text-generation-webui/modules/exllamav2_hf.py

182 lines
6.9 KiB
Python
Raw Normal View History

import os
2023-10-21 02:53:24 -04:00
import traceback
from pathlib import Path
from typing import Any, Dict, Optional, Union
import torch
2023-11-02 14:23:04 -04:00
from exllamav2 import (
ExLlamaV2,
ExLlamaV2Cache,
ExLlamaV2Cache_8bit,
ExLlamaV2Cache_Q4,
2023-11-02 14:23:04 -04:00
ExLlamaV2Config
)
from torch.nn import CrossEntropyLoss
from transformers import GenerationConfig, PretrainedConfig, PreTrainedModel
from transformers.modeling_outputs import CausalLMOutputWithPast
from modules import shared
from modules.logging_colors import logger
try:
import flash_attn
except ModuleNotFoundError:
logger.warning(
'You are running ExLlamaV2 without flash-attention. This will cause the VRAM usage '
'to be a lot higher than it could be.\n'
'Try installing flash-attention following the instructions here: '
'https://github.com/Dao-AILab/flash-attention#installation-and-features'
)
pass
2023-10-21 02:53:24 -04:00
except Exception:
logger.warning('Failed to load flash-attention due to the following error:\n')
traceback.print_exc()
class Exllamav2HF(PreTrainedModel):
def __init__(self, config: ExLlamaV2Config):
super().__init__(PretrainedConfig())
self.ex_config = config
self.loras = None
self.generation_config = GenerationConfig()
self.ex_model = ExLlamaV2(config)
if not shared.args.autosplit:
split = None
if shared.args.gpu_split:
split = [float(alloc) for alloc in shared.args.gpu_split.split(",")]
self.ex_model.load(split)
if shared.args.cache_8bit:
self.ex_cache = ExLlamaV2Cache_8bit(self.ex_model, lazy=shared.args.autosplit)
elif shared.args.cache_4bit:
self.ex_cache = ExLlamaV2Cache_Q4(self.ex_model, lazy=shared.args.autosplit)
else:
self.ex_cache = ExLlamaV2Cache(self.ex_model, lazy=shared.args.autosplit)
if shared.args.autosplit:
self.ex_model.load_autosplit(self.ex_cache)
2023-11-02 14:23:04 -04:00
self.past_seq = None
if shared.args.cfg_cache:
2023-11-02 14:23:04 -04:00
if shared.args.cache_8bit:
self.ex_cache_negative = ExLlamaV2Cache_8bit(self.ex_model)
elif shared.args.cache_4bit:
self.ex_cache_negative = ExLlamaV2Cache_Q4(self.ex_model)
2023-11-02 14:23:04 -04:00
else:
self.ex_cache_negative = ExLlamaV2Cache(self.ex_model)
self.past_seq_negative = None
def _validate_model_class(self):
pass
def _validate_model_kwargs(self, model_kwargs: Dict[str, Any]):
pass
def prepare_inputs_for_generation(self, input_ids, **kwargs):
return {'input_ids': input_ids, **kwargs}
@property
def device(self) -> torch.device:
return torch.device(0)
def __call__(self, *args, **kwargs):
use_cache = kwargs.get('use_cache', True)
labels = kwargs.get('labels', None)
past_key_values = kwargs.get('past_key_values', None)
if len(args) > 0:
if not shared.args.cfg_cache:
logger.error("Please enable the cfg-cache option to use CFG with ExLlamav2_HF.")
return
input_ids = args[0]
is_negative = True
past_seq = self.past_seq_negative
ex_cache = self.ex_cache_negative
else:
input_ids = kwargs['input_ids']
is_negative = False
past_seq = self.past_seq
ex_cache = self.ex_cache
seq = input_ids[0].tolist()
if is_negative and past_key_values is not None:
seq = past_key_values + seq
seq_tensor = torch.tensor(seq)
2023-09-19 16:12:19 -04:00
reset = True
# Make the forward call
if labels is None:
2023-09-19 16:12:19 -04:00
if past_seq is not None:
min_length = min(past_seq.shape[0], seq_tensor.shape[0])
indices = torch.nonzero(~torch.eq(past_seq[:min_length], seq_tensor[:min_length]))
if len(indices) > 0:
longest_prefix = indices[0].item()
else:
longest_prefix = min_length
if longest_prefix > 0:
reset = False
ex_cache.current_seq_len = longest_prefix
if len(seq_tensor) - longest_prefix > 1:
self.ex_model.forward(seq_tensor[longest_prefix:-1].view(1, -1), ex_cache, preprocess_only=True, loras=self.loras)
elif len(seq_tensor) == longest_prefix:
# Very tricky: if the prefix we are reusing *is* the input_ids, then we have to back up the cache pointer by one,
# because we feed input_ids[-1] to forward() below, but that last token is already in the cache!
ex_cache.current_seq_len -= 1
2023-09-19 16:12:19 -04:00
if reset:
ex_cache.current_seq_len = 0
2023-09-19 16:12:19 -04:00
if len(seq_tensor) > 1:
self.ex_model.forward(seq_tensor[:-1].view(1, -1), ex_cache, preprocess_only=True, loras=self.loras)
logits = self.ex_model.forward(seq_tensor[-1:].view(1, -1), ex_cache, loras=self.loras).to(input_ids.device).float()
else:
ex_cache.current_seq_len = 0
logits = self.ex_model.forward(seq_tensor.view(1, -1), ex_cache, last_id_only=False, loras=self.loras).float()
if is_negative:
self.past_seq_negative = seq_tensor
else:
self.past_seq = seq_tensor
loss = None
if labels is not None:
# Shift so that tokens < n predict n
shift_logits = logits[..., :-1, :].contiguous()
shift_labels = labels[..., 1:].contiguous()
# Flatten the tokens
loss_fct = CrossEntropyLoss()
shift_logits = shift_logits.view(-1, logits.shape[-1])
shift_labels = shift_labels.view(-1)
# Enable model parallelism
shift_labels = shift_labels.to(shift_logits.device)
loss = loss_fct(shift_logits, shift_labels)
return CausalLMOutputWithPast(logits=logits, past_key_values=seq if use_cache else None, loss=loss)
@classmethod
def from_pretrained(cls, pretrained_model_name_or_path: Optional[Union[str, os.PathLike]], *model_args, **kwargs):
assert len(model_args) == 0 and len(kwargs) == 0, "extra args is currently not supported"
if isinstance(pretrained_model_name_or_path, str):
pretrained_model_name_or_path = Path(pretrained_model_name_or_path)
pretrained_model_name_or_path = Path(f'{shared.args.model_dir}') / Path(pretrained_model_name_or_path)
config = ExLlamaV2Config()
2023-09-12 20:42:22 -04:00
config.model_dir = str(pretrained_model_name_or_path)
config.prepare()
config.max_seq_len = shared.args.max_seq_len
config.scale_pos_emb = shared.args.compress_pos_emb
config.scale_alpha_value = shared.args.alpha_value
2023-11-02 11:19:42 -04:00
config.no_flash_attn = shared.args.no_flash_attn
config.num_experts_per_token = int(shared.args.num_experts_per_token)
return Exllamav2HF(config)