コード例 #1
0
ファイル: cross_entropy.py プロジェクト: myleott/Megatron-LM
    def forward(ctx, vocab_parallel_logits, target):

        # Copy so the input remains unchanged.
        logits = vocab_parallel_logits.clone()
        # Maximum value along vocab dimension across all GPUs.
        logits_max = torch.max(logits, dim=-1)[0]
        all_reduce(logits_max, op='max', group=get_model_parallel_group())
        # Subtract the maximum value.
        logits.sub_(logits_max.unsqueeze(dim=-1))
        # Sum of exponential of logits along vocab dimension across all GPUs.
        exp_logits = logits.exp()
        sum_exp_logits = exp_logits.sum(dim=-1)
        all_reduce(sum_exp_logits, op='sum', group=get_model_parallel_group())

        # Get the partition's vocab indecies
        get_vocab_range = VocabUtility.vocab_range_from_per_partition_vocab_size
        partition_vocab_size = vocab_parallel_logits.size()[-1]
        rank = get_model_parallel_rank()
        world_size = get_model_parallel_world_size()
        vocab_start_index, vocab_end_index = get_vocab_range(
            partition_vocab_size, rank, world_size)

        # Create a mask of valid vocab ids (1 means it needs to be masked).
        target_mask = (target < vocab_start_index) | (target >=
                                                      vocab_end_index)
        masked_target = target.clone() - vocab_start_index
        masked_target[target_mask] = 0

        # Get predicted-logits = logits[target].
        # For Simplicity, we convert logits to a 2-D tensor with size
        # [*, partition-vocab-size] and target to a 1-D tensor of size [*].
        logits_2d = logits.view(-1, partition_vocab_size)
        masked_target_1d = masked_target.view(-1)
        arange_1d = torch.arange(start=0,
                                 end=logits_2d.size()[0],
                                 device=logits_2d.device)
        predicted_logits_1d = logits_2d[arange_1d, masked_target_1d]
        predicted_logits = predicted_logits_1d.view_as(target)
        predicted_logits[target_mask] = 0.0
        # All reduce is needed to get the chunks from other GPUs.
        all_reduce(predicted_logits,
                   op='sum',
                   group=get_model_parallel_group())

        # Loss = log(sum(exp(logits))) - predicted-logit.
        loss = torch.log(sum_exp_logits) - predicted_logits

        # Store softmax, target-mask and masked-target for backward pass.
        exp_logits.div_(sum_exp_logits.unsqueeze(dim=-1))
        ctx.save_for_backward(exp_logits, target_mask, masked_target_1d)

        return loss
コード例 #2
0
def _reduce(input_):
    """All-reduce the the input tensor across model parallel group."""
    group = get_model_parallel_group()

    # Bypass the function if we are using only 1 GPU.
    if get_world_size(group=group) == 1:
        return input_

    # All-reduce.
    all_reduce(input_, group=group)

    return input_
コード例 #3
0
ファイル: mappings.py プロジェクト: fairseq/Megatron-LM
def _gather(input_):
    """Gather tensors and concatinate along the last dimension."""
    group = get_model_parallel_group()

    # Bypass the function if we are using only 1 GPU.
    if get_world_size(group=group) == 1:
        return input_

    # Size and dimension.
    last_dim = input_.dim() - 1

    tensor_list = all_gather(tensor=input_, group=group)

    # Note: torch.cat already creates a contiguous tensor.
    output = torch.cat(tensor_list, dim=last_dim).contiguous()

    return output
コード例 #4
0
def _split(input_):
    """Split the tensor along its last dimension and keep the
    corresponding slice."""
    group = get_model_parallel_group()

    # Bypass the function if we are using only 1 GPU.
    if get_world_size(group=group) == 1:
        return input_

    # Split along last dimension.
    world_size = get_world_size(group=group)
    input_list = split_tensor_along_last_dim(input_, world_size)

    # Note: torch.split does not create contiguous tensors by default.
    rank = get_rank(group=group)
    output = input_list[rank].contiguous()

    return output
コード例 #5
0
 def _aggregate_model_parallel_grad_norm(total_norm):
     total_norm = total_norm**2
     distributed_utils.all_reduce(
         total_norm, group=distributed_utils.get_model_parallel_group())
     total_norm = total_norm**0.5
     return total_norm