gpt2.py

from dataclasses import dataclass
from typing import Optional, Tuple, Union

import torch
import torch.utils.checkpoint
from torch import nn
from torch.cuda.amp import autocast
from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss

from transformers import GPT2LMHeadModel
from transformers.models.gpt2 import GPT2Model
from transformers.models.gpt2.modeling_gpt2 import GPT2Block, GPT2Attention, GPT2MLP
from transformers.modeling_outputs import CausalLMOutputWithCrossAttentions, BaseModelOutputWithPastAndCrossAttentions
from transformers.pytorch_utils import Conv1D

from ttc import map_values_to_closest_labels_nolog


def one_hot_encode(input_ids):
    unique_values, unique_indices = torch.unique(input_ids, return_inverse=True)
    one_hot_tensor = torch.zeros(input_ids.shape[0], input_ids.shape[1], unique_values.numel())
    one_hot_tensor = one_hot_tensor.to(input_ids.device)
    one_hot_tensor = one_hot_tensor.scatter_(2, unique_indices.unsqueeze(2), 1)
    return one_hot_tensor, unique_values

class GPT2TTCAttention(GPT2Attention):
    def __init__(self, config, is_cross_attention=False, layer_idx=None):
        super().__init__(config, is_cross_attention, layer_idx)

        # TODO: This is a hack to avoid overrunning the size of bias with TTC tokens
        # need to move to model config
        MAX_UNIQUE_TTC_TOKENS = 20
        max_positions = config.max_position_embeddings
        tril_dim = max_positions+MAX_UNIQUE_TTC_TOKENS
        self.register_buffer(
            "bias",
            torch.tril(torch.ones((tril_dim, tril_dim), dtype=torch.bool)).view(1,1,tril_dim,tril_dim),
            persistent=False,
        )
        self.register_buffer("masked_bias", torch.tensor(-1e4), persistent=False)

        self.embed_dim = config.hidden_size
        self.num_heads = config.num_attention_heads
        self.head_dim = self.embed_dim // self.num_heads
        self.split_size = self.embed_dim
        if self.head_dim * self.num_heads != self.embed_dim:
            raise ValueError(
                f"`embed_dim` must be divisible by num_heads (got `embed_dim`: {self.embed_dim} and `num_heads`:"
                f" {self.num_heads})."
            )

        self.scale_attn_weights = config.scale_attn_weights
        self.is_cross_attention = is_cross_attention

        # Layer-wise attention scaling, reordering, and upcasting
        self.scale_attn_by_inverse_layer_idx = config.scale_attn_by_inverse_layer_idx
        self.layer_idx = layer_idx
        self.reorder_and_upcast_attn = config.reorder_and_upcast_attn

        if self.is_cross_attention:
            self.c_attn = Conv1D(2 * self.embed_dim, self.embed_dim)
            self.q_attn = Conv1D(self.embed_dim, self.embed_dim)
        else:
            self.c_attn = Conv1D(3 * self.embed_dim, self.embed_dim)
        self.c_proj = Conv1D(self.embed_dim, self.embed_dim)

        self.attn_dropout = nn.Dropout(config.attn_pdrop)
        self.resid_dropout = nn.Dropout(config.resid_pdrop)

        self.pruned_heads = set()

    def _attn(self, query, key, value, one_hot_ttc=None, attention_mask=None, head_mask=None):
        attn_weights = torch.matmul(query, key.transpose(-1, -2))
        #print(f'attn_weights (source): {attn_weights.shape}')

        if self.scale_attn_weights:
            attn_weights = attn_weights / torch.full(
                [], value.size(-1) ** 0.5, dtype=attn_weights.dtype, device=attn_weights.device
            )

        # Layer-wise attention scaling
        if self.scale_attn_by_inverse_layer_idx:
            attn_weights = attn_weights / float(self.layer_idx + 1)

        if not self.is_cross_attention:
            # if only "normal" attention layer implements causal mask
            query_length, key_length = query.size(-2), key.size(-2)
            causal_mask = self.bias[:, :, key_length - query_length : key_length, :key_length]
            # NOTE: The above logic makes the lower tril of the causal mask "shift right" when there are fewer queries than keys. Removing queries chops rows off of the top of the causal mask.
            
            if one_hot_ttc != None:
                #print(f'one_hot_ttc: {one_hot_ttc.shape}')
                #print(f'causal_mask: {causal_mask.shape}')
                n_ttc = one_hot_ttc.shape[-1]
                # NOTE: Overwrite end of causal mask to attend TTC in appropriate locations
                causal_mask = causal_mask.expand(attn_weights.shape[0], -1, -1, -1).clone()
                causal_mask[:,:,:,:n_ttc] = one_hot_ttc.unsqueeze(1)

            mask_value = torch.finfo(attn_weights.dtype).min
            # Need to be a tensor, otherwise we get error: `RuntimeError: expected scalar type float but found double`.
            # Need to be on the same device, otherwise `RuntimeError: ..., x and y to be on the same device`
            mask_value = torch.full([], mask_value, dtype=attn_weights.dtype, device=attn_weights.device)
            attn_weights = torch.where(causal_mask, attn_weights.to(attn_weights.dtype), mask_value)

        if attention_mask is not None:
            # Apply the attention mask
            if one_hot_ttc != None:
                # TODO: This seems to be breaking because of past_key_values on second forward call
                #print(f'attn_weights: {attn_weights.shape}')
                #print(f'attn_weights[:,:,:,n_ttc:]: {attn_weights[:,:,:,n_ttc:].shape}')
                #print(f'attention_mask: {attention_mask.shape}')
                #print(f'n_ttc: {n_ttc}')
                attn_weights[:,:,:,n_ttc:] = attn_weights[:,:,:,n_ttc:] + attention_mask
            else:
                attn_weights = attn_weights + attention_mask

        attn_weights = nn.functional.softmax(attn_weights, dim=-1)


        # Downcast (if necessary) back to V's dtype (if in mixed-precision) -- No-Op otherwise
        attn_weights = attn_weights.type(value.dtype)
        attn_weights = self.attn_dropout(attn_weights)

        # Mask heads if we want to
        if head_mask is not None:
            attn_weights = attn_weights * head_mask

        attn_output = torch.matmul(attn_weights, value)

        return attn_output, attn_weights

    def _upcast_and_reordered_attn(self, query, key, value, attention_mask=None, head_mask=None):
        # Use `torch.baddbmm` (a bit more efficient w/ alpha param for scaling -- from Megatron-LM)
        bsz, num_heads, q_seq_len, dk = query.size()
        _, _, k_seq_len, _ = key.size()

        # Preallocate attn_weights for `baddbmm`
        attn_weights = torch.empty(bsz * num_heads, q_seq_len, k_seq_len, dtype=torch.float32, device=query.device)

        # Compute Scale Factor
        scale_factor = 1.0
        if self.scale_attn_weights:
            scale_factor /= float(value.size(-1)) ** 0.5

        if self.scale_attn_by_inverse_layer_idx:
            scale_factor /= float(self.layer_idx + 1)

        # Upcast (turn off autocast) and reorder (Scale K by 1 / root(dk))
        with autocast(enabled=False):
            q, k = query.reshape(-1, q_seq_len, dk), key.transpose(-1, -2).reshape(-1, dk, k_seq_len)
            attn_weights = torch.baddbmm(attn_weights, q.float(), k.float(), beta=0, alpha=scale_factor)
            attn_weights = attn_weights.reshape(bsz, num_heads, q_seq_len, k_seq_len)

        if not self.is_cross_attention:
            # if only "normal" attention layer implements causal mask
            query_length, key_length = query.size(-2), key.size(-2)
            causal_mask = self.bias[:, :, key_length - query_length : key_length, :key_length]
            mask_value = torch.finfo(attn_weights.dtype).min
            # Need to be a tensor, otherwise we get error: `RuntimeError: expected scalar type float but found double`.
            # Need to be on the same device, otherwise `RuntimeError: ..., x and y to be on the same device`
            mask_value = torch.tensor(mask_value, dtype=attn_weights.dtype).to(attn_weights.device)
            attn_weights = torch.where(causal_mask, attn_weights, mask_value)

        if attention_mask is not None:
            # Apply the attention mask
            attn_weights = attn_weights + attention_mask

        attn_weights = nn.functional.softmax(attn_weights, dim=-1)

        # Downcast (if necessary) back to V's dtype (if in mixed-precision) -- No-Op if otherwise
        if attn_weights.dtype != torch.float32:
            raise RuntimeError("Error with upcasting, attn_weights does not have dtype torch.float32")
        attn_weights = attn_weights.type(value.dtype)
        attn_weights = self.attn_dropout(attn_weights)

        # Mask heads if we want to
        if head_mask is not None:
            attn_weights = attn_weights * head_mask

        attn_output = torch.matmul(attn_weights, value)

        return attn_output, attn_weights

    def forward(
        self,
        hidden_states: Optional[Tuple[torch.FloatTensor]],
        one_hot_ttc: Optional[Tuple[torch.FloatTensor]] = None,
        layer_past: Optional[Tuple[torch.Tensor]] = None,
        attention_mask: Optional[torch.FloatTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        encoder_hidden_states: Optional[torch.Tensor] = None,
        encoder_attention_mask: Optional[torch.FloatTensor] = None,
        use_cache: Optional[bool] = False,
        output_attentions: Optional[bool] = False,
    ) -> Tuple[Union[torch.Tensor, Tuple[torch.Tensor]], ...]:
        if encoder_hidden_states is not None:
            if not hasattr(self, "q_attn"):
                raise ValueError(
                    "If class is used as cross attention, the weights `q_attn` have to be defined. "
                    "Please make sure to instantiate class with `GPT2Attention(..., is_cross_attention=True)`."
                )

            query = self.q_attn(hidden_states)
            key, value = self.c_attn(encoder_hidden_states).split(self.split_size, dim=2)
            attention_mask = encoder_attention_mask
        else:
            query, key, value = self.c_attn(hidden_states).split(self.split_size, dim=2)

        query = self._split_heads(query, self.num_heads, self.head_dim)
        key = self._split_heads(key, self.num_heads, self.head_dim)
        value = self._split_heads(value, self.num_heads, self.head_dim)

        if one_hot_ttc != None:
            # NOTE: No query vectors should be created for TTC tokens, since they don't exist in the sequence
            # query: [B, heads, unique(TTC)+T, qk_dim]
            query = query[:,:,one_hot_ttc.shape[-1]:,:]

        if layer_past is not None:
            past_key, past_value = layer_past
            key = torch.cat((past_key, key), dim=-2)
            value = torch.cat((past_value, value), dim=-2)

        if use_cache is True:
            present = (key, value)
        else:
            present = None

        if self.reorder_and_upcast_attn:
            attn_output, attn_weights = self._upcast_and_reordered_attn(query, key, value, attention_mask, head_mask)
        else:
            attn_output, attn_weights = self._attn(query, key, value, one_hot_ttc, attention_mask, head_mask)

        attn_output = self._merge_heads(attn_output, self.num_heads, self.head_dim)
        attn_output = self.c_proj(attn_output)
        attn_output = self.resid_dropout(attn_output)

        outputs = (attn_output, present)
        if output_attentions:
            outputs += (attn_weights,)

        return outputs  # a, present, (attentions)


class GPT2TTCBlock(GPT2Block):
    def __init__(self, config, layer_idx=None):
        super().__init__(config, layer_idx)
        hidden_size = config.hidden_size
        inner_dim = config.n_inner if config.n_inner is not None else 4 * hidden_size

        self.ln_1 = nn.LayerNorm(hidden_size, eps=config.layer_norm_epsilon)
        self.attn = GPT2TTCAttention(config, layer_idx=layer_idx)
        self.ln_2 = nn.LayerNorm(hidden_size, eps=config.layer_norm_epsilon)

        if config.add_cross_attention:
            self.crossattention = GPT2TTCAttention(config, is_cross_attention=True, layer_idx=layer_idx)
            self.ln_cross_attn = nn.LayerNorm(hidden_size, eps=config.layer_norm_epsilon)

        self.mlp = GPT2MLP(inner_dim, config)

    def forward(
        self,
        hidden_states: Optional[Tuple[torch.FloatTensor]],
        one_hot_ttc: Optional[Tuple[torch.FloatTensor]] = None,
        layer_past: Optional[Tuple[torch.Tensor]] = None,
        attention_mask: Optional[torch.FloatTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        encoder_hidden_states: Optional[torch.Tensor] = None,
        encoder_attention_mask: Optional[torch.FloatTensor] = None,
        use_cache: Optional[bool] = False,
        output_attentions: Optional[bool] = False,
    ) -> Union[Tuple[torch.Tensor], Optional[Tuple[torch.Tensor, Tuple[torch.FloatTensor, ...]]]]:
        residual = hidden_states
        hidden_states = self.ln_1(hidden_states)
        attn_outputs = self.attn(
            hidden_states, # [B, T + unique(TTC), E]
            one_hot_ttc=one_hot_ttc,
            layer_past=layer_past,
            attention_mask=attention_mask,
            head_mask=head_mask,
            use_cache=use_cache,
            output_attentions=output_attentions,
        )
        attn_output = attn_outputs[0]  # output_attn: a, present, (attentions)
        outputs = attn_outputs[1:]
        # residual connection
        if one_hot_ttc != None:
            hidden_states = attn_output + residual[:,-attn_output.shape[1]:,:]
        else:
            hidden_states = attn_output + residual

        if encoder_hidden_states is not None:
            # add one self-attention block for cross-attention
            if not hasattr(self, "crossattention"):
                raise ValueError(
                    f"If `encoder_hidden_states` are passed, {self} has to be instantiated with "
                    "cross-attention layers by setting `config.add_cross_attention=True`"
                )
            residual = hidden_states
            hidden_states = self.ln_cross_attn(hidden_states)
            cross_attn_outputs = self.crossattention(
                hidden_states,
                attention_mask=attention_mask,
                head_mask=head_mask,
                encoder_hidden_states=encoder_hidden_states,
                encoder_attention_mask=encoder_attention_mask,
                output_attentions=output_attentions,
            )
            attn_output = cross_attn_outputs[0]
            # residual connection
            hidden_states = residual + attn_output
            outputs = outputs + cross_attn_outputs[2:]  # add cross attentions if we output attention weights

        residual = hidden_states
        hidden_states = self.ln_2(hidden_states)
        feed_forward_hidden_states = self.mlp(hidden_states)
        # residual connection
        hidden_states = residual + feed_forward_hidden_states

        if use_cache:
            outputs = (hidden_states,) + outputs
        else:
            outputs = (hidden_states,) + outputs[1:]

        return outputs  # hidden_states, present, (attentions, cross_attentions)



class GPT2TTCModel(GPT2Model):
    def __init__(self, config):
        super().__init__(config)
        self.embed_dim = config.hidden_size

        self.wte = nn.Embedding(config.vocab_size, self.embed_dim)
        self.wpe = nn.Embedding(config.max_position_embeddings, self.embed_dim)

        self.drop = nn.Dropout(config.embd_pdrop)
        self.h = nn.ModuleList([GPT2TTCBlock(config, layer_idx=i) for i in range(config.num_hidden_layers)])
        self.ln_f = nn.LayerNorm(self.embed_dim, eps=config.layer_norm_epsilon)

        # Model parallel
        self.model_parallel = False
        self.device_map = None
        self.gradient_checkpointing = False

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

    def forward(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        ttc_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None,
        attention_mask: Optional[torch.FloatTensor] = None,
        token_type_ids: Optional[torch.LongTensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        encoder_hidden_states: Optional[torch.Tensor] = None,
        encoder_attention_mask: Optional[torch.FloatTensor] = None,
        use_cache: Optional[bool] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        return_dict: Optional[bool] = None,
    ) -> Union[Tuple, BaseModelOutputWithPastAndCrossAttentions]:
        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
        )
        use_cache = use_cache if use_cache is not None else self.config.use_cache
        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()
            input_ids = input_ids.view(-1, input_shape[-1])
            batch_size = input_ids.shape[0]
        elif inputs_embeds is not None:
            input_shape = inputs_embeds.size()[:-1]
            batch_size = inputs_embeds.shape[0]
        else:
            raise ValueError("You have to specify either input_ids or inputs_embeds")

        device = input_ids.device if input_ids is not None else inputs_embeds.device

        if token_type_ids is not None:
            token_type_ids = token_type_ids.view(-1, input_shape[-1])

        if past_key_values is None:
            past_length = 0
            past_key_values = tuple([None] * len(self.h))
        else:
            past_length = past_key_values[0][0].size(-2)
        if position_ids is None:
            position_ids = torch.arange(past_length, input_shape[-1] + past_length, dtype=torch.long, device=device)
            position_ids = position_ids.unsqueeze(0)

        # GPT2Attention mask.
        if attention_mask is not None:
            if batch_size <= 0:
                raise ValueError("batch_size has to be defined and > 0")
            attention_mask = attention_mask.view(batch_size, -1)
            # We create a 3D attention mask from a 2D tensor mask.
            # Sizes are [batch_size, 1, 1, to_seq_length]
            # So we can broadcast to [batch_size, num_heads, from_seq_length, to_seq_length]
            # this attention mask is more simple than the triangular masking of causal attention
            # used in OpenAI GPT, we just need to prepare the broadcast dimension here.
            attention_mask = attention_mask[:, None, None, :]

            # Since attention_mask is 1.0 for positions we want to attend and 0.0 for
            # masked positions, this operation will create a tensor which is 0.0 for
            # positions we want to attend and the dtype's smallest value for masked positions.
            # Since we are adding it to the raw scores before the softmax, this is
            # effectively the same as removing these entirely.
            attention_mask = attention_mask.to(dtype=self.dtype)  # fp16 compatibility
            attention_mask = (1.0 - attention_mask) * torch.finfo(self.dtype).min

        # If a 2D or 3D attention mask is provided for the cross-attention
        # we need to make broadcastable to [batch_size, num_heads, seq_length, seq_length]
        if self.config.add_cross_attention and encoder_hidden_states is not None:
            encoder_batch_size, encoder_sequence_length, _ = encoder_hidden_states.size()
            encoder_hidden_shape = (encoder_batch_size, encoder_sequence_length)
            if encoder_attention_mask is None:
                encoder_attention_mask = torch.ones(encoder_hidden_shape, device=device)
            encoder_attention_mask = self.invert_attention_mask(encoder_attention_mask)
        else:
            encoder_attention_mask = None

        # 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
        # head_mask has shape n_layer x batch x n_heads x N x N
        head_mask = self.get_head_mask(head_mask, self.config.n_layer)

        if inputs_embeds is None:
            inputs_embeds = self.wte(input_ids)
        position_embeds = self.wpe(position_ids)
        hidden_states = inputs_embeds + position_embeds

        if token_type_ids is not None:
            token_type_embeds = self.wte(token_type_ids)
            hidden_states = hidden_states + token_type_embeds

        hidden_states = self.drop(hidden_states)

        output_shape = (-1,) + input_shape[1:] + (hidden_states.size(-1),)

        one_hot_ttc, unique_ttc_ids = one_hot_encode(ttc_ids)
        ttc_embeds = self.wte(unique_ttc_ids).unsqueeze(0).expand(hidden_states.shape[0], unique_ttc_ids.shape[0], hidden_states.shape[-1])

        # NOTE: Combining input hidden states and ttc hidden states here
        hidden_states = torch.cat([ttc_embeds, hidden_states], dim=1)

        if self.gradient_checkpointing and self.training:
            if use_cache:
                logger.warning_once(
                    "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
                )
                use_cache = False

        presents = () if use_cache else None
        all_self_attentions = () if output_attentions else None
        all_cross_attentions = () if output_attentions and self.config.add_cross_attention else None
        all_hidden_states = () if output_hidden_states else None
        for i, (block, layer_past) in enumerate(zip(self.h, past_key_values)):
            # Model parallel
            if self.model_parallel:
                torch.cuda.set_device(hidden_states.device)
                # Ensure layer_past is on same device as hidden_states (might not be correct)
                if layer_past is not None:
                    layer_past = tuple(past_state.to(hidden_states.device) for past_state in layer_past)
                # Ensure that attention_mask is always on the same device as hidden_states
                if attention_mask is not None:
                    attention_mask = attention_mask.to(hidden_states.device)
                if isinstance(head_mask, torch.Tensor):
                    head_mask = head_mask.to(hidden_states.device)
            if output_hidden_states:
                all_hidden_states = all_hidden_states + (hidden_states,)

            if self.gradient_checkpointing and self.training:
                outputs = self._gradient_checkpointing_func(
                    block.__call__,
                    hidden_states,
                    None,
                    attention_mask,
                    head_mask[i],
                    encoder_hidden_states,
                    encoder_attention_mask,
                    use_cache,
                    output_attentions,
                )
            else:
                outputs = block(
                    hidden_states,
                    one_hot_ttc=one_hot_ttc if i==0 else None,
                    layer_past=layer_past,
                    attention_mask=attention_mask,
                    head_mask=head_mask[i],
                    encoder_hidden_states=encoder_hidden_states,
                    encoder_attention_mask=encoder_attention_mask,
                    use_cache=use_cache,
                    output_attentions=output_attentions,
                )

            hidden_states = outputs[0]
            if use_cache is True:
                presents = presents + (outputs[1],)

            if output_attentions:
                all_self_attentions = all_self_attentions + (outputs[2 if use_cache else 1],)
                if self.config.add_cross_attention:
                    all_cross_attentions = all_cross_attentions + (outputs[3 if use_cache else 2],)

            # Model Parallel: If it's the last layer for that device, put things on the next device
            if self.model_parallel:
                for k, v in self.device_map.items():
                    if i == v[-1] and "cuda:" + str(k) != self.last_device:
                        hidden_states = hidden_states.to("cuda:" + str(k + 1))

        hidden_states = self.ln_f(hidden_states)

        hidden_states = hidden_states.view(output_shape)
        # Add last hidden state
        if output_hidden_states:
            all_hidden_states = all_hidden_states + (hidden_states,)

        if not return_dict:
            return tuple(
                v
                for v in [hidden_states, presents, all_hidden_states, all_self_attentions, all_cross_attentions]
                if v is not None
            )

        return BaseModelOutputWithPastAndCrossAttentions(
            last_hidden_state=hidden_states,
            past_key_values=presents,
            hidden_states=all_hidden_states,
            attentions=all_self_attentions,
            cross_attentions=all_cross_attentions,
        )


class GPT2TTC(GPT2LMHeadModel):

    def __init__(self, config):
        super().__init__(config)
        self.transformer = GPT2TTCModel(config)
        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)

        # Model parallel
        self.model_parallel = False
        self.device_map = None

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

    def prepare_inputs_for_generation(self, input_ids, past_key_values=None, inputs_embeds=None, **kwargs):
        token_type_ids = kwargs.get("token_type_ids", None)
        # Omit tokens covered by past_key_values
        if past_key_values:
            past_length = past_key_values[0][0].shape[2]

            # Some generation methods already pass only the last input ID
            if input_ids.shape[1] > past_length:
                remove_prefix_length = past_length
            else:
                # Default to old behavior: keep only final ID
                remove_prefix_length = input_ids.shape[1] - 1

            input_ids = input_ids[:, remove_prefix_length:]
            if token_type_ids is not None:
                token_type_ids = token_type_ids[:, -input_ids.shape[1] :]

        attention_mask = kwargs.get("attention_mask", None)
        position_ids = kwargs.get("position_ids", None)

        if attention_mask is not None and position_ids is None:
            # create position_ids on the fly for batch generation
            position_ids = attention_mask.long().cumsum(-1) - 1
            position_ids.masked_fill_(attention_mask == 0, 1)
            if past_key_values:
                position_ids = position_ids[:, -input_ids.shape[1] :]
        else:
            position_ids = None

        # if `inputs_embeds` are passed, we only want to use them in the 1st generation step
        if inputs_embeds is not None and past_key_values is None:
            model_inputs = {"inputs_embeds": inputs_embeds}
        else:
            model_inputs = {"input_ids": input_ids}

        # gather kwargs for ttc decoding
        target_length = kwargs.get("target_length", None)
        ttc_ids = kwargs.get("ttc_ids", None)
        prompt_length_offset = kwargs.get("prompt_length_offset", None)
        ttc_values_to_ids = kwargs.get("ttc_values_to_ids", None)

        model_inputs.update(
            {
                "past_key_values": past_key_values,
                "use_cache": kwargs.get("use_cache"),
                "position_ids": position_ids,
                "attention_mask": attention_mask,
                "token_type_ids": token_type_ids,
                "target_length": target_length,
                "ttc_ids": ttc_ids,
                "prompt_length_offset": prompt_length_offset,
                "ttc_values_to_ids": ttc_values_to_ids,
            }
        )

        return model_inputs


    def forward(
        self,
        input_ids: Optional[torch.LongTensor] = None,
        ttc_ids: Optional[torch.LongTensor] = None,
        past_key_values: Optional[Tuple[Tuple[torch.Tensor]]] = None,
        attention_mask: Optional[torch.FloatTensor] = None,
        token_type_ids: Optional[torch.LongTensor] = None,
        position_ids: Optional[torch.LongTensor] = None,
        head_mask: Optional[torch.FloatTensor] = None,
        inputs_embeds: Optional[torch.FloatTensor] = None,
        encoder_hidden_states: Optional[torch.Tensor] = None,
        encoder_attention_mask: Optional[torch.FloatTensor] = None,
        labels: Optional[torch.LongTensor] = None,
        use_cache: Optional[bool] = None,
        output_attentions: Optional[bool] = None,
        output_hidden_states: Optional[bool] = None,
        return_dict: Optional[bool] = None,
        target_length: Optional[int] = None,
        prompt_length_offset: Optional[int] = None,
        ttc_values_to_ids: Optional[dict[str,int]] = None,
    ) -> Union[Tuple, CausalLMOutputWithCrossAttentions]:
        r"""
        labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
            Labels for language modeling. Note that the labels **are shifted** inside the model, i.e. you can set
            `labels = input_ids` Indices are selected in `[-100, 0, ..., config.vocab_size]` All labels set to `-100`
            are ignored (masked), the loss is only computed for labels in `[0, ..., config.vocab_size]`
        """
        #print(f'== Begin Forward ==')
        ##print(f'Forward Args:\n{locals()}')
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict

        assert ((ttc_ids==None) or (target_length==None)), f'Only one of ttc_ids or target_length may be defined.'

        if prompt_length_offset == None:
            prompt_length_offset = 0

        if target_length != None:
            assert (len(input_ids.shape) == 1) or ((input_ids.shape[0] == 1) and (len(input_ids.shape) == 2)), 'batch size must be 1 to use target_length inference'
            assert ttc_values_to_ids != None, 'If target_length is specified, then the ttc_values_to_ids mapping should be provided'
            #  input_ids = [5,4,7,8,9,8,7,34,2,43]
            # prompt_ids = [5,4,7,8]
            # prompt_length_offset = 4
            # target_length = 10
            # ttc_values = [13,12,11,10,9,8,7,6,5,4] ...3,2,1,0
            ttc_tokens = list(ttc_values_to_ids.keys())
            # [8,16,32,64,128,256,512,1024,1024]
            ttc_numbers = list(reversed(range(target_length + prompt_length_offset)))
            if input_ids.shape[1] > len(ttc_numbers):
                #print(f'input_ids.shape[1]: {input_ids.shape[1]}')
                #print(f'len ttc_numbers: {len(ttc_numbers)}')
                ttc_numbers += [0]*(input_ids.shape[1]-len(ttc_numbers))
                #print(f'new len ttc_numbers: {len(ttc_numbers)}')
            ttc_numbers = ttc_numbers[:input_ids.shape[1]]
            ttc_values = map_values_to_closest_labels_nolog(ttc_numbers, ttc_tokens)
            ttc_ids = list(map(lambda x: ttc_values_to_ids[x], ttc_values))
            ttc_ids = torch.tensor(ttc_ids).long()
            ttc_ids = ttc_ids.view(input_ids.shape[0],input_ids.shape[1]).to(input_ids.device)
            #print(f'input_ids: {input_ids.shape}')
            #print(f'ttc_ids: {ttc_ids.shape}')

        transformer_outputs = self.transformer(
            input_ids,
            ttc_ids=ttc_ids,
            past_key_values=past_key_values,
            attention_mask=attention_mask,
            token_type_ids=token_type_ids,
            position_ids=position_ids,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            encoder_hidden_states=encoder_hidden_states,
            encoder_attention_mask=encoder_attention_mask,
            use_cache=use_cache,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
        )
        hidden_states = transformer_outputs[0]

        # Set device for model parallelism
        if self.model_parallel:
            torch.cuda.set_device(self.transformer.first_device)
            hidden_states = hidden_states.to(self.lm_head.weight.device)

        lm_logits = self.lm_head(hidden_states)

        loss = None
        if labels is not None:
            # move labels to correct device to enable model parallelism
            labels = labels.to(lm_logits.device)
            # Shift so that tokens < n predict n
            # NOTE: This does the offset by 1 so that labels line up with logits
            shift_logits = lm_logits[..., :-1, :].contiguous()
            shift_labels = labels[..., 1:].contiguous()
            # Flatten the tokens
            loss_fct = CrossEntropyLoss()
            loss = loss_fct(shift_logits.view(-1, shift_logits.size(-1)), shift_labels.view(-1))

        if not return_dict:
            output = (lm_logits,) + transformer_outputs[1:]
            return ((loss,) + output) if loss is not None else output

        return CausalLMOutputWithCrossAttentions(
            loss=loss,
            logits=lm_logits,
            past_key_values=transformer_outputs.past_key_values,
            hidden_states=transformer_outputs.hidden_states,
            attentions=transformer_outputs.attentions,
            cross_attentions=transformer_outputs.cross_attentions,
        )