DeciLM-7B / modeling_decilm.py
tomer-deci's picture
added support for text-generation pipeline
83e1abe
# coding=utf-8
# Copyright and license in the repo.
""" PyTorch DeciLM model."""
from .version_check import check_transformers_version
check_transformers_version()
from typing import List, Optional, Tuple, Union
import torch
import torch.nn.functional as F
import torch.utils.checkpoint
from torch import nn
from transformers.models.auto.modeling_auto import MODEL_FOR_CAUSAL_LM_MAPPING_NAMES
from transformers.utils import add_start_docstrings, add_start_docstrings_to_model_forward, logging
from .configuration_decilm import DeciLMConfig
from .transformers_v4_35_2__modeling_attn_mask_utils import _prepare_4d_causal_attention_mask
from .transformers_v4_35_2__modeling_llama import LlamaMLP, LlamaRMSNorm, LlamaAttention, apply_rotary_pos_emb, \
repeat_kv, LlamaPreTrainedModel, LLAMA_START_DOCSTRING, LlamaDecoderLayer, LlamaForCausalLM, LlamaModel, \
BaseModelOutputWithPast, LLAMA_INPUTS_DOCSTRING
MODEL_FOR_CAUSAL_LM_MAPPING_NAMES["deci"] = "DeciLMForCausalLM"
_CONFIG_FOR_DOC = "DeciLMConfig"
logger = logging.get_logger(__name__)
class DeciLMAttention(LlamaAttention):
"""Multi-headed attention from 'Attention Is All You Need' paper"""
def __init__(self, config: DeciLMConfig, layer_idx: int):
nn.Module.__init__(self)
self.config = config
self.hidden_size = config.hidden_size
self.num_heads = config.num_attention_heads
self.head_dim = self.hidden_size // self.num_heads
self.layer_idx = layer_idx
self.num_key_value_heads = config.num_key_value_heads_per_layer[layer_idx]
self.num_key_value_groups = self.num_heads // self.num_key_value_heads
self.pretraining_tp = config.pretraining_tp
self.max_position_embeddings = config.max_position_embeddings
self.rope_theta = getattr(config, 'rope_theta', None)
if (self.head_dim * self.num_heads) != self.hidden_size:
raise ValueError(
f"hidden_size must be divisible by num_heads (got `hidden_size`: {self.hidden_size}"
f" and `num_heads`: {self.num_heads})."
)
self.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=False)
self.k_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
self.v_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
self.o_proj = nn.Linear(self.num_heads * self.head_dim, self.hidden_size, bias=False)
self._init_rope()
def forward(
self,
hidden_states: torch.Tensor,
attention_mask: Optional[torch.Tensor] = None,
position_ids: Optional[torch.LongTensor] = None,
past_key_value: Optional[Tuple[torch.Tensor]] = None,
output_attentions: bool = False,
use_cache: bool = False,
**kwargs,
) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
bsz, q_len, _ = hidden_states.size()
is_decode = past_key_value is not None
if self.pretraining_tp > 1:
key_value_slicing = (self.num_key_value_heads * self.head_dim) // self.pretraining_tp
query_slices = self.q_proj.weight.split((self.num_heads * self.head_dim) // self.pretraining_tp, dim=0)
key_slices = self.k_proj.weight.split(key_value_slicing, dim=0)
value_slices = self.v_proj.weight.split(key_value_slicing, dim=0)
query_states = [F.linear(hidden_states, query_slices[i]) for i in range(self.pretraining_tp)]
query_states = torch.cat(query_states, dim=-1)
key_states = [F.linear(hidden_states, key_slices[i]) for i in range(self.pretraining_tp)]
key_states = torch.cat(key_states, dim=-1)
value_states = [F.linear(hidden_states, value_slices[i]) for i in range(self.pretraining_tp)]
value_states = torch.cat(value_states, dim=-1)
else:
query_states = self.q_proj(hidden_states)
key_states = self.k_proj(hidden_states)
value_states = self.v_proj(hidden_states)
query_states = query_states.view(bsz, q_len, self.num_heads, self.head_dim).transpose(1, 2)
key_states = key_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
value_states = value_states.view(bsz, q_len, self.num_key_value_heads, self.head_dim).transpose(1, 2)
kv_seq_len = key_states.shape[-2]
if past_key_value is not None:
kv_seq_len += past_key_value[0].shape[-2]
cos, sin = self.rotary_emb(value_states, seq_len=kv_seq_len)
query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, position_ids)
if past_key_value is not None:
# reuse k, v, self_attention
key_states = torch.cat([past_key_value[0], key_states], dim=2)
value_states = torch.cat([past_key_value[1], value_states], dim=2)
past_key_value = (key_states, value_states) if use_cache else None
# repeat k/v heads if n_kv_heads < n_heads
key_states = repeat_kv(key_states, self.num_key_value_groups)
value_states = repeat_kv(value_states, self.num_key_value_groups)
if is_decode:
with torch.backends.cuda.sdp_kernel(enable_math=True, enable_flash=True,
enable_mem_efficient=attention_mask is None):
attn_output = F.scaled_dot_product_attention(query_states, key_states, value_states,
is_causal=False,
attn_mask=attention_mask)
attn_output = attn_output.contiguous().view(bsz, q_len, self.hidden_size)
else:
with torch.backends.cuda.sdp_kernel(enable_math=True, enable_flash=False, enable_mem_efficient=False):
attn_output = F.scaled_dot_product_attention(query_states, key_states, value_states,
is_causal=attention_mask is None,
attn_mask=attention_mask)
if attn_output.size() != (bsz, self.num_heads, q_len, self.head_dim):
raise ValueError(
f"`attn_output` should be of size {(bsz, self.num_heads, q_len, self.head_dim)}, but is"
f" {attn_output.size()}"
)
attn_output = attn_output.transpose(1, 2).contiguous().view(bsz, q_len, self.hidden_size)
if self.pretraining_tp > 1:
attn_output = attn_output.split(self.hidden_size // self.pretraining_tp, dim=2)
o_proj_slices = self.o_proj.weight.split(self.hidden_size // self.pretraining_tp, dim=1)
attn_output = sum([F.linear(attn_output[i], o_proj_slices[i]) for i in range(self.pretraining_tp)])
else:
attn_output = self.o_proj(attn_output)
attn_weights = None
return attn_output, attn_weights, past_key_value
class DeciLMDecoderLayer(LlamaDecoderLayer):
def __init__(self, config: DeciLMConfig, layer_idx: int):
nn.Module.__init__(self)
self.hidden_size = config.hidden_size
self.layer_idx = layer_idx
self.self_attn = DeciLMAttention(config=config, layer_idx=layer_idx)
self.mlp = LlamaMLP(config)
self.input_layernorm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
self.post_attention_layernorm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
@add_start_docstrings(
"The bare DeciLM Model outputting raw hidden-states without any specific head on top.",
LLAMA_START_DOCSTRING,
)
class DeciLMPreTrainedModel(LlamaPreTrainedModel):
config_class = DeciLMConfig
_no_split_modules = ["DeciLMDecoderLayer"]
_keys_to_ignore_on_load_missing = ["self_attn.rotary_emb.inv_freq"]
@add_start_docstrings(
"The bare DeciLM Model outputting raw hidden-states without any specific head on top.",
LLAMA_START_DOCSTRING,
)
class DeciLMModel(LlamaModel, DeciLMPreTrainedModel):
"""
Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`DeciLMDecoderLayer`]
Args:
config: DeciLMConfig
"""
def __init__(self, config: DeciLMConfig):
DeciLMPreTrainedModel.__init__(self, config)
self.padding_idx = config.pad_token_id
self.vocab_size = config.vocab_size
self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
self.layers = nn.ModuleList([DeciLMDecoderLayer(config, layer_idx) for layer_idx
in range(config.num_hidden_layers)])
self.norm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
self.gradient_checkpointing = False
# Initialize weights and apply final processing
self.post_init()
@add_start_docstrings_to_model_forward(LLAMA_INPUTS_DOCSTRING)
def forward(
self,
input_ids: torch.LongTensor = None,
attention_mask: Optional[torch.Tensor] = None,
position_ids: Optional[torch.LongTensor] = None,
past_key_values: Optional[List[torch.FloatTensor]] = None,
inputs_embeds: 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, BaseModelOutputWithPast]:
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
# retrieve input_ids and inputs_embeds
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:
batch_size, seq_length = input_ids.shape[:2]
elif inputs_embeds is not None:
batch_size, seq_length = inputs_embeds.shape[:2]
else:
raise ValueError("You have to specify either input_ids or inputs_embeds")
past_key_values_length = 0
if past_key_values is not None:
past_key_values_length = past_key_values[0][0].shape[2]
if position_ids is None:
device = input_ids.device if input_ids is not None else inputs_embeds.device
position_ids = torch.arange(
past_key_values_length, seq_length + past_key_values_length, dtype=torch.long, device=device
)
position_ids = position_ids.unsqueeze(0)
if inputs_embeds is None:
inputs_embeds = self.embed_tokens(input_ids)
attention_mask = attention_mask if (attention_mask is not None and 0 in attention_mask) else None
if attention_mask is not None:
# 4d mask is passed through the layers
attention_mask = _prepare_4d_causal_attention_mask(
attention_mask, (batch_size, seq_length), inputs_embeds, past_key_values_length
)
# embed positions
hidden_states = inputs_embeds
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
# decoder layers
all_hidden_states = () if output_hidden_states else None
all_self_attns = () if output_attentions else None
next_decoder_cache = () if use_cache else None
for idx, decoder_layer in enumerate(self.layers):
if output_hidden_states:
all_hidden_states += (hidden_states,)
past_key_value = past_key_values[idx] if past_key_values is not None else None
if self.gradient_checkpointing and self.training:
layer_outputs = self._gradient_checkpointing_func(
decoder_layer.__call__,
hidden_states,
attention_mask,
position_ids,
past_key_value,
output_attentions,
use_cache,
)
else:
layer_outputs = decoder_layer(
hidden_states,
attention_mask=attention_mask,
position_ids=position_ids,
past_key_value=past_key_value,
output_attentions=output_attentions,
use_cache=use_cache,
)
hidden_states = layer_outputs[0]
if use_cache:
next_decoder_cache += (layer_outputs[2 if output_attentions else 1],)
if output_attentions:
all_self_attns += (layer_outputs[1],)
hidden_states = self.norm(hidden_states)
# add hidden states from the last decoder layer
if output_hidden_states:
all_hidden_states += (hidden_states,)
next_cache = next_decoder_cache if use_cache else None
if not return_dict:
return tuple(v for v in [hidden_states, next_cache, all_hidden_states, all_self_attns] if v is not None)
return BaseModelOutputWithPast(
last_hidden_state=hidden_states,
past_key_values=next_cache,
hidden_states=all_hidden_states,
attentions=all_self_attns,
)
class DeciLMForCausalLM(LlamaForCausalLM, DeciLMPreTrainedModel):
def __init__(self, config):
DeciLMPreTrainedModel.__init__(self, config)
self.model = DeciLMModel(config)
self.pretraining_tp = config.pretraining_tp
self.vocab_size = config.vocab_size
self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
# Initialize weights and apply final processing
self.post_init()