).numpy().reshape(2 * 16, qlen, 64)
        q = ggml.to_numpy(nodes[b"q"])
        assert q.shape == q_exp.shape
        assert np.allclose(q_exp, q, atol=1e-5)

        attn_weights_hook = fairseq2.nn.transformer.AttentionWeightStoreHook([])
        self_attn.register_attn_weight_hook(attn_weights_hook)

    y_exp = self_attn(xq, None, xk, None, xk).numpy()

    # with flash_attn we don't have attn_weights
    naive_attn = b"attn_weights" in nodes
    if naive_attn and not overlap:
        attn_weights = ggml.to_numpy(nodes[b"attn_weights"]).reshape(-1, 16, qlen, klen)
        [(_, attn_weights_exp)] = attn_weights_hook._storage
        attn_weights_exp = attn_weights_exp.numpy()
        assert attn_weights_exp.shape == attn_weights.shape
        # GGML is very agressively reducing small softmax weights to 0,
        # so the error isn't that small
        assert np.allclose(attn_weights_exp, attn_weights, atol=1e-3)
        # But the sums should be close to 1
        assert np.allclose(np.sum(attn_weights, axis=-1), np.ones((2, 16, qlen)))
        # And the maximum index should match the original ones.
        assert np.allclose(
            np.argmax(attn_weights_exp, axis=-1), np.argmax(attn_weights, axis=-1)
        )
    assert y.shape == y_exp.shape
    assert np.allclose(y_exp, y, atol=1e-2 if naive_attn else 1e-4)


def test_MultiheadAttention_forward_self_attn_with_cache(
    ctx: Ctx, g_model: c_void_p
) -> None:
    pt_model = load_pt_model()
    attn = pt_model.text_decoder.layers[0].self_attn

    x = torch.empty((2, 21, 1024))
    torch.random.manual_seed(0)
    torch.nn.init.uniform_(x, -1, 1)

    state_bag = fairseq2.nn.IncrementalStateBag(100)

    with ggml.fairseq2_kv_cache_alloc(g_model, 16 * MB, 2, 21):
        # Incremental decoding
        for t in range(3):
            xq = x[:, t : t + 1]

            gxq = ggml.from_numpy(ctx, xq.contiguous())
            ggml.ggml_set_name(gxq, b"xq")
            gy = ggml.forward(
                "MultiheadAttention",
                g_model,
                "text_decoder.layers.0.self_attn",
                gxq,
                gxq,
                gxq,
                None,  # type: ignore
            )
            gf = ggml.build_and_compute(
                ctx,
                gy,
                dump=f"dot/test_MultiheadAttention_forward_self_attn_with_cache_{t}.dot",
            )
            nodes = ggml.nodes(gf)
            gk_cache = ggml.to_numpy(
                nodes[b"text_decoder.layers.0.self_attn.k (step=%d)" % t]
            )
            assert gk_cache.shape == (2, t + 1, 1024)
            gk_cache = gk_cache.reshape(2, t + 1, 16, 64).transpose(0, 2, 1, 3)
            assert gk_cache.shape == (2, 16, t + 1, 64)

            y_exp = attn(xq, None, xq, None, xq, state_bag=state_bag).numpy()
            assert y_exp.shape == (2, 1, 1024)
            state = state_bag.get_state(attn, fairseq2.nn.transformer.AttentionState)
            state_bag.increment_step_nr()
            assert state is not None

            k_cache = state.get()[0].numpy()
            assert k_cache.shape == (2, 16, t + 1, 64)
            assert np.allclose(gk_cache, k_cache, atol=1e-3)

            y = ggml.to_numpy(gy)
            assert np.allclose(y, y_exp, atol=1e-2)


def test_MultiheadAttention_forward_cross_attn_with_cache(
    ctx: Ctx, g_model: c_void_p
) -> None:
    pt_model = load_pt_model()
    attn = pt_model.text_decoder.layers[0].encoder_decoder_attn

    x = torch.empty((2, 21, 1024))
    torch.random.manual_seed(0)
    torch.nn.init.uniform_(x, -1, 1)

    state_bag = fairseq2.nn.IncrementalStateBag(100)

    with ggml.fairseq2_kv_cache_alloc(g_model, 16 * MB, 2, 21):
        # Incremental decoding, the keys come from the encoder, and don't change during decoding
        xk = x[:, :11]
        gxk = ggml.from_numpy(ctx, xk.contiguous(), name=b"xk")

        for t in range(3):
            xq = x[:, t : t + 1]

            gxq = ggml.from_numpy(ctx, xq.contiguous())
            ggml.ggml_set_name(gxq, b"xq")
            gy = ggml.forward(
                "MultiheadAttention",
                g_model,
                "text_decoder.layers.0.encoder_decoder_attn",
                gxq,
                gxk,
                gxk,
                None,  # type: ignore
            )
            gf = ggml.build_and_compute(
                ctx,
                gy,
                dump=f"dot/test_MultiheadAttention_forward_cross_attn_with_cache_{t}.dot",
            )
            y = ggml.to_numpy(gy)
            nodes = ggml.nodes(gf)
            leaves = ggml.leafs(gf)

            if t > 0:
                # the cache only appear in the graph during the second call
                state = state_bag.get_state(
                    attn, fairseq2.nn.transformer.AttentionState
                )
                assert state is not None
                assert np.allclose(
                    state.get()[0].transpose(1, 2).numpy(),
                    ggml.to_numpy(
                        nodes[
                            b"text_decoder.layers.0.encoder_decoder_attn.k_cache (view)"
                        ]
                    ),
                    atol=1e-3,
                )

            state_bag.increment_step_nr()
            y_exp = attn(xq, None, xk, None, xk, state_bag=state_bag).numpy()
            assert y_exp.shape == (2, 1, 1024)
            assert np.allclose(y, y_exp, atol=1e-2)


def test_StandardTransformerEncoderLayer_forward(ctx: Ctx, g_model: c_void_p) -> None:
    x = torch.empty((2, 21, 1024))
    torch.random.manual_seed(0)
    torch.nn.init.uniform_(x, -1, 1)

    pt_model = load_pt_model()
    layer = pt_model.text_encoder.layers[0]

    gx = ggml.from_numpy(ctx, x)
    ggml.ggml_set_name(gx, b"x")
    gy = ggml.forward(
        "StandardTransformerEncoderLayer",
        g_model,
        "text_encoder.layers.0",
        gx,
        None,  # TODO support padding mask
    )
    gf = ggml.build_and_compute(ctx, gy)

    y = ggml.to_numpy(gy)

    y_exp, _ = layer(x, padding_mask=None)
    y_exp = y_exp.numpy()

    assert y.shape == y_exp.shape
    assert np.allclose(y_exp, y, atol=1e-4 if UNITY_FLASH_ATTN else 1e-2)


def test_StandardConformerEncoderLayer_forward(ctx: Ctx, g_model: c_void_p) -> None:
    pt_model = load_pt_model()
    x = torch.rand(1, 137, 1024)

    layer = pt_model.speech_encoder.inner.layers[0]
    gx = ggml.from_numpy(ctx, x[0])
    ggml.ggml_set_name(gx, b"x")
    gy = ggml.forward(
        "StandardConformerEncoderLayer",
        g_model,
        "speech_encoder.inner.layers.0",
        gx,
        None,  # TODO support padding mask
    )
    gf = ggml.build_and_compute(ctx, gy)

    y = ggml.to_numpy(gy)

    y_exp, _ = layer(x, padding_mask=None)
    y_exp = y_exp.squeeze(0).numpy()
    assert y.shape == y_exp.shape
    assert np.allclose(y_exp, y, atol=2e-3)


def test_StandardConformerEncoderAdaptorLayer_forward(
    ctx: Ctx, g_model: c_void_p
) -> None:
    pt_model = load_pt_model()
    torch.random.manual_seed(0)
    x = torch.rand(1, 137, 1024)
    layer = pt_model.speech_encoder.adaptor_layers[0]
    gx = ggml.from_numpy(ctx, x[0])
    ggml.ggml_set_name(gx, b"x")
    gy = ggml.forward(
        "StandardConformerEncoderAdaptorLayer",
        g_model,
        "speech_encoder.adaptor_layers.0",
        gx,
        None,  # TODO support padding mask
    )
    gf = ggml.build_and_compute(ctx, gy)

    y = ggml.to_numpy(gy)

    y_exp, _ = layer(x, None)
    y_exp = y_exp.numpy()

    assert y.shape == y_exp.shape
    assert np.allclose(y_exp, y, atol=2e-3)


def test_StandardTransformerEncoder_forward(ctx: Ctx, g_model: c_void_p) -> None:
    x = torch.empty((2, 21, 1024))
    padding_mask = fairseq2.nn.padding.PaddingMask(torch.tensor([21, 21]), 21)
    torch.random.manual_seed(0)
    torch.nn.init.uniform_(x, -1, 1)

    gx = ggml.from_numpy(ctx, x)
    ggml.ggml_set_name(gx, b"x")
    gpad = ggml.from_numpy(ctx, padding_mask.materialize())
    ggml.ggml_set_name(gpad, b"padding_mask")
    gy = ggml.forward(
        "StandardTransformerEncoder",
        g_model,
        "text_encoder",
        gx,
        None,  # TODO support padding mask
    )
    gf = ggml.build_and_compute(ctx, gy)

    y = ggml.to_numpy(gy)

    pt_model = load_pt_model()
    y_exp, _ = pt_model.text_encoder(x, padding_mask)
    y_exp = y_exp.numpy()

    assert y.shape == y_exp.shape
    assert np.allclose(y_exp, y, atol=5e-3)


def test_StandardConformerEncoder_forward(ctx: Ctx, g_model: c_void_p) -> None:
    pt_model = load_pt_model()
    if not LOCAL_AUDIO_SAMPLE_PATH.exists():
        download_sample_audio()
    wav, _ = torchaudio.load(LOCAL_AUDIO_SAMPLE_PATH)
    gx = ggml.from_numpy(ctx, wav * 2**15)  # Apply scale before sending into ggml!
    ggml.ggml_set_name(gx, b"x")
    gy = ggml.forward(
        "StandardConformerEncoder",
        g_model,
        "speech_encoder",
        gx,
        None,  # TODO support padding mask
    )
    gf = ggml.build_and_compute(ctx, gy)

    y = ggml.to_numpy(gy)

    cache = DATA / "test_StandardConformerEncoder_forward.npy"
    if not cache.exists():
        converter = WaveformToFbankConverter(
            num_mel_bins=80,
            waveform_scale=2**15,
            channel_last=True,
            standardize=True,
        )
        converter_input = {
            "waveform": wav.transpose(0, 1),
            "sample_rate": 16000.0,
            "format": -1,
        }

        pt_model = load_pt_model()
        speech_encoder_input = pt_model.speech_encoder_frontend(
            converter(converter_input)["fbank"].unsqueeze(0), None
        )[0]

        y_exp, _ = pt_model.speech_encoder(speech_encoder_input, None)
        y_exp = y_exp.numpy()
        np.save(cache, y_exp)
    else:
        y_exp = np.load(cache)

    assert y.shape == y_exp.shape
    assert np.allclose(y_exp, y, atol=1e-2)


def test_WaveformToFbank_forward(ctx: Ctx, g_model: c_void_p) -> None:
    converter = WaveformToFbankConverter(
        num_mel_bins=80,
        waveform_scale=2**15,
        channel_last=True,
        standardize=True,
    )
    extractor = Wav2Vec2FbankFeatureExtractor(80, stride=2, sample_every_k=1)
    if not LOCAL_AUDIO_SAMPLE_PATH.exists():
        download_sample_audio()
    wav, _ = torchaudio.load(LOCAL_AUDIO_SAMPLE_PATH)
    gx = ggml.from_numpy(ctx, wav * 2**15)  # Apply scale before sending into ggml!
    ggml.ggml_set_name(gx, b"x")

    gy = ggml.forward("WaveformToFbank", g_model, "", gx)
    gf = ggml.build_and_compute(ctx, gy)

    y = ggml.to_numpy(gy)
    converter_input = {
        "waveform": wav.transpose(0, 1),
        "sample_rate": 16000.0,
        "format": -1,
    }
    y_exp, _ = extractor(converter(converter_input)["fbank"].unsqueeze(0), None)
    y_exp = y_exp.squeeze(0).numpy()

    assert y.shape == y_exp.shape
    assert np.allclose(y_exp, y, atol=4e-3)  # reduce? error is from standardization


def test_PositionalEmbedding_forward(ctx: Ctx, g_model: c_void_p) -> None:
    seq = torch.zeros((4, 20, 1024), dtype=torch.float32)

    pos_encoder = fairseq2.nn.SinusoidalPositionEncoder(1024, 55, _legacy_pad_idx=1)
    y_exp = pos_encoder(seq, None)[0].numpy()

    gseq = ggml.from_numpy(ctx, seq[0].clone().numpy())
    ggml.ggml_set_name(gseq, b"seq")
    gy = ggml.forward(
        "PositionalEmbedding", g_model, "text_decoder_frontend.pos_encoder", gseq
    )
    gf = ggml.build_and_compute(ctx, gy, dump=True)
    y = ggml.to_numpy(gy)

    assert y.shape == y_exp.shape
    assert np.allclose(y_exp, y, atol=1e-6)


def test_PositionalEmbedding_forward_with_cache(ctx: Ctx, g_model: c_void_p) -> None:
    seq = torch.zeros((4, 20, 1024), dtype=torch.float32)
    pos_encoder = fairseq2.nn.SinusoidalPositionEncoder(1024, 55, _legacy_pad_idx=1)
    pos_encoder.eval()
    state_bag = fairseq2.nn.IncrementalStateBag(100)

    with ggml.fairseq2_kv_cache_alloc(g_model, 16 * MB, 2, 21):
        # Incremental decoding
        for t in range(20):
            gseq = ggml.from_numpy(ctx, seq[:, t : t + 1, :].numpy())
            ggml.ggml_set_name(gseq, b"seq")
            gy = ggml.forward(
                "PositionalEmbedding",
                g_model,
                "text_decoder_frontend.pos_encoder",
                gseq,
            )
            ggml.build_and_compute(ctx, gy, dump=t == 1)
            y = ggml.to_numpy(gy)

            y_exp = pos_encoder(seq[:, t : t + 1, :], None, state_bag=state_bag).numpy()
            state_bag.increment_step_nr()
            assert y.shape == y_exp.shape
            assert np.allclose(y_exp, y, atol=1e-6)


def test_TransformerEmbeddingFrontend_forward(ctx: Ctx, g_model: c_void_p) -> None:
    seq = torch.arange(2 * 20).reshape(2, 20)
    seq[1, 15:] = 0  # padding for second sentence
    seq_len = torch.tensor([20, 15])
    gseq = ggml.from_numpy(ctx, seq.numpy().astype(np.int32))

    ggml.ggml_set_name(gseq, b"seq")
    gy = ggml.forward(
        "TransformerEmbeddingFrontend", g_model, "text_decoder_frontend", gseq
    )
    ggml.build_and_compute(ctx, gy)
    y = ggml.to_numpy(gy)

    pt_model = load_pt_model()
    y_exp, _ = pt_model.text_decoder_frontend(seq, seq_len)
    y_exp = y_exp.numpy()

    assert y.shape == y_exp.shape
    assert np.allclose(y_exp, y, atol=1e-6)


def test_StandardTransformerDecoderLayer_forward(ctx: Ctx, g_model: c_void_p) -> None:
    x = torch.empty((2, 13, 1024))
    encoder_out = torch.empty((2, 21, 1024))
    torch.random.manual_seed(0)
    torch.nn.init.uniform_(x, -1, 1)
    torch.nn.init.uniform_(encoder_out, -1, 1)

    self_attn_mask = fairseq2.nn.transformer.CausalAttentionMaskFactory()(x, x)
    gx = ggml.from_numpy(ctx, x)
    ggml.ggml_set_name(gx, b"x")
    gself_attn_mask = ggml.from_numpy(ctx, self_attn_mask.materialize().numpy())
    ggml.ggml_set_name(gself_attn_mask, b"self_attn_mask")
    genc = ggml.from_numpy(ctx, encoder_out)
    ggml.ggml_set_name(genc, b"encoder_out")
    gy = ggml.forward(
        "StandardTransformerDecoderLayer",
        g_model,
        "text_decoder.layers.0",
        gx,
        gself_attn_mask,
        genc,
        NULLPTR,  # TODO support padding mask,
    )
    ggml.build_and_compute(ctx, gy, dump=True)
    y = ggml.to_numpy(gy)

    pt_model = load_pt_model()
    y_exp, _ = pt_model.text_decoder.layers[0](x, None, encoder_output=encoder_out, self_attn_mask=self_attn_mask)
    y_exp = y_exp.numpy()

    assert y.shape == y_exp.shape
    # We still have some numerical imprecision
    assert np.allclose(y_exp, y, atol=0.1)
    assert np.sum(np.abs(y_exp-y) > 1e-2) < 20


def test_StandardTransformerDecoder_forward(ctx: Ctx, g_model: c_void_p) -> None:
    x = torch.empty((2, 13, 1024))
    encoder_out = torch.empty((2, 21, 1024))
    padding_mask = fairseq2.nn.padding.PaddingMask(torch.tensor([13, 13]), 13)
    torch.random.manual_seed(0)
    torch.nn.init.uniform_(x, -1, 1)
    torch.nn.init.uniform_(encoder_out, -1, 1)
    gx = ggml.from_numpy(ctx, x)
    ggml.ggml_set_name(gx, b"x")
    gpad = ggml.from_numpy(ctx, padding_mask.materialize())
    ggml.ggml_set_name(gpad, b"padding_mask")
    genc = ggml.from_numpy(ctx, encoder_out)
    gy = ggml.forward(
        "StandardTransformerDecoder",
        g_model,
        "text_decoder",
        gx,
        None,  # TODO support padding mask,
        genc,
        None,
    )
    ggml.build_and_compute(ctx, gy)
    y = ggml.to_numpy(gy)

    pt_model = load_pt_model()
    y_exp, _ = pt_model.text_decoder(x, padding_mask, encoder_out, None)
    y_exp = y_exp.numpy()

    assert y.shape == y_exp.shape
    assert np.allclose(y_exp, y, atol=1e-3)  # TODO: those tests are failing now


def test_s2tt(ctx: Ctx, g_model: c_void_p):
    if not LOCAL_AUDIO_SAMPLE_PATH.exists():
        download_sample_audio()
    src_audio_wav, _ = torchaudio.load(LOCAL_AUDIO_SAMPLE_PATH)
    sample_file = DATA / "LJ037-0171_sr16k.wav.trans"
    translator = load_translator()
    if not sample_file.exists():
        decoded_audio = {
            "waveform": src_audio_wav.t(),
            "sample_rate": 16000.0,
            "format": -1,
        }
        src = translator.collate(translator.convert_to_fbank(decoded_audio))["fbank"]

        text_out, _ = translator.get_prediction(
            translator.model,
            translator.text_tokenizer,
            translator.unit_tokenizer,
            src["seqs"],
            padding_mask=None,
            input_modality=Modality.SPEECH,
            output_modality=Modality.TEXT,
            tgt_lang="cmn",
            text_generation_opts=SequenceGeneratorOptions(),
            unit_generation_opts=None,
        )

        tgt_text = str(text_out[0])
        assert tgt_text == "专家的检查和证据使该委员会得出了结论,可能有五次枪击."
        with open(sample_file, "w") as f:
            f.write(tgt_text)

    with open(sample_file, "r") as exp:
        exp_tgt_text = exp.readlines()[0].strip()

    # Apply scale before sending into ggml!
    gx = ggml.from_numpy(ctx, src_audio_wav * 2**15)
    ggml.ggml_set_name(gx, b"x")
    encoder_out = ggml.forward(
        "StandardConformerEncoder",
        g_model,
        "speech_encoder",
        gx,
        NULLPTR,  # TODO support padding mask
    )
    gf = ggml.build_and_compute(ctx, encoder_out)

    beam_size = 5
    opts = ggml.SequenceGeneratorOptions(
        beam_size=beam_size,
        soft_max_seq_len_a=1,
        soft_max_seq_len_b=200,
        hard_max_seq_len=500,
    )
    job = ggml.SequenceGeneratorJob(
        opts=opts,
        prefix_seq=ggml.from_numpy(ctx, np.array([3, 256200]).astype(np.int32)),
        pad_idx=0,
        unk_idx=1,
        bos_idx=2,
        eos_idx=3,
    )
    result_ptr = ggml.generate_sequence(g_model, Ptr(job), encoder_out, NULLPTR, ctx)
    results = [result_ptr[i] for i in range(beam_size) if result_ptr[i].seq != None]
    tokens = [
        translator.text_tokenizer.model.index_to_token(id)
        for id in ggml.to_numpy(results[0].seq).tolist()
    ][2:-1]
    tokens = "".join(tokens).replace("▁", " ")[1:]
    assert tokens == exp_tgt_text