verl/workers/reward_model/megatron/reward_model.py

# Copyright 2024 Bytedance Ltd. and/or its affiliates
#
# 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.
"""
Megatron Reward Model.
"""

import torch
import torch.distributed
from megatron.core import parallel_state as mpu
from megatron.core.pipeline_parallel import get_forward_backward_func
from tensordict import TensorDict

from verl import DataProto
from verl.utils.megatron.pipeline_parallel import make_batch_generator
from verl.utils.torch_functional import broadcast_dict_tensor, pad_sequence_to_length, split_dict_tensor_into_batches
from verl.workers.reward_model.base import BasePPORewardModel


class MegatronRewardModel(BasePPORewardModel):
    def __init__(
        self,
        config,
        model_config,
        reward_model_module: torch.nn.ModuleList,
        hf_config,
        tf_config,
        sft_tokenizer=None,
        rm_tokenizer=None,
    ):
        self.config = config
        self.reward_model_module = reward_model_module
        self.hf_config = hf_config
        self.tf_config = tf_config
        self.model_config = model_config
        self.device = "cuda"
        self.sft_tokenizer = sft_tokenizer
        self.rm_tokenizer = rm_tokenizer
        self.use_different_tokenizer = rm_tokenizer is not None

        print(f"MegatronRewardModel.config: {self.config}")

        if self.config.megatron.param_offload:
            self.offload_params_to_cpu()

    def re_encode_by_rm_tokenizer(self, data: DataProto) -> DataProto:
        assert self.use_different_tokenizer, "re-encode need rm tokenizer not be None!"
        # need to use rm tokenizer to re-generate input_ids, attention_mask and position_ids
        # 1. remove pad for each sequence
        # 2. decode by sft_tokenizer, remove sft system prompts
        # 3. encode by rm_tokenizer with rm system prompts, get rm_input_ids
        # 4. generate attention_mask and position_ids
        input_ids = data.batch["input_ids"]  # (bs, seq_len)
        attention_mask = data.batch["attention_mask"]
        position_ids = data.batch["position_ids"]
        ori_values = {"input_ids": input_ids, "attention_mask": attention_mask, "position_ids": position_ids}
        _, ori_seqlen = input_ids.size(0), input_ids.size(1)
        input_ids_for_rm = []
        attention_mask_for_rm = []
        position_ids_for_rm = []
        print_decode = True
        ori_seqlen = ori_seqlen + 128
        for id, mask in zip(input_ids, attention_mask):
            # 1. remove pad for each sequence
            non_zero_indices = torch.nonzero(mask).view(-1)
            begin_pos, end_pos = non_zero_indices[0].item(), non_zero_indices[-1].item()
            valid_id = id[begin_pos : end_pos + 1]
            # 2. decode by sft_tokenizer, remove sft system prompts
            decode_result = self.sft_tokenizer.decode(valid_id)
            # workaround
            decode_with_rm_chat = decode_result.replace("<|user|>\n", "[INST] ").replace("</s>\n<|assistant|>\n", " [/INST]").replace("</s> \n<|assistant|>\n", " [/INST]") + "</s>"
            if print_decode and torch.distributed.get_rank() == 0:
                # only print first decode result
                print(
                    f"device {torch.cuda.current_device()}: sft decode result:\n{decode_result}\n \
                        \ndevice {torch.cuda.current_device()}: sft decode result with \
                        rm chat template:\n{decode_with_rm_chat}\n\n"
                )
                print_decode = False
            # 3. encode by rm_tokenizer
            rm_input_ids = self.rm_tokenizer(decode_with_rm_chat, return_tensors="pt")["input_ids"][0].to(input_ids.device)
            # 4. generate attention_mask and position_ids
            rm_attention_mask = torch.ones_like(rm_input_ids, device=input_ids.device)
            cur_seqlen = rm_input_ids.shape[-1]
            # NOTE(gh): the later reward compute will process the shape (bs, seqlen_pad_128)
            if cur_seqlen > ori_seqlen:
                print(f"warninig: rm encode seqlen {cur_seqlen} > sft encode seqlen {ori_seqlen}")
                rm_input_ids = rm_input_ids[:ori_seqlen]
                rm_attention_mask = rm_attention_mask[:ori_seqlen]
            else:
                # right padding
                rm_input_ids = pad_sequence_to_length(rm_input_ids, ori_seqlen, self.rm_tokenizer.pad_token_id)
                rm_attention_mask = pad_sequence_to_length(rm_attention_mask, ori_seqlen, 0)
            rm_position_ids = torch.arange(0, ori_seqlen, device=input_ids.device)
            input_ids_for_rm.append(torch.unsqueeze(rm_input_ids, dim=0))
            attention_mask_for_rm.append(torch.unsqueeze(rm_attention_mask, dim=0))
            position_ids_for_rm.append(torch.unsqueeze(rm_position_ids, dim=0))
        input_ids_for_rm = torch.cat(input_ids_for_rm, dim=0)
        attention_mask_for_rm = torch.cat(attention_mask_for_rm, dim=0)
        position_ids_for_rm = torch.cat(position_ids_for_rm, dim=0)

        # (bs, seqlen) will not change, but input_ids, attention_mask and position_ids will change
        # NOTE(gh): need to replace into origin values after compute reward!
        data.batch["input_ids"] = input_ids_for_rm
        data.batch["attention_mask"] = attention_mask_for_rm
        data.batch["position_ids"] = position_ids_for_rm

        return data, ori_values

    @torch.no_grad()
    def compute_reward(self, data: DataProto) -> DataProto:
        if self.config.megatron.param_offload:
            self.load_params_to_cuda()

        if self.use_different_tokenizer:
            data, ori_values = self.re_encode_by_rm_tokenizer(data)

        input_ids = data.batch["input_ids"]  # (bs, seq_len')
        attention_mask = data.batch["attention_mask"]
        position_ids = data.batch["position_ids"]

        responses = data.batch["responses"]
        batch_size = responses.size(0)
        response_length = responses.size(1)

        with torch.no_grad():
            output = self.forward_batch(data)
            if mpu.is_pipeline_last_stage(ignore_virtual=True):
                logits = torch.cat(output, dim=0)
            else:
                logits = torch.empty(
                    (input_ids.shape[0], input_ids.shape[1]),
                    device=input_ids.device,
                )
            logits = logits.to(torch.float32)

            # broadcast across pp ranks
            torch.distributed.broadcast(
                tensor=logits,
                src=mpu.get_pipeline_model_parallel_last_rank(),
                group=mpu.get_pipeline_model_parallel_group(),
                async_op=False,
            )

        # (bs, seqlen', hidden_size) -> (bs, seqlen', 1) -> (bs, seqlen')
        token_level_rewards = logits
        # find the last token reward
        ends = attention_mask.cumsum(dim=-1).argmax(dim=-1).view(-1, 1)  # (bs, 1)
        rewards = torch.gather(token_level_rewards, dim=1, index=ends)  # (bs, 1)

        if self.use_different_tokenizer:
            data.batch.update(ori_values)
            input_ids = ori_values["input_ids"]
            attention_mask = ori_values["attention_mask"]
            position_ids = ori_values["position_ids"]

        token_level_rewards = rewards.expand(attention_mask.shape[0], attention_mask.shape[1])  # (bs, ori_seqlen)

        # assign last valid token reward to ori position
        eos_mask_idx = torch.argmax(position_ids * attention_mask, dim=-1)  # (bs,)
        eos_mask = torch.zeros_like(attention_mask)
        eos_mask[torch.arange(batch_size), eos_mask_idx] = 1.0

        token_level_rewards = token_level_rewards * eos_mask
        token_level_rewards = token_level_rewards[:, -response_length:]

        if self.config.megatron.param_offload:
            self.offload_params_to_cpu()
        else:
            # add empty cache after each compute
            torch.cuda.empty_cache()

        batch = TensorDict({"rm_scores": token_level_rewards}, batch_size=input_ids.shape[0])

        return DataProto(batch=batch)

    def forward_batch(self, data: DataProto):
        """
        We assume:
        - The model takes input: (input_ids, attention_mask, position_ids). No rmpad for the input
        - The communication shape is (total_nnz_pad_to_sp // tp_size, 1, hidden_size) if sequence parallel is enabled
        """
        # broadcast from last pp rank to all other pp ranks
        # TODO: actually, we just need to control the sampling order.
        data.batch = data.batch.contiguous()
        broadcast_dict_tensor(data.batch, src=mpu.get_pipeline_model_parallel_last_rank(), group=mpu.get_pipeline_model_parallel_group())

        # split into micro-batches
        if self.config is not None and "micro_batch_size_per_gpu" in self.config:
            infer_batch_size = self.config.micro_batch_size_per_gpu
        else:
            infer_batch_size = data.batch.batch_size[0]

        data.batch["attention_mask"] = data.batch["attention_mask"].to(bool)
        batches = split_dict_tensor_into_batches(data.batch, batch_size=infer_batch_size)
        n_micro_batch = len(batches)
        seq_len = batches[0]["input_ids"].shape[1]

        # compute input shapes for pp stages
        forward_backward_func = get_forward_backward_func()

        def loss_func(output):
            return 1.0, output

        def forward_step(batch_iter, model):
            batch = next(batch_iter)
            input_ids = batch["input_ids"]
            attention_mask = batch["attention_mask"]
            position_ids = batch["position_ids"]
            from verl.models.mcore import get_mcore_forward_fn

            forward_fn = get_mcore_forward_fn(self.hf_config)

            output = forward_fn(
                model,
                input_ids,
                attention_mask,
                position_ids,
                sequence_parallel=self.tf_config.sequence_parallel,
                value_model=True,
            )

            return output, loss_func

        # batch should be a list of batches inside micro-batches
        batch_generator = make_batch_generator(batches, vpp_size=len(self.reward_model_module))

        # TODO: we may use the new schedule instead
        # for flash-attn: (seq_len, batch_size, hidden_size) = (mbs*seq_len, 1, hidden_size)
        if mpu.get_pipeline_model_parallel_world_size() > 1:
            losses_reduced = forward_backward_func(
                forward_step_func=forward_step,
                data_iterator=batch_generator,
                model=self.reward_model_module,
                num_microbatches=n_micro_batch,
                seq_length=infer_batch_size * seq_len,  # no use when input_shapes was set
                micro_batch_size=1,  # no use when input_shapes was set
                forward_only=True,
            )
        else:
            losses_reduced = forward_backward_func(
                forward_step_func=forward_step,
                data_iterator=batch_generator,
                model=self.reward_model_module,
                num_microbatches=n_micro_batch,
                seq_length=infer_batch_size * seq_len,  # in use for pp = 1
                micro_batch_size=1,  # in use for pp = 1
                forward_only=True,
            )
        # loss_reduces contains the stats returned from loss_func

        return losses_reduced

    def offload_params_to_cpu(self):
        if self.device == "cuda":
            for reward_model_module in self.reward_model_module:
                for name, param in reward_model_module.named_parameters():
                    param.data = param.data.to("cpu", non_blocking=True)
            self.device = "cpu"
            torch.cuda.empty_cache()

    def load_params_to_cuda(self):
        if self.device == "cpu":
            for reward_model_module in self.reward_model_module:
                for name, param in reward_model_module.named_parameters():
                    param.data = param.data.to(torch.cuda.current_device(), non_blocking=True)
            self.device = "cuda"