rust-bert/utils/convert_model.py

# Copyright 2019-2023 Guillaume Becquin
# Copyright 2023 https://github.com/starkat99/half-rs
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#     http://www.apache.org/licenses/LICENSE-2.0
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
# Copyright (c) 2005-2023, NumPy Developers.
# All rights reserved.
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
#
# * Redistributions in binary form must reproduce the above
# copyright notice, this list of conditions and the following
# disclaimer in the documentation and/or other materials provided
# with the distribution.
#
# * Neither the name of the NumPy Developers nor the names of any
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.


import argparse
import glob
import logging
import subprocess
import sys
import zipfile
from pathlib import Path
from typing import Dict
import os
import numpy as np
import torch
from numpy.lib.format import write_array
# from numpy.lib.npyio import zipfile_factory
from torch import Tensor

def zipfile_factory(file, *args, **kwargs):
    if not hasattr(file, 'read'):
        file = os.fspath(file)
    import zipfile
    kwargs['allowZip64'] = True
    kwargs['compresslevel'] = 4
    return zipfile.ZipFile(file, *args, **kwargs)

def get_bf16_repr(input_tensor: torch.Tensor) -> np.ndarray:
    """Convert a bfloat16 tensor to an equivalent byte representation in Numpy.
    This is a vectorized implementation inspired from https://github.com/starkat99/half-rs/blob/main/src/bfloat/convert.rs
    (shared under Apache 2.0 license at https://github.com/starkat99/half-rs/blob/main/LICENSES/Apache-2.0.txt)
    """
    v_fp32 = input_tensor.cpu().float().numpy()
    byte_array = np.frombuffer(v_fp32.tobytes(), dtype=np.uint32)
    nan_value = np.logical_or(np.right_shift(byte_array, 16), 0x0040)
    nan_mask = np.logical_and(byte_array, 0x7FFF_FFFF) > 0x7F80_0000
    round_bit = 0x0000_8000
    output_val = np.right_shift(byte_array, 16)
    threshold_mask = (np.logical_and(byte_array, round_bit) != 0) & (
        np.logical_and(byte_array, (3 * round_bit - 1)) != 0
    )
    output = np.where(
        nan_mask, nan_value, np.where(threshold_mask, output_val + 1, output_val)
    ).astype(np.uint16)
    return output


def append_to_zipf(
    array_dict: Dict[str, np.ndarray], parent_zipfile: zipfile.ZipFile
) -> None:
    """Append a dictionary of arrays to a parent zipfile.

    Inspired from https://github.com/numpy/numpy/blob/main/numpy/lib/npyio.py
    shared under BSD 3-Clause license by the numpy team
    """
    for key, array in array_dict.items():
        internal_filename = key + ".npy"
        array = np.asanyarray(array)
        with parent_zipfile.open(internal_filename, "w", force_zip64=True) as f_in:
            write_array(f_in, array, allow_pickle=True, pickle_kwargs=None)


if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument(
        "source_file",
        nargs="+",
        help="""Absolute path (or file pattern) to the Pytorch weights file(s) to convert.
        A single file, list of files, glob pattern or list of glob patterns can be provided.""",
    )
    parser.add_argument(
        "--skip_embeddings",
        action="store_true",
        help="Skip shared embeddings",
    )
    parser.add_argument(
        "--skip_lm_head", action="store_true", help="Skip language model head"
    )
    parser.add_argument("--prefix", help="Add a prefix on weight names")
    parser.add_argument(
        "--suffix",
        action="store_true",
        help="Split weight names on '.' and keep only last part",
    )
    parser.add_argument(
        "--dtype",
        help="Convert weights to a specific numpy DataType (float32, float16, ...)",
    )
    parser.add_argument(
        "--download_libtorch",
        action="store_true",
        help="Use this flag to enable automatic download of the libtorch library.",
    )
    args = parser.parse_args()
    
    logger = logging.getLogger('convert_model')
    logger.setLevel(logging.DEBUG)
    fh = logging.FileHandler('convert_model.log')
    fh.setLevel(logging.DEBUG)
    logger.addHandler(fh)

    target_folder = Path(args.source_file[0]).parent
    with zipfile_factory(
        target_folder / "model.npz", mode="w", compression=False
    ) as output_zipfile:
        for source_file_or_pattern in args.source_file:
            source_files = glob.glob(source_file_or_pattern)
            for source_file in source_files:
                logger.info(f"Processing source file {source_file}")
                nps = {}
                source_file = Path(source_file)
                weights = torch.load(str(source_file), map_location="cpu")

                for k, v in weights.items():
                    k = k.replace("gamma", "weight").replace("beta", "bias")
                    if args.skip_embeddings:
                        if k in {
                            "model.encoder.embed_tokens.weight",
                            "encoder.embed_tokens.weight",
                            "model.decoder.embed_tokens.weight",
                            "decoder.embed_tokens.weight",
                        }:
                            continue
                    if args.skip_lm_head:
                        if k in {
                            "lm_head.weight",
                        }:
                            continue
                    if args.prefix:
                        k = args.prefix + k
                    if args.suffix:
                        k = k.split(".")[-1]
                    if isinstance(v, Tensor):
                        if v.dtype == torch.bfloat16:
                            tensor = get_bf16_repr(v)
                        else:
                            tensor = v.cpu().numpy()
                        if args.dtype is not None:
                            nps[k] = np.ascontiguousarray(
                                tensor.astype(np.dtype(args.dtype))
                            )
                        else:
                            nps[k] = np.ascontiguousarray(tensor)
                        logger.info(
                            f"converted {k} - {str(sys.getsizeof(nps[k]))} bytes"
                        )
                    else:
                        logger.info(f"skipped non-tensor object: {k}")
            append_to_zipf(nps, output_zipfile)

    source = str(target_folder / "model.npz")
    target = str(target_folder / "rust_model.ot")

    toml_location = (Path(__file__).resolve() / ".." / ".." / "Cargo.toml").resolve()
    cargo_args = [
        "cargo",
        "run",
        "--bin=convert-tensor",
        "--manifest-path=%s" % toml_location,
        "--",
        source,
        target,
    ]
    if args.download_libtorch:
        cargo_args += ["--features", "download-libtorch"]
    subprocess.run(cargo_args)