797 lines
29 KiB
Python
797 lines
29 KiB
Python
# coding=utf-8
|
|
# Copyright 2023 Michael Feil.
|
|
#
|
|
# This code is loosely based on Huggingface text-generation-inference v0.9.3's causal_lm.py implementation.
|
|
# While it remains licensed under Apache License, Version 2.0,
|
|
# text-generation-inference itself on 7/28/2023 has changed its license.
|
|
# This code remains unaffected by this change.
|
|
#
|
|
# 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.
|
|
|
|
import torch
|
|
import numpy as np
|
|
import os
|
|
import multiprocessing
|
|
from pathlib import Path
|
|
from dataclasses import dataclass
|
|
|
|
from huggingface_hub.constants import HUGGINGFACE_HUB_CACHE
|
|
from opentelemetry import trace
|
|
from transformers import (
|
|
AutoTokenizer,
|
|
AutoConfig,
|
|
PreTrainedTokenizerBase
|
|
)
|
|
from text_generation_server.models.types import (
|
|
Batch,
|
|
PrefillTokens,
|
|
Generation,
|
|
GeneratedText,
|
|
)
|
|
from typing import Optional, Tuple, List, Type, Dict
|
|
|
|
from text_generation_server.models import Model
|
|
from text_generation_server.models.types import (
|
|
PrefillTokens,
|
|
Generation,
|
|
GeneratedText,
|
|
)
|
|
from text_generation_server.pb import generate_pb2
|
|
from text_generation_server.utils import NextTokenChooser, StoppingCriteria, Sampling
|
|
|
|
from text_generation_server.utils import Sampling
|
|
|
|
try:
|
|
import ctranslate2
|
|
except ImportError:
|
|
ctranslate2 = None
|
|
|
|
|
|
tracer = trace.get_tracer(__name__)
|
|
|
|
|
|
@dataclass
|
|
class CT2CausalLMBatch(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
|
|
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,
|
|
) -> "CT2CausalLMBatch":
|
|
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.inputs)
|
|
next_token_choosers.append(NextTokenChooser.from_pb(r.parameters, device))
|
|
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
|
|
)
|
|
|
|
tokenized_inputs = tokenizer(
|
|
inputs,
|
|
return_tensors="pt",
|
|
padding=True,
|
|
return_token_type_ids=False,
|
|
truncation=True,
|
|
max_length=max_truncation,
|
|
).to(device)
|
|
for _ in pb.requests:
|
|
input_len = tokenized_inputs["input_ids"].shape[1]
|
|
prefix_offsets.append(input_len - 5)
|
|
read_offsets.append(input_len)
|
|
|
|
input_lengths = tokenized_inputs["attention_mask"].sum(1)
|
|
max_input_length = input_lengths.max()
|
|
|
|
input_ids = tokenized_inputs["input_ids"]
|
|
# 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"]
|
|
|
|
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)
|
|
|
|
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,
|
|
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["CT2CausalLMBatch"]:
|
|
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,
|
|
]
|
|
|
|
# Ensure that past_key_values tensors can be updated in-place
|
|
if type(self.past_key_values[0]) == 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.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["CT2CausalLMBatch"]) -> "CT2CausalLMBatch":
|
|
# Used for padding
|
|
total_batch_size = 0
|
|
max_input_length = 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)
|
|
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
|
|
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),
|
|
)
|
|
|
|
# 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,
|
|
]
|
|
|
|
# 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 type(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,
|
|
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 CT2CausalLM(Model):
|
|
def __init__(
|
|
self,
|
|
model_id: str,
|
|
revision: Optional[str] = None,
|
|
quantize: Optional[str] = None,
|
|
dtype: Optional[torch.dtype] = None,
|
|
trust_remote_code: bool = False,
|
|
):
|
|
if ctranslate2 is None:
|
|
raise ValueError(
|
|
"for quantization with ct2, the installation requires the pip package ctranslate2. "
|
|
"install via `text-generation-server[ct2]` or `pip install ctranslate2` is required.",
|
|
)
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(
|
|
model_id,
|
|
revision=revision,
|
|
padding_side="left",
|
|
truncation_side="left",
|
|
trust_remote_code=trust_remote_code,
|
|
)
|
|
|
|
# Start CT2
|
|
ct2_generator_kwargs = {
|
|
"inter_threads": int(os.environ.get("TGI_CT2_INTER_THREADS", 1))
|
|
}
|
|
if torch.cuda.is_available():
|
|
self.ct2_device = "cuda"
|
|
ct2_generator_kwargs["intra_threads"] = int(
|
|
os.environ.get("TGI_CT2_INTRA_THREADS", 1)
|
|
)
|
|
else:
|
|
self.ct2_device = "cpu"
|
|
ct2_generator_kwargs["intra_threads"] = int(
|
|
os.environ.get(
|
|
"TGI_CT2_INTRA_THREADS", multiprocessing.cpu_count() // 2
|
|
)
|
|
)
|
|
|
|
if dtype == torch.float16 and self.ct2_device == "cuda":
|
|
ct2_compute_type = "float16"
|
|
elif dtype == torch.bfloat16 and self.ct2_device == "cuda":
|
|
ct2_compute_type = "bfloat16"
|
|
elif self.ct2_device == "cpu" and dtype in [torch.float16, torch.bfloat16]:
|
|
# float16 is not available on CPU
|
|
# and int16 has no stable implementation
|
|
ct2_compute_type = "float32"
|
|
else:
|
|
# default, int8 quantization.
|
|
|
|
if "cuda" in self.ct2_device:
|
|
# int8 for int8 layers, float16 for non-quantized layers
|
|
ct2_compute_type = "int8_float16"
|
|
else:
|
|
# int8 for int8 layers, float32 for non-quantized layers
|
|
ct2_compute_type = "int8"
|
|
|
|
# Start CT2 - conversion
|
|
out_dir = (
|
|
Path(HUGGINGFACE_HUB_CACHE)
|
|
/ "ct2models" / f"{model_id.replace('/','--')}--{ct2_compute_type}"
|
|
)
|
|
|
|
if not os.path.exists(out_dir / "model.bin"):
|
|
try:
|
|
converter = ctranslate2.converters.TransformersConverter(
|
|
model_id,
|
|
activation_scales=None,
|
|
load_as_float16=ct2_compute_type != "bfloat16",
|
|
revision=revision,
|
|
low_cpu_mem_usage=True,
|
|
trust_remote_code=trust_remote_code,
|
|
)
|
|
converter.convert(
|
|
output_dir=out_dir,
|
|
vmap=None,
|
|
quantization=ct2_compute_type,
|
|
force=True,
|
|
)
|
|
except Exception as ex:
|
|
raise ValueError(
|
|
f"conversion with ctranslate2 for {model_id} failed : Error {ex}"
|
|
)
|
|
if not os.path.exists(out_dir / "model.bin"):
|
|
raise ValueError(
|
|
f"no ctranslate2 model for {model_id} found after conversion in {out_dir}"
|
|
)
|
|
|
|
# Start CT2
|
|
self.ct2_model = ctranslate2.Generator(
|
|
str(out_dir),
|
|
device=self.ct2_device,
|
|
compute_type=ct2_compute_type,
|
|
**ct2_generator_kwargs,
|
|
)
|
|
|
|
class DummyModel(torch.nn.Module):
|
|
def __init__(self, *args, **kwargs) -> None:
|
|
super().__init__(*args, **kwargs)
|
|
self.config = AutoConfig.from_pretrained(
|
|
model_id, revision=revision, trust_remote_code=trust_remote_code
|
|
)
|
|
|
|
model = DummyModel()
|
|
|
|
if tokenizer.pad_token_id is None:
|
|
if model.config.pad_token_id is not None:
|
|
tokenizer.pad_token_id = model.config.pad_token_id
|
|
elif model.config.eos_token_id is not None:
|
|
tokenizer.pad_token_id = model.config.eos_token_id
|
|
elif tokenizer.eos_token_id is not None:
|
|
tokenizer.pad_token_id = tokenizer.eos_token_id
|
|
else:
|
|
tokenizer.add_special_tokens({"pad_token": "[PAD]"})
|
|
|
|
super().__init__(
|
|
model=model,
|
|
tokenizer=tokenizer,
|
|
requires_padding=True,
|
|
dtype=torch.int8 if "int8" in ct2_compute_type else torch.float16,
|
|
device=torch.device(self.ct2_device),
|
|
)
|
|
|
|
@property
|
|
def batch_type(self) -> Type[CT2CausalLMBatch]:
|
|
return CT2CausalLMBatch
|
|
|
|
def decode(self, generated_ids: List[int]) -> str:
|
|
return self.tokenizer.decode(
|
|
generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False
|
|
)
|
|
|
|
def forward_ct2(
|
|
self,
|
|
all_input_ids,
|
|
input_lengths,
|
|
) -> Tuple[torch.Tensor, List[Tuple[torch.Tensor, torch.Tensor]]]:
|
|
# CT2 forward requires a list of list of input tokens ids and lengths
|
|
ids_input = (
|
|
torch.nested.to_padded_tensor(
|
|
torch.nested.nested_tensor(all_input_ids), 1234567
|
|
)
|
|
.flatten(1)
|
|
.to(torch.int32)
|
|
)
|
|
# lengths of the padded ids_input, i.e. how often not pad=1234567 is used.
|
|
lengths = np.array(input_lengths, dtype=np.int32)
|
|
|
|
if self.ct2_device == "cuda":
|
|
lengths = torch.from_numpy(lengths).to(self.ct2_device)
|
|
elif self.ct2_device == "cpu":
|
|
ids_input = ids_input.numpy()
|
|
|
|
ids_input = ctranslate2.StorageView.from_array(ids_input)
|
|
lengths = ctranslate2.StorageView.from_array(lengths)
|
|
# now, forward through the network
|
|
logits = self.ct2_model.forward_batch(ids_input, lengths)
|
|
|
|
# continue with logits as torch tensor
|
|
if self.ct2_device == "cpu":
|
|
# logits is a float32 torch cpu tensor
|
|
logits = torch.from_numpy(np.asarray(logits))
|
|
else:
|
|
# logits is a float16 torch cuda tensor
|
|
logits = torch.as_tensor(logits, device=self.ct2_device)
|
|
return logits, None
|
|
|
|
@tracer.start_as_current_span("generate_token")
|
|
def generate_token(
|
|
self, batch: CT2CausalLMBatch
|
|
) -> Tuple[List[Generation], Optional[CT2CausalLMBatch]]:
|
|
logits, past = self.forward_ct2(batch.all_input_ids, batch.input_lengths)
|
|
|
|
# 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(
|
|
all_input_ids[-stopping_criteria.current_tokens :, 0]
|
|
)
|
|
# 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 = PrefillTokens(
|
|
prefill_token_ids, prefill_logprobs, prefill_texts
|
|
)
|
|
else:
|
|
prefill_tokens = None
|
|
|
|
generation = Generation(
|
|
request.id,
|
|
prefill_tokens,
|
|
next_token_id_squeezed,
|
|
next_token_logprob,
|
|
next_token_text,
|
|
next_token_id_squeezed.item() in self.all_special_ids,
|
|
generated_text,
|
|
)
|
|
|
|
generations.append(generation)
|
|
|
|
# Update values
|
|
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:
|
|
return generations, None
|
|
|
|
# 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
|
|
# 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
|
|
|
|
return generations, batch
|