modeling_pp.py

# Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved.
#
# 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.

# pass
import paddle
import paddle.distributed.fleet as fleet
import paddle.nn as nn
from paddle.distributed.fleet.meta_parallel import LayerDesc, PipelineLayer

from paddlenlp.transformers import PretrainedModel
from paddlenlp.transformers.llama.modeling import (
    LlamaConfig,
    LlamaDecoderLayer,
    LlamaLMHead,
    LlamaModel,
    LlamaPretrainedModel,
    LlamaPretrainingCriterion,
    LlamaRMSNorm,
)


def get_hcg():
    return fleet.get_hybrid_communicate_group()


def parse_args(args):
    if isinstance(args, tuple):
        if len(args) == 3:
            hidden_states, attention_mask, position_ids = args
        elif len(args) == 2:
            hidden_states, attention_mask = args
            position_ids = None
    else:
        hidden_states = args
        attention_mask, position_ids = None, None

    if position_ids is not None:
        position_ids.stop_gradient = True

    if attention_mask is not None:
        attention_mask.stop_gradient = True

    return hidden_states, attention_mask, position_ids


def return_args(hidden_states, attention_mask=None, position_ids=None):
    ret = (hidden_states,)

    if attention_mask is not None:
        ret += (attention_mask.clone(),)
    if position_ids is not None:
        ret += (position_ids.clone(),)
    if len(ret) == 1:
        ret = ret[0]

    return ret


class LlamaEmbeddingPipe(nn.Layer):
    """Extends LlamaEmbeddings to forward attention_mask through the pipeline."""

    def __init__(self, config):
        super(LlamaEmbeddingPipe, self).__init__()
        self.sequence_parallel = config.sequence_parallel
        self.hidden_size = config.hidden_size
        if config.tensor_parallel_degree > 1:
            self.embed_tokens = fleet.meta_parallel.VocabParallelEmbedding(
                config.vocab_size,
                config.hidden_size,
                weight_attr=paddle.ParamAttr(initializer=nn.initializer.XavierNormal()),
            )
        else:
            self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size)

    def forward(self, args):
        """_summary_

        Args:
            input (_type_): _description_

        Returns:
            _type_: _description_
        """
        input_ids, attention_mask, position_ids = parse_args(args)
        input_embeds = self.embed_tokens(input_ids)
        if self.sequence_parallel:
            from paddlenlp.transformers import ScatterOp

            # [bs, seq_len, num_head * head_dim] -> [bs * seq_len, num_head * head_dim]
            bs, seq_len, hidden_size = input_embeds.shape
            input_embeds = paddle.reshape_(input_embeds, [bs * seq_len, hidden_size])
            # [seq_len * bs / n, num_head * head_dim] (n is mp parallelism)
            input_embeds = ScatterOp.apply(input_embeds)

        batch_size, seq_length = input_ids.shape
        if attention_mask is not None:
            attention_mask = LlamaModel._prepare_decoder_attention_mask(
                attention_mask, (batch_size, seq_length), 0, input_embeds.dtype
            )
            attention_mask.stop_gradient = True

        return return_args(input_embeds, attention_mask, position_ids)


class LlamaDecoderLayerPipe(LlamaDecoderLayer):
    def forward(self, args):
        hidden_states, attention_mask, position_ids = parse_args(args)
        hidden_states = super().forward(hidden_states, attention_mask=attention_mask)
        return return_args(hidden_states, attention_mask, position_ids)


class LlamaRMSNormPipe(LlamaRMSNorm):
    def forward(self, args):
        hidden_states, attention_mask, position_ids = parse_args(args)
        return super().forward(hidden_states)


class PipelinePretrainedModel(PretrainedModel):
    _sequential_layers = []
    _pipeline_name_mapping = None

    def __init__(self, config, *args, **kwargs):
        super().__init__(config, *args, **kwargs)

    def add_sequential_layer(self, layer_desc, name_prefix=""):
        self._sequential_layers.append({"layer": layer_desc, "name_prefix": name_prefix})

    def get_sequential_layers(self):
        return [x["layer"] for x in self._sequential_layers]

    def get_sequential_name_prefixs(self):
        return {str(index): x["name_prefix"] for index, x in enumerate(self._sequential_layers)}

    def _set_pipeline_name_mapping(self, mappings=None):
        if mappings is not None:
            self._pipeline_name_mapping = mappings
        else:
            mapping = {}
            state_dict_keys = list(super().state_dict().keys())
            first_key = state_dict_keys[0].split(".")
            # if use virtual pp_degree, the prefix is like 0.0.xxx
            # else it will be like 0.xxx
            use_virtual_pp_degree = first_key[0].isdigit() and first_key[1].isdigit()

            prefixs = self.get_sequential_name_prefixs()
            for k in state_dict_keys:
                name_splited = k.split(".")
                if use_virtual_pp_degree:
                    idx = str(int(name_splited[0]) + int(name_splited[1]))
                    single_name = [prefixs[idx]]
                    single_name.extend(name_splited[2:])
                else:
                    idx = name_splited[0]
                    single_name = [prefixs[idx]]
                    single_name.extend(name_splited[1:])
                mapping[".".join(single_name)] = k

            self._pipeline_name_mapping = mapping

        return self._pipeline_name_mapping

    def state_dict(self, *args, **kwargs):
        state_dict = super().state_dict(*args, **kwargs)

        if self._pipeline_name_mapping is None:
            self._set_pipeline_name_mapping()
        assert len(self._pipeline_name_mapping) > 0, "The pipeline stage must have parameters!"
        pp_to_single_mapping = {v: k for k, v in self._pipeline_name_mapping.items()}

        for k in list(state_dict.keys()):
            v = state_dict.pop(k)
            state_dict[pp_to_single_mapping[k]] = v

        return state_dict

    def set_state_dict(self, state_dict, *args, **kwargs):
        if self._pipeline_name_mapping is None:
            self._set_pipeline_name_mapping()
        assert len(self._pipeline_name_mapping) > 0, "The pipeline stage must have parameters!"

        for k in list(state_dict.keys()):
            v = state_dict.pop(k)
            if k not in self._pipeline_name_mapping:
                continue
            state_dict[self._pipeline_name_mapping[k]] = v

        ret = super().set_state_dict(state_dict, *args, **kwargs)
        return ret


class LlamaForCausalLMPipe(PipelinePretrainedModel, PipelineLayer):
    """LlamaForPretraining adapted for pipeline parallelism.

    The largest change is flattening the LlamaModel class so we can express it as a
    sequence of layers including embedding, transformer layers, and output.
    """

    config_class = LlamaConfig

    _get_tensor_parallel_mappings = LlamaPretrainedModel._get_tensor_parallel_mappings
    _init_weights = LlamaPretrainedModel._init_weights

    # NO base_model_prefix !!!!

    def __init__(
        self,
        config,
        # scale_qk_by_layer_num=True,
        # virtual_pp_degree=4,
    ):
        self.config = config

        self.use_recompute = self.config.use_recompute
        self.recompute_granularity = self.config.recompute_granularity
        self.pp_recompute_interval = self.config.pp_recompute_interval
        self.no_recompute_layers = config.no_recompute_layers if config.no_recompute_layers is not None else []
        if self.recompute_granularity == "full":
            assert len(self.no_recompute_layers) == 0, "for pp with full recompute, no_recompute_layers is not support"

        # virtual_pp_degree = self.config.virtual_pp_degree
        virtual_pp_degree = getattr(self.config, "virtual_pp_degree", 1)

        hcg = get_hcg()
        tensor_parallel_degree = max(hcg.get_model_parallel_world_size(), 1)
        tensor_parallel_rank = max(hcg.get_model_parallel_rank(), 0)

        config.tensor_parallel_degree = tensor_parallel_degree
        config.tensor_parallel_rank = tensor_parallel_rank

        self.add_sequential_layer(LayerDesc(LlamaEmbeddingPipe, config=config), "llama")
        for i in range(config.num_hidden_layers):
            self.add_sequential_layer(
                LayerDesc(LlamaDecoderLayerPipe, config=config, layerwise_recompute=i not in self.no_recompute_layers),
                f"llama.layers.{i}",
            )

        self.add_sequential_layer(LayerDesc(LlamaRMSNormPipe, config=config), "llama.norm")
        self.add_sequential_layer(LayerDesc(LlamaLMHead, config=config), "lm_head")

        recompute_interval = 0
        if self.use_recompute and self.recompute_granularity == "full":
            assert self.config.pp_recompute_interval <= config.num_hidden_layers // (
                virtual_pp_degree * get_hcg().topology().get_dim_size("pipe")
            ), "pp recompute interval should smaller than num layers of each pp chunk"
            recompute_interval = self.config.pp_recompute_interval

        seg_method = "layer:LlamaDecoderLayer"
        if config.num_hidden_layers % get_hcg().topology().get_dim_size("pipe") != 0:
            seg_method = "uniform"

        PipelineLayer.__init__(
            self,
            layers=self.get_sequential_layers(),
            loss_fn=LlamaPretrainingCriterion(config),
            topology=get_hcg().topology(),
            seg_method=seg_method,
            recompute_interval=recompute_interval,
            recompute_ctx={
                "mp_group": get_hcg().get_model_parallel_group(),
                "offload": False,
                "partition": False,
            },
            num_virtual_pipeline_stages=virtual_pp_degree,
        )
        self.apply(self._init_weights)
        # DON'T init PipelinePretrainedModel
        # PipelinePretrainedModel.__init__(self.super(), config=config)