text-generation-inference/server/text_generation_server/models/idefics_causal_lm.py

from io import BytesIO
from PIL import Image
import torch
import time

from dataclasses import dataclass
from opentelemetry import trace
from transformers import (
    AutoConfig,
    AutoProcessor,
    AutoTokenizer,
    PreTrainedTokenizerBase,
    ProcessorMixin,
)
from typing import Optional, Tuple, List, Type, Dict

from text_generation_server.models import Model
from text_generation_server.models.types import (
    Batch,
    Tokens,
    Generation,
    GeneratedText,
)
from text_generation_server.pb import generate_pb2
from text_generation_server.utils import NextTokenChooser, StoppingCriteria, Sampling
import torch.distributed
from text_generation_server.models.custom_modeling.idefics_modeling import (
    IdeficsForVisionText2Text,
)
from text_generation_server.utils import (
    initialize_torch_distributed,
    weight_files,
    Weights,
)
from text_generation_server.utils.quantization import get_loader

from text_generation_server.utils.import_utils import SYSTEM


tracer = trace.get_tracer(__name__)


@dataclass
class IdeficsCausalLMBatch(Batch):
    batch_id: int
    requests: List[generate_pb2.Request]
    requests_idx_mapping: Dict[int, int]

    # Decoder values
    input_ids: torch.Tensor
    attention_mask: torch.Tensor
    position_ids: torch.Tensor
    pixel_values: Optional[torch.Tensor]
    image_hidden_states: Optional[torch.Tensor]
    image_attention_mask: Optional[torch.Tensor]
    past_key_values: Optional[List[Tuple]]

    # All tokens
    all_input_ids: List[torch.Tensor]

    # Lengths of all generations present in the batch
    input_lengths: List[int]
    prefix_offsets: List[int]
    read_offsets: List[int]

    # Generation helpers
    next_token_choosers: List[NextTokenChooser]
    stopping_criterias: List[StoppingCriteria]

    # Metadata used for padding
    max_input_length: int
    padding_right_offset: int

    # Maximum number of tokens this batch will grow to
    max_tokens: int

    # Past metadata
    keys_head_dim_last: bool = True

    def to_pb(self) -> generate_pb2.CachedBatch:
        return generate_pb2.CachedBatch(
            id=self.batch_id,
            request_ids=[r.id for r in self.requests],
            size=len(self),
            max_tokens=self.max_tokens,
        )

    @classmethod
    def from_pb(
        cls,
        pb: generate_pb2.Batch,
        tokenizer: PreTrainedTokenizerBase,
        dtype: torch.dtype,
        device: torch.device,
    ) -> "IdeficsCausalLMBatch":
        raise NotImplementedError

    @classmethod
    def from_pb_processor(
        cls,
        pb: generate_pb2.Batch,
        tokenizer: PreTrainedTokenizerBase,
        processor: ProcessorMixin,  # Hack
        config,
        dtype: torch.dtype,
        device: torch.device,
    ) -> "IdeficsCausalLMBatch":
        inputs = []
        next_token_choosers = []
        stopping_criterias = []
        prefix_offsets = []
        read_offsets = []
        requests_idx_mapping = {}

        # Parse batch
        max_truncation = 0
        padding_right_offset = 0
        max_decode_tokens = 0
        for i, r in enumerate(pb.requests):
            requests_idx_mapping[r.id] = i
            inputs.append(r.input_chunks.chunks)
            next_token_choosers.append(
                NextTokenChooser.from_pb(r.parameters, device, tokenizer)
            )
            stopping_criteria = StoppingCriteria.from_pb(
                r.stopping_parameters, tokenizer
            )
            stopping_criterias.append(stopping_criteria)
            max_truncation = max(max_truncation, r.truncate)
            max_decode_tokens += stopping_criteria.max_new_tokens
            padding_right_offset = max(
                padding_right_offset, stopping_criteria.max_new_tokens
            )

        # TODO Check impact on idefics
        prompts = []
        for inp in inputs:
            # Each input is encoded into a list, where each element of this input list is either a string or a URL
            prompt = []
            for chunk in inp:
                chunk_type = chunk.WhichOneof("chunk")
                if chunk_type == "text":
                    prompt.append(chunk.text)
                elif chunk_type == "image":
                    image = Image.open(BytesIO(chunk.image.data))
                    prompt.append(image)
                else:
                    raise RuntimeError(f"Invalid chunk type {chunk_type}")
            prompts.append(prompt)

        # The processor replaces the call to tokenizer, and
        # a/ takes care of fetching images from the URL
        # b/ generate the correct input_ids, attention_mask, pixel_values, image_attention_mask to feed to the model
        tokenized_inputs = processor(
            prompts,
            return_tensors="pt",
            padding=True,
            truncation=True,
            max_length=max_truncation,
            # TODO Check impact on idefics
            # add_end_of_utterance_token=False,  # Already taken care of inside the prompts, so bypassing the processor's handling of this token
        ).to(device)
        for _ in pb.requests:
            input_len = tokenized_inputs["input_ids"].shape[1]
            prefix_offsets.append(
                input_len - 5
            )  # To decode without potential fallbacks errors
            read_offsets.append(
                input_len
            )  # To decode without potential fallbacks errors

        input_lengths = tokenized_inputs["attention_mask"].sum(1)
        max_input_length = input_lengths.max()

        input_ids = tokenized_inputs["input_ids"]
        pixel_values = tokenized_inputs.get("pixel_values", None)
        image_hidden_states = None
        # Allocate maximum attention_mask
        attention_mask = input_ids.new_zeros(
            (pb.size, max_input_length + padding_right_offset)
        )
        # Copy tokenizer attention_mask into fully allocated attention_mask
        attention_mask[:, :max_input_length] = tokenized_inputs["attention_mask"]
        # Do the same for image_attention_mask
        if pixel_values is None:
            image_attention_mask = None
        else:
            image_attention_mask = input_ids.new_zeros(
                (
                    pb.size,
                    max_input_length + padding_right_offset,
                    pixel_values.size(1),
                )
            )
            image_attention_mask[:, :max_input_length, :] = tokenized_inputs[
                "image_attention_mask"
            ]

        position_ids = tokenized_inputs["attention_mask"].long().cumsum(-1) - 1
        position_ids.masked_fill_(tokenized_inputs["attention_mask"] == 0, 1)
        all_input_ids = tokenized_inputs["input_ids"].T.split(
            1, dim=1
        )  # It's input_ids but splitted into a tuple of tensors where each tensor is (seq_len, 1) size. It is then transformed into a list

        max_tokens = len(inputs) * (max_input_length + max_decode_tokens)

        return cls(
            batch_id=pb.id,
            requests=pb.requests,
            requests_idx_mapping=requests_idx_mapping,
            input_ids=input_ids,
            attention_mask=attention_mask,
            position_ids=position_ids,
            pixel_values=pixel_values,
            image_hidden_states=image_hidden_states,
            image_attention_mask=image_attention_mask,
            past_key_values=None,
            all_input_ids=list(all_input_ids),
            input_lengths=input_lengths.tolist(),
            prefix_offsets=prefix_offsets,
            read_offsets=read_offsets,
            next_token_choosers=next_token_choosers,
            stopping_criterias=stopping_criterias,
            max_input_length=max_input_length.item(),
            padding_right_offset=padding_right_offset,
            max_tokens=max_tokens,
        )

    @tracer.start_as_current_span("filter")
    def filter(self, request_ids: List[int]) -> Optional["IdeficsCausalLMBatch"]:
        # It deletes requests from the batch. For instance when client lost connection
        if len(request_ids) == 0:
            raise ValueError("Batch must have at least one request")
        if len(request_ids) == len(self):
            return self

        keep_indices = []

        # New values after filtering
        requests_idx_mapping = {}
        requests = []
        input_lengths = []
        prefix_offsets = []
        read_offsets = []
        all_input_ids = []
        max_input_length = 0

        next_token_choosers = []
        stopping_criterias = []

        total_remaining_decode_tokens = 0
        new_padding_right_offset = 0

        for i, request_id in enumerate(request_ids):
            idx = self.requests_idx_mapping[request_id]
            requests_idx_mapping[request_id] = i
            keep_indices.append(idx)

            requests.append(self.requests[idx])
            prefix_offsets.append(self.prefix_offsets[idx])
            read_offsets.append(self.read_offsets[idx])
            all_input_ids.append(self.all_input_ids[idx])

            request_input_length = self.input_lengths[idx]
            input_lengths.append(request_input_length)
            max_input_length = max(max_input_length, request_input_length)

            next_token_choosers.append(self.next_token_choosers[idx])
            stopping_criteria = self.stopping_criterias[idx]
            stopping_criterias.append(stopping_criteria)
            remaining_decode_tokens = (
                stopping_criteria.max_new_tokens - stopping_criteria.current_tokens
            )
            total_remaining_decode_tokens += remaining_decode_tokens
            new_padding_right_offset = max(
                new_padding_right_offset, remaining_decode_tokens
            )

        # Apply indices to input_ids, attention mask, past key values and other items that need to be cached
        input_ids = self.input_ids[keep_indices]
        position_ids = self.position_ids[keep_indices]
        self.attention_mask = self.attention_mask[
            keep_indices,
            -(self.padding_right_offset + max_input_length) : (
                self.attention_mask.shape[1] - self.padding_right_offset
            )
            + new_padding_right_offset,
        ]
        # Do the same for pixel_values and image_attention_mask
        pixel_values = self.pixel_values[keep_indices]
        self.image_attention_mask = self.image_attention_mask[
            keep_indices,
            -(self.padding_right_offset + max_input_length) : (
                self.image_attention_mask.shape[1] - self.padding_right_offset
            )
            + new_padding_right_offset,
            :,
        ]
        if self.image_hidden_states is None:
            image_hidden_states = None
        else:
            image_hidden_states = self.image_hidden_states[keep_indices]

        # Ensure that past_key_values tensors can be updated in-place
        if type(self.past_key_values[0]) is tuple:
            self.past_key_values = [list(layer) for layer in self.past_key_values]

        # Update tensors in-place to allow incremental garbage collection
        past_kv_length = max_input_length - 1
        for layer in self.past_key_values:
            past_keys, past_values = layer
            if len(past_keys.shape) == 3:
                # Force past to be of dim [self_size, num_heads, ...] for easy indexing
                past_keys = past_keys.view(len(self), -1, *past_keys.shape[-2:])
                past_values = past_values.view(len(self), -1, *past_values.shape[-2:])
            if self.keys_head_dim_last:
                layer[0] = past_keys[keep_indices, :, -past_kv_length:, :]
            else:
                layer[0] = past_keys[keep_indices, :, :, -past_kv_length:]
            del past_keys
            layer[1] = past_values[keep_indices, :, -past_kv_length:, :]
            del past_values

        max_tokens = len(request_ids) * max_input_length + total_remaining_decode_tokens

        self.requests = requests
        self.requests_idx_mapping = requests_idx_mapping
        self.input_ids = input_ids
        self.pixel_values = pixel_values
        self.image_hidden_states = image_hidden_states
        self.position_ids = position_ids
        self.all_input_ids = all_input_ids
        self.input_lengths = input_lengths
        self.prefix_offsets = prefix_offsets
        self.read_offsets = read_offsets
        self.next_token_choosers = next_token_choosers
        self.stopping_criterias = stopping_criterias
        self.max_input_length = max_input_length
        self.padding_right_offset = new_padding_right_offset
        self.max_tokens = max_tokens

        return self

    @classmethod
    @tracer.start_as_current_span("concatenate")
    def concatenate(
        cls, batches: List["IdeficsCausalLMBatch"]
    ) -> "IdeficsCausalLMBatch":
        # It adds new requests to the batch
        # Used for padding
        total_batch_size = 0
        max_input_length = 0
        max_num_images = 0
        padding_right_offset = 0
        for batch in batches:
            total_batch_size += len(batch)
            max_input_length = max(max_input_length, batch.max_input_length)
            max_num_images = max(max_num_images, batch.pixel_values.size(1))
            padding_right_offset = max(padding_right_offset, batch.padding_right_offset)

        # Batch attributes
        requests = []
        requests_idx_mapping = {}
        input_lengths = []
        prefix_offsets = []
        read_offsets = []
        all_input_ids = []
        next_token_choosers = []
        stopping_criterias = []
        max_tokens = 0

        # Batch tensors
        input_ids = None
        attention_mask = None
        position_ids = None
        pixel_values = None
        image_hidden_states = None
        image_attention_mask = None
        past_key_values = []

        # Used for slicing correctly inside the tensors
        # Equivalent to a cumsum on batch sizes
        start_index = 0
        for i, batch in enumerate(batches):
            requests.extend(batch.requests)
            input_lengths.extend(batch.input_lengths)
            prefix_offsets.extend(batch.prefix_offsets)
            read_offsets.extend(batch.read_offsets)
            all_input_ids.extend(batch.all_input_ids)
            next_token_choosers.extend(batch.next_token_choosers)
            stopping_criterias.extend(batch.stopping_criterias)

            if i == 0:
                requests_idx_mapping = batch.requests_idx_mapping
            else:
                # We need to offset the mapping for each batch by the cumulative batch size
                for k, v in batch.requests_idx_mapping.items():
                    requests_idx_mapping[k] = v + start_index

            # Slicing end index for this batch
            end_index = start_index + len(batch)

            # We only concatenate batches that did at least one step
            if batch.past_key_values is None:
                raise ValueError("only concatenate prefilled batches")

            # Create empty tensor
            # input_ids is always of shape [batch_size, 1]
            # We do not need to pad it
            if input_ids is None:
                input_ids = batch.input_ids.new_empty((total_batch_size, 1))
            # Copy to correct indices
            input_ids[start_index:end_index] = batch.input_ids

            # Create padded tensor
            if attention_mask is None:
                attention_mask = batch.attention_mask.new_zeros(
                    (total_batch_size, max_input_length + padding_right_offset),
                )

            curr_batch_max_num_images = batch.pixel_values.size(1)
            if pixel_values is None:
                pixel_values = batch.pixel_values.new_zeros(
                    (total_batch_size, max_num_images, 3, 224, 224)
                )
            pixel_values[start_index:end_index, :curr_batch_max_num_images] = (
                batch.pixel_values
            )

            if image_attention_mask is None:
                image_attention_mask = batch.image_attention_mask.new_zeros(
                    (
                        total_batch_size,
                        max_input_length + padding_right_offset,
                        max_num_images,
                    )
                )

            # We need to slice the attention mask to remove padding from previous steps
            # and to remove unused allocated space
            left_offset = max_input_length - batch.max_input_length
            batch_left_offset = (
                batch.attention_mask.shape[1]
                - batch.max_input_length
                - batch.padding_right_offset
            )
            attention_mask[
                start_index:end_index,
                left_offset:-padding_right_offset,
            ] = batch.attention_mask[
                :,
                batch_left_offset : -batch.padding_right_offset,
            ]
            image_attention_mask[
                start_index:end_index,
                left_offset:-padding_right_offset,
                :curr_batch_max_num_images,
            ] = batch.image_attention_mask[
                :, batch_left_offset : -batch.padding_right_offset, :
            ]

            # Create empty tensor
            # position_ids is always of shape [batch_size, 1]
            if position_ids is None:
                position_ids = batch.position_ids.new_empty((total_batch_size, 1))
            position_ids[start_index:end_index] = batch.position_ids

            # Shenanigans to get dimensions because BLOOM outputs a past with a different shape
            # BLOOM Keys:   [batch_size * num_heads, head_dim, seq_length]
            # BLOOM Values: [batch_size * num_heads, seq_length, head_dim]
            # And ensure that we can update tensors in-place
            if isinstance(batch.past_key_values[0], tuple):
                batch.past_key_values = [
                    [t.view(len(batch), -1, *t.shape[-2:]) for t in layer]
                    for layer in batch.past_key_values
                ]
            elif len(batch.past_key_values[0][0].shape) == 3:
                for layer in batch.past_key_values:
                    for k, t in enumerate(layer):
                        layer[k] = t.view(len(batch), -1, *t.shape[-2:])

            # Add eventual padding tokens that were added while concatenating
            max_tokens += batch.max_tokens + (
                max_input_length - batch.max_input_length
            ) * len(batch)

            start_index = end_index

        first_past_kvs = batches[0].past_key_values
        _, num_heads, padded_sequence_length, head_dim = first_past_kvs[0][1].shape

        padded_past_values_shape = (
            total_batch_size,
            num_heads,
            max_input_length - 1,
            head_dim,
        )

        if batches[0].keys_head_dim_last:
            padded_past_keys_shape = padded_past_values_shape
        else:
            # seq_length is last for BLOOM
            padded_past_keys_shape = (
                total_batch_size,
                num_heads,
                head_dim,
                max_input_length - 1,
            )

        # Iterate over attention layers
        # Concatenate past key values layer by layer to allow incremental garbage collection
        for j in range(len(first_past_kvs)):
            padded_past_keys = first_past_kvs[j][0].new_zeros(padded_past_keys_shape)
            start_index = 0
            for batch in batches:
                past_keys = batch.past_key_values[j][0]
                # Clear reference to the original tensor
                batch.past_key_values[j][0] = None

                # Slicing end index for this batch
                end_index = start_index + len(batch)
                # We slice the keys to remove the padding from previous batches
                past_seq_len = batch.max_input_length - 1
                if batch.keys_head_dim_last:
                    padded_past_keys[start_index:end_index, :, -past_seq_len:, :] = (
                        past_keys[:, :, -past_seq_len:, :]
                    )
                else:
                    # BLOOM case
                    padded_past_keys[start_index:end_index, :, :, -past_seq_len:] = (
                        past_keys[:, :, :, -past_seq_len:]
                    )
                del past_keys

                start_index = end_index

            padded_past_values = first_past_kvs[j][1].new_zeros(
                padded_past_values_shape
            )
            start_index = 0
            for batch in batches:
                past_values = batch.past_key_values[j][1]
                # Clear reference to the original tensor
                batch.past_key_values[j][1] = None

                # Slicing end index for this batch
                end_index = start_index + len(batch)
                # We slice the past values to remove the padding from previous batches
                past_seq_len = batch.max_input_length - 1
                padded_past_values[start_index:end_index, :, -past_seq_len:, :] = (
                    past_values[:, :, -past_seq_len:, :]
                )
                del past_values

                # Update values
                start_index = end_index

            past_key_values.append([padded_past_keys, padded_past_values])

        return cls(
            batch_id=batches[0].batch_id,
            requests=requests,
            requests_idx_mapping=requests_idx_mapping,
            input_ids=input_ids,
            attention_mask=attention_mask,
            position_ids=position_ids,
            pixel_values=pixel_values,
            image_hidden_states=image_hidden_states,
            image_attention_mask=image_attention_mask,
            past_key_values=past_key_values,
            all_input_ids=all_input_ids,
            input_lengths=input_lengths,
            prefix_offsets=prefix_offsets,
            read_offsets=read_offsets,
            next_token_choosers=next_token_choosers,
            stopping_criterias=stopping_criterias,
            max_input_length=max_input_length,
            padding_right_offset=padding_right_offset,
            keys_head_dim_last=batches[0].keys_head_dim_last,
            max_tokens=max_tokens,
        )

    def __len__(self):
        return len(self.requests)


class IdeficsCausalLM(Model):
    def __init__(
        self,
        model_id: str,
        revision: Optional[str] = None,
        quantize: Optional[str] = None,
        speculator: Optional[str] = None,
        dtype: Optional[torch.dtype] = None,
        trust_remote_code: bool = False,
    ):
        self.quantize = quantize
        self.process_group, rank, world_size = initialize_torch_distributed()
        if torch.cuda.is_available():
            device = torch.device(f"cuda:{rank}")
            # 9b seems to work correctly enough in float16, but 80b seems
            # to be really saturating for f16.
            dtype = torch.float16 if dtype is None else dtype
        elif SYSTEM == "ipex":
            if hasattr(torch, "xpu") and torch.xpu.is_available():
                device = torch.device(f"xpu:{rank}")
                dtype = torch.float16 if dtype is None else dtype
            else:
                device = torch.device("cpu")
                # Float16 doesn't exist on target.
                dtype = torch.bfloat16 if dtype is None else dtype
        else:
            device = torch.device("cpu")
            dtype = torch.float32 if dtype is None else dtype
        self.device, self.dtype = device, dtype

        config = AutoConfig.from_pretrained(
            model_id,
            revision=revision,
            trust_remote_code=trust_remote_code,
        )
        config.quantize = quantize
        config.speculator = speculator
        config.vision_config.quantize = quantize

        tokenizer = AutoTokenizer.from_pretrained(
            model_id,
            revision=revision,
            padding_side="left",
            truncation_side="left",
            trust_remote_code=trust_remote_code,
        )
        self.processor = AutoProcessor.from_pretrained(
            model_id,
            revision=revision,
            padding_side="left",
            truncation_side="left",
            trust_remote_code=trust_remote_code,
        )

        weights_loader = get_loader(
            quantize=quantize, model_id=model_id, revision=revision
        )
        torch.distributed.barrier(group=self.process_group)
        filenames = weight_files(model_id, revision=revision, extension=".safetensors")
        weights = Weights(
            filenames,
            device=device,
            dtype=dtype,
            process_group=self.process_group,
            weights_loader=weights_loader,
        )

        model = IdeficsForVisionText2Text(config, weights)

        self.config = config

        torch.distributed.barrier(group=self.process_group)
        super().__init__(
            model_id=model_id,
            model=model,
            tokenizer=tokenizer,
            requires_padding=True,
            dtype=dtype,
            device=device,
            rank=rank,
            world_size=world_size,
        )

    @property
    def batch_type(self) -> Type[IdeficsCausalLMBatch]:
        return IdeficsCausalLMBatch

    def forward(
        self,
        input_ids,
        attention_mask,
        position_ids,
        pixel_values,
        image_hidden_states,
        image_attention_mask,
        past_key_values: Optional = None,
    ) -> Tuple[torch.Tensor, List[Tuple[torch.Tensor, torch.Tensor]]]:
        # Model Forward
        kwargs = {
            "input_ids": input_ids,
            "attention_mask": attention_mask,
            "pixel_values": pixel_values,
            "image_hidden_states": image_hidden_states,
            "image_attention_mask": image_attention_mask,
            "past_key_values": past_key_values,
            "use_cache": True,
            "return_dict": True,
        }
        if self.has_position_ids:
            kwargs["position_ids"] = position_ids

        outputs, speculative_logits = self.model.forward(**kwargs)
        return (
            outputs.logits,
            speculative_logits,
            outputs.past_key_values,
            outputs.image_hidden_states,
        )

    @tracer.start_as_current_span("generate_token")
    def generate_token(
        self, batch: IdeficsCausalLMBatch
    ) -> Tuple[List[Generation], Optional[IdeficsCausalLMBatch], Tuple[int, int]]:
        start = time.time_ns()
        # slice the attention mask to the correct shape
        attention_mask = batch.attention_mask[:, : -batch.padding_right_offset]
        if batch.image_attention_mask is None:
            image_attention_mask = None
        else:
            if batch.input_ids.size(1) == 1:
                # THIS is a hack: when calling idefics.generate, the first time, we need the whole image_attention_mask (size bs x max_seq_len x max_num_images),
                # but the subsequent times, we only need the last attention mask along the `max_seq_len` dimension
                # this is due to the nature IDEFICS: it's an encoder decoder, and so when decoding, only the currently generated
                # token need to attend to the encoder hidden states (i.e. the vision encoder)
                # Also see seq2seq_lm.Seq2SeqLM.generate_token which has roughly the same logic
                image_attention_mask = batch.image_attention_mask[
                    :, -(batch.padding_right_offset + 1)
                ].unsqueeze(1)
            else:
                image_attention_mask = batch.image_attention_mask[
                    :, : -batch.padding_right_offset
                ]

        logits, speculative_logits, past, image_hidden_states = self.forward(
            input_ids=batch.input_ids,
            attention_mask=attention_mask,
            position_ids=batch.position_ids,
            pixel_values=batch.pixel_values,
            image_hidden_states=batch.image_hidden_states,
            image_attention_mask=image_attention_mask,
            past_key_values=batch.past_key_values,
        )
        # Hardcoded remove image tokens
        logits[:, 32000:32001] = torch.finfo(logits.dtype).min

        start_decode = time.time_ns()

        # Results
        generations: List[Generation] = []
        stopped = True

        # Zipped iterator
        iterator = zip(
            batch.requests,
            batch.input_lengths,
            batch.prefix_offsets,
            batch.read_offsets,
            logits,
            batch.next_token_choosers,
            batch.stopping_criterias,
            batch.all_input_ids,
        )

        # For each member of the batch
        for i, (
            request,
            input_length,
            prefix_offset,
            read_offset,
            logits,
            next_token_chooser,
            stopping_criteria,
            all_input_ids,
        ) in enumerate(iterator):
            # Select next token
            next_token_id, logprobs = next_token_chooser(
                all_input_ids.view(1, -1), logits[-1:, :]
            )

            # Append next token to all tokens
            all_input_ids = torch.cat([all_input_ids, next_token_id])
            new_input_length = input_length + 1

            # Generated token
            next_token_logprob = logprobs[-1, next_token_id]
            next_token_id_squeezed = next_token_id.squeeze()
            next_token_text, prefix_offset, read_offset = self.decode_token(
                all_input_ids[:, 0], prefix_offset, read_offset
            )

            # Evaluate stopping criteria
            stop, reason = stopping_criteria(
                next_token_id_squeezed,
                next_token_text,
            )

            if not stop:
                stopped = False

            # Shard generations
            # All generations will be appended in the rust sharded client
            if i % self.world_size == self.rank:
                if stop:
                    # Decode generated tokens
                    output_text, _, _ = self.decode_token(
                        all_input_ids[:, 0],
                        prefix_offset=len(all_input_ids)
                        - stopping_criteria.current_tokens
                        - 1,
                        read_offset=len(all_input_ids)
                        - stopping_criteria.current_tokens,
                        skip_special_tokens=True,
                    )
                    # Get seed
                    if isinstance(next_token_chooser.choice, Sampling):
                        seed = next_token_chooser.choice.seed
                    else:
                        seed = None

                    generated_text = GeneratedText(
                        output_text, stopping_criteria.current_tokens, reason, seed
                    )
                else:
                    generated_text = None

                # Prefill
                if stopping_criteria.current_tokens == 1 and request.prefill_logprobs:
                    # Remove generated token to only have prefill and add nan for first prompt token
                    prefill_logprobs = [float("nan")] + torch.log_softmax(
                        logits, -1
                    ).gather(1, all_input_ids[1:]).squeeze(1)[
                        -new_input_length:-1
                    ].tolist()
                    prefill_token_ids = all_input_ids[-new_input_length:-1]
                    prefill_texts = self.tokenizer.batch_decode(
                        prefill_token_ids,
                        clean_up_tokenization_spaces=False,
                        skip_special_tokens=False,
                    )
                    prefill_tokens = Tokens(
                        prefill_token_ids,
                        prefill_logprobs,
                        prefill_texts,
                        is_special=[],
                    )
                else:
                    prefill_tokens = None

                top_tokens = None

                generation = Generation(
                    request.id,
                    prefill_tokens,
                    Tokens(
                        [next_token_id_squeezed],
                        [next_token_logprob],
                        [next_token_text],
                        [next_token_id_squeezed.item() in self.all_special_ids],
                    ),
                    generated_text,
                    top_tokens,
                )

                generations.append(generation)

            # Update values
            batch.next_token_choosers[i] = batch.next_token_choosers[i].advance_grammar(
                next_token_id_squeezed.item()
            )
            batch.input_ids[i, 0] = next_token_id
            batch.all_input_ids[i] = all_input_ids
            batch.input_lengths[i] = new_input_length
            batch.prefix_offsets[i] = prefix_offset
            batch.read_offsets[i] = read_offset
            batch.max_input_length = max(batch.max_input_length, new_input_length)

        # We finished all generations in the batch; there is no next batch
        if stopped:
            forward_ns = start_decode - start
            decode_ns = time.time_ns() - start_decode
            return generations, None, (forward_ns, decode_ns)

        # Slice unused values from prefill
        batch.input_ids = batch.input_ids[:, :1]

        # Update attention_mask as we added a new token to input_ids
        batch.attention_mask[:, -batch.padding_right_offset] = 1
        batch.image_attention_mask[:, -batch.padding_right_offset, :] = (
            batch.image_attention_mask[:, -(batch.padding_right_offset + 1), :]
        )
        # Decrease right offset
        batch.padding_right_offset -= 1

        # Update position_ids
        batch.position_ids = batch.position_ids[:, -1:] + 1

        # Update past key values
        batch.past_key_values = past
        batch.image_hidden_states = image_hidden_states

        forward_ns = start_decode - start
        decode_ns = time.time_ns() - start_decode
        return generations, batch, (forward_ns, decode_ns)