modeling_molformer.py

# coding=utf-8
# Copyright 2023 The HuggingFace Inc. team. 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.
""" PyTorch Molformer model."""


import math
from typing import Optional, Tuple, Union

import torch
import torch.utils.checkpoint
from torch import nn
from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss

from transformers.activations import ACT2FN
from transformers.modeling_outputs import (
    BaseModelOutput,
    BaseModelOutputWithPooling,
    MaskedLMOutput,
    SequenceClassifierOutput,
)
from transformers.modeling_utils import PreTrainedModel
from transformers.pytorch_utils import apply_chunking_to_forward, find_pruneable_heads_and_indices, prune_linear_layer
from transformers.utils import (
    add_code_sample_docstrings,
    add_start_docstrings,
    add_start_docstrings_to_model_forward,
    logging,
)
from .configuration_molformer import MolformerConfig


logger = logging.get_logger(__name__)

_CHECKPOINT_FOR_DOC = "ibm/MoLFormer-XL-both-10pct"
_CONFIG_FOR_DOC = "MolformerConfig"

MOLFORMER_PRETRAINED_MODEL_ARCHIVE_LIST = [
    "ibm/MoLFormer-XL-both-10pct",
    # See all MoLFormer models at https://huggingface.co/models?filter=molformer
]


# Copied from transformers.models.esm.modeling_esm.rotate_half
def rotate_half(x):
    x1, x2 = x.chunk(2, dim=-1)
    return torch.cat((-x2, x1), dim=-1)


# Copied from transformers.models.llama.modeling_llama.apply_rotary_pos_emb
def apply_rotary_pos_emb(q, k, cos, sin, position_ids):
    cos = cos[position_ids].unsqueeze(1)  # [seq_len, dim] -> [batch_size, 1, seq_len, head_dim]
    sin = sin[position_ids].unsqueeze(1)
    q_embed = (q * cos) + (rotate_half(q) * sin)
    k_embed = (k * cos) + (rotate_half(k) * sin)
    return q_embed, k_embed


# Copied from transformers.models.llama.modeling_llama.LlamaRotaryEmbedding with Llama->Molformer
class MolformerRotaryEmbedding(nn.Module):
    def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None):
        super().__init__()

        self.dim = dim
        self.max_position_embeddings = max_position_embeddings
        self.base = base
        inv_freq = 1.0 / (self.base ** (torch.arange(0, self.dim, 2).float().to(device) / self.dim))
        self.register_buffer("inv_freq", inv_freq, persistent=False)

        # Build here to make `torch.jit.trace` work.
        self._set_cos_sin_cache(
            seq_len=max_position_embeddings, device=self.inv_freq.device, dtype=torch.get_default_dtype()
        )

    def _set_cos_sin_cache(self, seq_len, device, dtype):
        self.max_seq_len_cached = seq_len
        t = torch.arange(self.max_seq_len_cached, device=device, dtype=self.inv_freq.dtype)

        freqs = torch.einsum("i,j->ij", t, self.inv_freq)
        # Different from paper, but it uses a different permutation in order to obtain the same calculation
        emb = torch.cat((freqs, freqs), dim=-1)
        self.register_buffer("cos_cached", emb.cos().to(dtype), persistent=False)
        self.register_buffer("sin_cached", emb.sin().to(dtype), persistent=False)

    def forward(self, x, seq_len=None):
        # x: [bs, num_attention_heads, seq_len, head_size]
        if seq_len > self.max_seq_len_cached:
            self._set_cos_sin_cache(seq_len=seq_len, device=x.device, dtype=x.dtype)

        return (
            self.cos_cached[:seq_len].to(dtype=x.dtype),
            self.sin_cached[:seq_len].to(dtype=x.dtype),
        )


class MolformerEmbeddings(nn.Module):
    """Construct the embeddings from word embeddings."""

    def __init__(self, config):
        super().__init__()
        self.word_embeddings = nn.Embedding(config.vocab_size, config.hidden_size, padding_idx=config.pad_token_id)
        self.dropout = nn.Dropout(config.embedding_dropout_prob)

    def forward(
        self, input_ids: Optional[torch.LongTensor] = None, inputs_embeds: Optional[torch.FloatTensor] = None
    ) -> torch.Tensor:
        if inputs_embeds is None:
            inputs_embeds = self.word_embeddings(input_ids)

        embeddings = inputs_embeds
        embeddings = self.dropout(embeddings)
        return embeddings


class MolformerFeatureMap(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.query_size = config.hidden_size // config.num_attention_heads
        self.num_components = config.num_random_features
        self.orthogonal_random_weights()
        if isinstance(config.feature_map_kernel, str):
            self.kernel = ACT2FN[config.feature_map_kernel]
        else:
            self.kernel = config.feature_map_kernel
        self.deterministic = config.deterministic_eval

    def orthogonal_random_weights(self, device=None):
        # make sure query size evenly divides feature size (round up)
        num_batches = math.ceil(self.num_components / self.query_size)

        def orthogonal_batch(size):
            block = torch.randn(size, size, device=device)
            norms = torch.linalg.norm(block, dim=1).unsqueeze(0)
            Q, _ = torch.linalg.qr(block)
            return Q * norms

        random_weights = torch.cat([orthogonal_batch(self.query_size) for _ in range(num_batches)], dim=1)
        random_weights = random_weights[:, : self.num_components]
        self.register_buffer("weight", random_weights)

    def forward(self, query, key):
        if not self.deterministic or self.training:
            self.orthogonal_random_weights(query.device)
        # generalized random fourier features
        query = torch.matmul(query, self.weight)
        key = torch.matmul(key, self.weight)
        return self.kernel(query), self.kernel(key)


class MolformerSelfAttention(nn.Module):
    def __init__(self, config):
        super().__init__()
        if config.hidden_size % config.num_attention_heads != 0 and not hasattr(config, "embedding_size"):
            raise ValueError(
                f"The hidden size ({config.hidden_size}) is not a multiple of the number of attention "
                f"heads ({config.num_attention_heads})"
            )

        self.num_attention_heads = config.num_attention_heads
        self.attention_head_size = int(config.hidden_size / config.num_attention_heads)
        self.all_head_size = self.num_attention_heads * self.attention_head_size

        self.query = nn.Linear(config.hidden_size, self.all_head_size)
        self.key = nn.Linear(config.hidden_size, self.all_head_size)
        self.value = nn.Linear(config.hidden_size, self.all_head_size)

        self.eps = config.linear_attention_eps

        self.rotary_embeddings = MolformerRotaryEmbedding(
            dim=self.attention_head_size, max_position_embeddings=config.max_position_embeddings
        )
        self.feature_map = MolformerFeatureMap(config)

    # Copied from transformers.models.bert.modeling_bert.BertSelfAttention.transpose_for_scores
    def transpose_for_scores(self, x: torch.Tensor) -> torch.Tensor:
        new_x_shape = x.size()[:-1] + (self.num_attention_heads, self.attention_head_size)
        x = x.view(new_x_shape)
        return x.permute(0, 2, 1, 3)

    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.FloatTensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        output_attentions: Optional[bool] = False,
    ) -> Tuple[torch.Tensor]:
        query_layer = self.transpose_for_scores(self.query(hidden_states))
        key_layer = self.transpose_for_scores(self.key(hidden_states))
        value_layer = self.transpose_for_scores(self.value(hidden_states))

        kv_seq_len = key_layer.shape[-2]
        cos, sin = self.rotary_embeddings(value_layer, seq_len=kv_seq_len)
        query_layer, key_layer = apply_rotary_pos_emb(query_layer, key_layer, cos, sin, position_ids)
        # Apply the feature map to the queries and keys
        query_layer, key_layer = self.feature_map(query_layer, key_layer)

        if attention_mask is not None:
            # since we don't use softmax, we need to reconvert this mask to 1/0
            attention_mask = (attention_mask == 0).to(attention_mask.dtype)
            # separate original mask from causal mask
            per_query_attn = attention_mask[:, 0, -1]
            per_query_extended = per_query_attn[:, None, None, :]
            if not torch.equal(attention_mask, per_query_extended):
                raise ValueError(
                    "MolformerSelfAttention does not support arbitrary 3D attention. attention_mask must be 2D (i.e., [batch size, sequence length])"
                )

            key_layer = key_layer * per_query_attn[:, None, -kv_seq_len:, None]

        # linear attention
        key_value = torch.matmul(key_layer.transpose(-1, -2), value_layer)
        norm = torch.matmul(query_layer, key_layer.sum(dim=-2).unsqueeze(-1)).clamp(min=self.eps)
        context_layer = torch.matmul(query_layer, key_value) / norm

        if head_mask is not None:
            context_layer = context_layer * head_mask

        if output_attentions:
            logger.warning(
                "Outputting attentions in linear attention negates the efficiency gains! Only use for visualization/debugging."
            )
            attention_scores = torch.matmul(query_layer, key_layer.transpose(-1, -2))
            if attention_mask is not None:
                attention_scores = attention_scores * attention_mask
            attention_probs = nn.functional.normalize(attention_scores, p=1, dim=-1, eps=self.eps)
            if head_mask is not None:
                attention_probs = attention_probs * head_mask
            # recompute context_layer for grad
            context_layer = torch.matmul(attention_probs, value_layer)

        context_layer = context_layer.permute(0, 2, 1, 3).contiguous()
        new_context_layer_shape = context_layer.size()[:-2] + (self.all_head_size,)
        context_layer = context_layer.view(*new_context_layer_shape)

        outputs = (context_layer, attention_probs) if output_attentions else (context_layer,)

        return outputs


# Copied from transformers.models.bert.modeling_bert.BertSelfOutput
class MolformerSelfOutput(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)

    def forward(self, hidden_states: torch.Tensor, input_tensor: torch.Tensor) -> torch.Tensor:
        hidden_states = self.dense(hidden_states)
        hidden_states = self.dropout(hidden_states)
        hidden_states = self.LayerNorm(hidden_states + input_tensor)
        return hidden_states


class MolformerAttention(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.self = MolformerSelfAttention(config)
        self.output = MolformerSelfOutput(config)
        self.pruned_heads = set()

    # Copied from transformers.models.bert.modeling_bert.BertAttention.prune_heads
    def prune_heads(self, heads):
        if len(heads) == 0:
            return
        heads, index = find_pruneable_heads_and_indices(
            heads, self.self.num_attention_heads, self.self.attention_head_size, self.pruned_heads
        )

        # Prune linear layers
        self.self.query = prune_linear_layer(self.self.query, index)
        self.self.key = prune_linear_layer(self.self.key, index)
        self.self.value = prune_linear_layer(self.self.value, index)
        self.output.dense = prune_linear_layer(self.output.dense, index, dim=1)

        # Update hyper params and store pruned heads
        self.self.num_attention_heads = self.self.num_attention_heads - len(heads)
        self.self.all_head_size = self.self.attention_head_size * self.self.num_attention_heads
        self.pruned_heads = self.pruned_heads.union(heads)

    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.FloatTensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        output_attentions: Optional[bool] = False,
    ) -> Tuple[torch.Tensor]:
        self_outputs = self.self(
            hidden_states,
            attention_mask,
            position_ids,
            head_mask,
            output_attentions,
        )
        attention_output = self.output(self_outputs[0], hidden_states)
        outputs = (attention_output,) + self_outputs[1:]  # add attentions if we output them
        return outputs


# Copied from transformers.models.bert.modeling_bert.BertIntermediate
class MolformerIntermediate(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.dense = nn.Linear(config.hidden_size, config.intermediate_size)
        if isinstance(config.hidden_act, str):
            self.intermediate_act_fn = ACT2FN[config.hidden_act]
        else:
            self.intermediate_act_fn = config.hidden_act

    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
        hidden_states = self.dense(hidden_states)
        hidden_states = self.intermediate_act_fn(hidden_states)
        return hidden_states


# Copied from transformers.models.bert.modeling_bert.BertOutput
class MolformerOutput(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.dense = nn.Linear(config.intermediate_size, config.hidden_size)
        self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)

    def forward(self, hidden_states: torch.Tensor, input_tensor: torch.Tensor) -> torch.Tensor:
        hidden_states = self.dense(hidden_states)
        hidden_states = self.dropout(hidden_states)
        hidden_states = self.LayerNorm(hidden_states + input_tensor)
        return hidden_states


class MolformerLayer(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.chunk_size_feed_forward = config.chunk_size_feed_forward
        self.seq_len_dim = 1
        self.attention = MolformerAttention(config)
        self.intermediate = MolformerIntermediate(config)
        self.output = MolformerOutput(config)

    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.FloatTensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        output_attentions: Optional[bool] = False,
    ) -> Tuple[torch.Tensor]:
        self_attention_outputs = self.attention(
            hidden_states,
            attention_mask,
            position_ids,
            head_mask,
            output_attentions=output_attentions,
        )
        attention_output = self_attention_outputs[0]
        outputs = self_attention_outputs[1:]  # add self attentions if we output attention weights

        layer_output = apply_chunking_to_forward(
            self.feed_forward_chunk, self.chunk_size_feed_forward, self.seq_len_dim, attention_output
        )
        outputs = (layer_output,) + outputs

        return outputs

    def feed_forward_chunk(self, attention_output):
        intermediate_output = self.intermediate(attention_output)
        layer_output = self.output(intermediate_output, attention_output)
        return layer_output


class MolformerEncoder(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.config = config
        self.layer = nn.ModuleList([MolformerLayer(config) for _ in range(config.num_hidden_layers)])
        self.gradient_checkpointing = False

    def forward(
        self,
        hidden_states: torch.Tensor,
        attention_mask: Optional[torch.FloatTensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        output_attentions: Optional[bool] = False,
        output_hidden_states: Optional[bool] = False,
        return_dict: Optional[bool] = True,
    ) -> Union[Tuple[torch.Tensor], BaseModelOutput]:
        all_hidden_states = () if output_hidden_states else None
        all_self_attentions = () if output_attentions else None

        for i, layer_module in enumerate(self.layer):
            if output_hidden_states:
                all_hidden_states = all_hidden_states + (hidden_states,)

            layer_head_mask = head_mask[i] if head_mask is not None else None

            if self.gradient_checkpointing and self.training:

                def create_custom_forward(module):
                    def custom_forward(*inputs):
                        return module(*inputs, output_attentions)

                    return custom_forward

                layer_outputs = torch.utils.checkpoint.checkpoint(
                    create_custom_forward(layer_module),
                    hidden_states,
                    attention_mask,
                    position_ids,
                    layer_head_mask,
                )
            else:
                layer_outputs = layer_module(
                    hidden_states,
                    attention_mask,
                    position_ids,
                    layer_head_mask,
                    output_attentions,
                )

            hidden_states = layer_outputs[0]
            if output_attentions:
                all_self_attentions = all_self_attentions + (layer_outputs[1],)

        if output_hidden_states:
            all_hidden_states = all_hidden_states + (hidden_states,)

        if not return_dict:
            return tuple(
                v
                for v in [
                    hidden_states,
                    all_hidden_states,
                    all_self_attentions,
                ]
                if v is not None
            )
        return BaseModelOutput(
            last_hidden_state=hidden_states,
            hidden_states=all_hidden_states,
            attentions=all_self_attentions,
        )


# Copied from transformers.models.bert.modeling_bert.BertPredictionHeadTransform
class MolformerPredictionHeadTransform(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        if isinstance(config.hidden_act, str):
            self.transform_act_fn = ACT2FN[config.hidden_act]
        else:
            self.transform_act_fn = config.hidden_act
        self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)

    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
        hidden_states = self.dense(hidden_states)
        hidden_states = self.transform_act_fn(hidden_states)
        hidden_states = self.LayerNorm(hidden_states)
        return hidden_states


class MolformerLMPredictionHead(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.transform = MolformerPredictionHeadTransform(config)
        self.decoder = nn.Linear(config.hidden_size, config.vocab_size, bias=False)

    def forward(self, hidden_states):
        hidden_states = self.transform(hidden_states)
        hidden_states = self.decoder(hidden_states)
        return hidden_states


# Copied from transformers.models.roberta.modeling_roberta.RobertaPreTrainedModel with Roberta->Molformer,roberta->molformer
class MolformerPreTrainedModel(PreTrainedModel):
    """
    An abstract class to handle weights initialization and a simple interface for downloading and loading pretrained
    models.
    """

    config_class = MolformerConfig
    base_model_prefix = "molformer"
    supports_gradient_checkpointing = True
    _no_split_modules = ["MolformerEmbeddings", "MolformerSelfAttention"]

    # Copied from transformers.models.bert.modeling_bert.BertPreTrainedModel._init_weights
    def _init_weights(self, module):
        """Initialize the weights"""
        if isinstance(module, nn.Linear):
            # Slightly different from the TF version which uses truncated_normal for initialization
            # cf https://github.com/pytorch/pytorch/pull/5617
            module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
            if module.bias is not None:
                module.bias.data.zero_()
        elif isinstance(module, nn.Embedding):
            module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
            if module.padding_idx is not None:
                module.weight.data[module.padding_idx].zero_()
        elif isinstance(module, nn.LayerNorm):
            module.bias.data.zero_()
            module.weight.data.fill_(1.0)

    def _set_gradient_checkpointing(self, module, value=False):
        if isinstance(module, MolformerEncoder):
            module.gradient_checkpointing = value


def masked_avg_pool1d(hidden_states, attention_mask, eps=1e-9):
    attention_mask = attention_mask.unsqueeze(-1).expand_as(hidden_states).float()
    sum_embeddings = torch.sum(hidden_states * attention_mask, dim=1)
    sum_mask = torch.clamp(attention_mask.sum(dim=1), min=eps)
    embedding = sum_embeddings / sum_mask
    return embedding


MOLFORMER_START_DOCSTRING = r"""
    This model is a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) sub-class. Use
    it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and
    behavior.

    Parameters:
        config ([`MolformerConfig`]): Model configuration class with all the parameters of the model.
            Initializing with a config file does not load the weights associated with the model, only the
            configuration. Check out the [`~PreTrainedModel.from_pretrained`] method to load the model weights.
"""

MOLFORMER_INPUTS_DOCSTRING = r"""
    Args:
        input_ids (`torch.LongTensor` of shape `({0})`):
            Indices of input sequence tokens in the vocabulary.

            Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
            [`PreTrainedTokenizer.__call__`] for details.

            [What are input IDs?](../glossary#input-ids)
        attention_mask (`torch.FloatTensor` of shape `({0})`, *optional*):
            Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:

            - 1 for tokens that are **not masked**,
            - 0 for tokens that are **masked**.

            [What are attention masks?](../glossary#attention-mask)
        position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
            config.n_positions - 1]`.

            [What are position IDs?](../glossary#position-ids)
        head_mask (`torch.FloatTensor` of shape `(num_heads,)` or `(num_layers, num_heads)`, *optional*):
            Mask to nullify selected heads of the self-attention modules. Mask values selected in `[0, 1]`:

            - 1 indicates the head is **not masked**,
            - 0 indicates the head is **masked**.

        inputs_embeds (`torch.FloatTensor` of shape `({0}, hidden_size)`, *optional*):
            Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
            is useful if you want more control over how to convert *input_ids* indices into associated vectors than the
            model's internal embedding lookup matrix.
        output_attentions (`bool`, *optional*):
            Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
            tensors for more detail.
        output_hidden_states (`bool`, *optional*):
            Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
            more detail.
        return_dict (`bool`, *optional*):
            Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
"""


@add_start_docstrings(
    "The bare Molformer Model transformer outputting raw hidden-states without any specific head on top.",
    MOLFORMER_START_DOCSTRING,
    """
        add_pooling_layer (`bool`, *optional*, defaults to `True`):
            Whether or not to apply pooling layer.
    """,
)
class MolformerModel(MolformerPreTrainedModel):
    """

    The model can behave as an encoder (with only self-attention).
    """

    def __init__(self, config, add_pooling_layer=True):
        super().__init__(config)
        self.config = config

        self.embeddings = MolformerEmbeddings(config)
        self.encoder = MolformerEncoder(config)

        self.LayerNorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
        self.pooler = masked_avg_pool1d if add_pooling_layer else None

        # Initialize weights and apply final processing
        self.post_init()

    def get_input_embeddings(self):
        return self.embeddings.word_embeddings

    def set_input_embeddings(self, value):
        self.embeddings.word_embeddings = value

    def _prune_heads(self, heads_to_prune):
        """
        Prunes heads of the model. heads_to_prune: dict of {layer_num: list of heads to prune in this layer} See base
        class PreTrainedModel
        """
        for layer, heads in heads_to_prune.items():
            self.encoder.layer[layer].attention.prune_heads(heads)

    @add_start_docstrings_to_model_forward(MOLFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
    @add_code_sample_docstrings(
        checkpoint=_CHECKPOINT_FOR_DOC,
        output_type=BaseModelOutputWithPooling,
        config_class=_CONFIG_FOR_DOC,
    )
    def forward(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        attention_mask: Optional[torch.FloatTensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        return_dict: Optional[bool] = None,
    ) -> Union[BaseModelOutputWithPooling, Tuple[torch.Tensor]]:
        output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
        output_hidden_states = (
            output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
        )
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict

        if input_ids is not None and inputs_embeds is not None:
            raise ValueError("You cannot specify both input_ids and inputs_embeds at the same time")
        elif input_ids is not None:
            self.warn_if_padding_and_no_attention_mask(input_ids, attention_mask)
            input_shape = input_ids.size()
        elif inputs_embeds is not None:
            input_shape = inputs_embeds.size()[:-1]
        else:
            raise ValueError("You have to specify either input_ids or inputs_embeds")

        batch_size, seq_length = input_shape
        device = input_ids.device if input_ids is not None else inputs_embeds.device

        if position_ids is None:
            position_ids = torch.arange(seq_length, dtype=torch.long, device=device)
            position_ids = position_ids.unsqueeze(0).view(-1, seq_length)
        else:
            position_ids = position_ids.view(-1, seq_length).long()

        if attention_mask is None:
            attention_mask = torch.ones((batch_size, seq_length), device=device)

        # We can provide a self-attention mask of dimensions [batch_size, from_seq_length, to_seq_length]
        # ourselves in which case we just need to make it broadcastable to all heads.
        extended_attention_mask: torch.Tensor = self.get_extended_attention_mask(attention_mask, input_shape)

        # Prepare head mask if needed
        # 1.0 in head_mask indicate we keep the head
        # attention_probs has shape bsz x n_heads x N x N
        # input head_mask has shape [num_heads] or [num_hidden_layers x num_heads]
        # and head_mask is converted to shape [num_hidden_layers x batch x num_heads x seq_length x seq_length]
        head_mask = self.get_head_mask(head_mask, self.config.num_hidden_layers)

        embedding_output = self.embeddings(input_ids=input_ids, inputs_embeds=inputs_embeds)

        encoder_outputs = self.encoder(
            embedding_output,
            attention_mask=extended_attention_mask,
            position_ids=position_ids,
            head_mask=head_mask,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
        )
        sequence_output = encoder_outputs[0]
        sequence_output = self.LayerNorm(sequence_output)
        pooled_output = self.pooler(sequence_output, attention_mask) if self.pooler is not None else None

        if not return_dict:
            return (sequence_output, pooled_output) + encoder_outputs[1:]

        return BaseModelOutputWithPooling(
            last_hidden_state=sequence_output,
            pooler_output=pooled_output,
            hidden_states=encoder_outputs.hidden_states,
            attentions=encoder_outputs.attentions,
        )


@add_start_docstrings("""Molformer Model with a `language modeling` head on top.""", MOLFORMER_START_DOCSTRING)
class MolformerForMaskedLM(MolformerPreTrainedModel):
    _tied_weights_keys = ["lm_head.decoder.weight"]

    # Copied from transformers.models.roberta.modeling_roberta.RobertaForMaskedLM.__init__ with Roberta->Molformer,roberta->molformer,LMHead->LMPredictionHead
    def __init__(self, config):
        super().__init__(config)

        if config.is_decoder:
            logger.warning(
                "If you want to use `MolformerForMaskedLM` make sure `config.is_decoder=False` for "
                "bi-directional self-attention."
            )

        self.molformer = MolformerModel(config, add_pooling_layer=False)
        self.lm_head = MolformerLMPredictionHead(config)

        # Initialize weights and apply final processing
        self.post_init()

    def get_output_embeddings(self):
        return self.lm_head.decoder

    def set_output_embeddings(self, new_embeddings):
        self.lm_head.decoder = new_embeddings

    @add_start_docstrings_to_model_forward(MOLFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
    @add_code_sample_docstrings(
        checkpoint=_CHECKPOINT_FOR_DOC,
        output_type=MaskedLMOutput,
        config_class=_CONFIG_FOR_DOC,
        mask="P<mask>",  # add extra token so labels line up
    )
    def forward(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        attention_mask: Optional[torch.FloatTensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        encoder_hidden_states: Optional[torch.FloatTensor] = None,
        encoder_attention_mask: Optional[torch.FloatTensor] = None,
        labels: Optional[torch.LongTensor] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        return_dict: Optional[bool] = None,
    ) -> Union[MaskedLMOutput, Tuple[torch.Tensor]]:
        r"""
        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Labels for computing the masked language modeling loss. Indices should be in `[-100, 0, ...,
            config.vocab_size]` (see `input_ids` docstring) Tokens with indices set to `-100` are ignored (masked), the
            loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.
        """
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict

        outputs = self.molformer(
            input_ids,
            attention_mask=attention_mask,
            position_ids=position_ids,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
        )

        sequence_output = outputs[0]
        prediction_scores = self.lm_head(sequence_output)

        masked_lm_loss = None
        if labels is not None:
            # move labels to correct device to enable model parallelism
            labels = labels.to(prediction_scores.device)
            loss_fct = CrossEntropyLoss()  # -100 index = padding token
            masked_lm_loss = loss_fct(prediction_scores.view(-1, self.config.vocab_size), labels.view(-1))

        if not return_dict:
            output = (prediction_scores,) + outputs[2:]
            return ((masked_lm_loss,) + output) if masked_lm_loss is not None else output

        return MaskedLMOutput(
            loss=masked_lm_loss,
            logits=prediction_scores,
            hidden_states=outputs.hidden_states,
            attentions=outputs.attentions,
        )


class MolformerClassificationHead(nn.Module):
    """Head for sequence-level classification tasks."""

    def __init__(self, config):
        super().__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.dense2 = nn.Linear(config.hidden_size, config.hidden_size)
        self.dropout = nn.Dropout(
            config.classifier_dropout_prob
            if config.classifier_dropout_prob is not None
            else config.hidden_dropout_prob
        )
        self.out_proj = nn.Linear(config.hidden_size, config.num_labels)
        if isinstance(config.hidden_act, str):
            self.classifier_act_fn = ACT2FN[config.hidden_act]
        else:
            self.classifier_act_fn = config.hidden_act
        self.skip_connection = config.classifier_skip_connection

    def forward(self, pooled_output):
        hidden_state = self.dense(pooled_output)
        hidden_state = self.dropout(hidden_state)
        hidden_state = self.classifier_act_fn(hidden_state)
        if self.skip_connection:
            hidden_state = residual = hidden_state + pooled_output
        hidden_state = self.dense2(hidden_state)
        hidden_state = self.dropout(hidden_state)
        hidden_state = self.classifier_act_fn(hidden_state)
        if self.skip_connection:
            hidden_state = hidden_state + residual
        logits = self.out_proj(hidden_state)
        return logits


@add_start_docstrings(
    """
    Molformer Model transformer with a sequence classification/regression head on top (a linear layer on top of the
    pooled output) e.g. for MoleculeNet tasks.
    """,
    MOLFORMER_START_DOCSTRING,
)
class MolformerForSequenceClassification(MolformerPreTrainedModel):
    def __init__(self, config):
        super().__init__(config)
        self.num_labels = config.num_labels
        self.config = config

        self.molformer = MolformerModel(config, add_pooling_layer=True)
        self.classifier = MolformerClassificationHead(config)

        # Initialize weights and apply final processing
        self.post_init()

    @add_start_docstrings_to_model_forward(MOLFORMER_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
    @add_code_sample_docstrings(
        checkpoint=_CHECKPOINT_FOR_DOC,
        output_type=SequenceClassifierOutput,
        config_class=_CONFIG_FOR_DOC,
    )
    def forward(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        attention_mask: Optional[torch.FloatTensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        labels: Optional[torch.LongTensor] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        return_dict: Optional[bool] = None,
    ) -> Union[Tuple[torch.Tensor], SequenceClassifierOutput]:
        r"""
        labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
            Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
            config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
            `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
        """
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict

        outputs = self.molformer(
            input_ids,
            attention_mask=attention_mask,
            position_ids=position_ids,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
        )

        pooled_output = outputs[1]
        logits = self.classifier(pooled_output)

        loss = None
        if labels is not None:
            # move labels to correct device to enable model parallelism
            labels = labels.to(logits.device)
            if self.config.problem_type is None:
                if self.num_labels == 1:
                    self.config.problem_type = "regression"
                elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int):
                    self.config.problem_type = "single_label_classification"
                else:
                    self.config.problem_type = "multi_label_classification"

            if self.config.problem_type == "regression":
                loss_fct = MSELoss()
                if self.num_labels == 1:
                    loss = loss_fct(logits.squeeze(), labels.squeeze())
                else:
                    loss = loss_fct(logits, labels)
            elif self.config.problem_type == "single_label_classification":
                loss_fct = CrossEntropyLoss()
                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
            elif self.config.problem_type == "multi_label_classification":
                loss_fct = BCEWithLogitsLoss()
                loss = loss_fct(logits, labels)

        if not return_dict:
            output = (logits,) + outputs[2:]
            return ((loss,) + output) if loss is not None else output

        return SequenceClassifierOutput(
            loss=loss,
            logits=logits,
            hidden_states=outputs.hidden_states,
            attentions=outputs.attentions,
        )