776 lines
33 KiB
Python
776 lines
33 KiB
Python
# coding=utf-8
|
|
# Copyright 2023 EleutherAI and the HuggingFace Inc. team. All rights reserved.
|
|
#
|
|
# This code is based on EleutherAI's GPT-NeoX library and the GPT-NeoX
|
|
# and OPT implementations in this library. It has been modified from its
|
|
# original forms to accommodate minor architectural differences compared
|
|
# to GPT-NeoX and OPT used by the Meta AI team that trained the model.
|
|
#
|
|
# 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.
|
|
"""PyTorch Persimmon model."""
|
|
|
|
from typing import Callable, Optional, Union
|
|
|
|
import torch
|
|
import torch.utils.checkpoint
|
|
from torch import nn
|
|
|
|
from ...activations import ACT2FN
|
|
from ...cache_utils import Cache, DynamicCache
|
|
from ...generation import GenerationMixin
|
|
from ...modeling_attn_mask_utils import AttentionMaskConverter
|
|
from ...modeling_flash_attention_utils import FlashAttentionKwargs
|
|
from ...modeling_layers import (
|
|
GenericForSequenceClassification,
|
|
GenericForTokenClassification,
|
|
GradientCheckpointingLayer,
|
|
)
|
|
from ...modeling_outputs import (
|
|
BaseModelOutputWithPast,
|
|
CausalLMOutputWithPast,
|
|
)
|
|
from ...modeling_rope_utils import ROPE_INIT_FUNCTIONS, dynamic_rope_update
|
|
from ...modeling_utils import ALL_ATTENTION_FUNCTIONS, PreTrainedModel
|
|
from ...processing_utils import Unpack
|
|
from ...utils import auto_docstring, can_return_tuple, is_torch_flex_attn_available, logging
|
|
from .configuration_persimmon import PersimmonConfig
|
|
|
|
|
|
if is_torch_flex_attn_available():
|
|
from torch.nn.attention.flex_attention import BlockMask
|
|
|
|
from ...integrations.flex_attention import make_flex_block_causal_mask
|
|
|
|
|
|
logger = logging.get_logger(__name__)
|
|
|
|
|
|
# Copied from transformers.models.llama.modeling_llama.LlamaRotaryEmbedding with Llama->Persimmon
|
|
class PersimmonRotaryEmbedding(nn.Module):
|
|
def __init__(self, config: PersimmonConfig, device=None):
|
|
super().__init__()
|
|
# BC: "rope_type" was originally "type"
|
|
if hasattr(config, "rope_scaling") and isinstance(config.rope_scaling, dict):
|
|
self.rope_type = config.rope_scaling.get("rope_type", config.rope_scaling.get("type"))
|
|
else:
|
|
self.rope_type = "default"
|
|
self.max_seq_len_cached = config.max_position_embeddings
|
|
self.original_max_seq_len = config.max_position_embeddings
|
|
|
|
self.config = config
|
|
self.rope_init_fn = ROPE_INIT_FUNCTIONS[self.rope_type]
|
|
|
|
inv_freq, self.attention_scaling = self.rope_init_fn(self.config, device)
|
|
self.register_buffer("inv_freq", inv_freq, persistent=False)
|
|
self.original_inv_freq = self.inv_freq
|
|
|
|
@torch.no_grad()
|
|
@dynamic_rope_update # power user: used with advanced RoPE types (e.g. dynamic rope)
|
|
def forward(self, x, position_ids):
|
|
inv_freq_expanded = self.inv_freq[None, :, None].float().expand(position_ids.shape[0], -1, 1).to(x.device)
|
|
position_ids_expanded = position_ids[:, None, :].float()
|
|
|
|
device_type = x.device.type if isinstance(x.device.type, str) and x.device.type != "mps" else "cpu"
|
|
with torch.autocast(device_type=device_type, enabled=False): # Force float32
|
|
freqs = (inv_freq_expanded.float() @ position_ids_expanded.float()).transpose(1, 2)
|
|
emb = torch.cat((freqs, freqs), dim=-1)
|
|
cos = emb.cos() * self.attention_scaling
|
|
sin = emb.sin() * self.attention_scaling
|
|
|
|
return cos.to(dtype=x.dtype), sin.to(dtype=x.dtype)
|
|
|
|
|
|
# Copied from transformers.models.llama.modeling_llama.rotate_half
|
|
def rotate_half(x):
|
|
"""Rotates half the hidden dims of the input."""
|
|
x1 = x[..., : x.shape[-1] // 2]
|
|
x2 = x[..., x.shape[-1] // 2 :]
|
|
return torch.cat((-x2, x1), dim=-1)
|
|
|
|
|
|
# Copied from transformers.models.llama.modeling_llama.apply_rotary_pos_emb
|
|
def apply_rotary_pos_emb(q, k, cos, sin, position_ids=None, unsqueeze_dim=1):
|
|
"""Applies Rotary Position Embedding to the query and key tensors.
|
|
|
|
Args:
|
|
q (`torch.Tensor`): The query tensor.
|
|
k (`torch.Tensor`): The key tensor.
|
|
cos (`torch.Tensor`): The cosine part of the rotary embedding.
|
|
sin (`torch.Tensor`): The sine part of the rotary embedding.
|
|
position_ids (`torch.Tensor`, *optional*):
|
|
Deprecated and unused.
|
|
unsqueeze_dim (`int`, *optional*, defaults to 1):
|
|
The 'unsqueeze_dim' argument specifies the dimension along which to unsqueeze cos[position_ids] and
|
|
sin[position_ids] so that they can be properly broadcasted to the dimensions of q and k. For example, note
|
|
that cos[position_ids] and sin[position_ids] have the shape [batch_size, seq_len, head_dim]. Then, if q and
|
|
k have the shape [batch_size, heads, seq_len, head_dim], then setting unsqueeze_dim=1 makes
|
|
cos[position_ids] and sin[position_ids] broadcastable to the shapes of q and k. Similarly, if q and k have
|
|
the shape [batch_size, seq_len, heads, head_dim], then set unsqueeze_dim=2.
|
|
Returns:
|
|
`tuple(torch.Tensor)` comprising of the query and key tensors rotated using the Rotary Position Embedding.
|
|
"""
|
|
cos = cos.unsqueeze(unsqueeze_dim)
|
|
sin = sin.unsqueeze(unsqueeze_dim)
|
|
q_embed = (q * cos) + (rotate_half(q) * sin)
|
|
k_embed = (k * cos) + (rotate_half(k) * sin)
|
|
return q_embed, k_embed
|
|
|
|
|
|
# Copied from transformers.models.gpt_neox.modeling_gpt_neox.GPTNeoXMLP with GPTNeoX->Persimmon
|
|
class PersimmonMLP(nn.Module):
|
|
def __init__(self, config):
|
|
super().__init__()
|
|
self.dense_h_to_4h = nn.Linear(config.hidden_size, config.intermediate_size)
|
|
self.dense_4h_to_h = nn.Linear(config.intermediate_size, config.hidden_size)
|
|
self.act = ACT2FN[config.hidden_act]
|
|
|
|
def forward(self, hidden_states):
|
|
hidden_states = self.dense_h_to_4h(hidden_states)
|
|
hidden_states = self.act(hidden_states)
|
|
hidden_states = self.dense_4h_to_h(hidden_states)
|
|
return hidden_states
|
|
|
|
|
|
def eager_attention_forward(
|
|
module: nn.Module,
|
|
query: torch.Tensor,
|
|
key: torch.Tensor,
|
|
value: torch.Tensor,
|
|
attention_mask: Optional[torch.Tensor],
|
|
scaling: float,
|
|
dropout: float = 0.0,
|
|
**kwargs,
|
|
):
|
|
attn_weights = torch.matmul(query, key.transpose(2, 3)) * scaling
|
|
if attention_mask is not None:
|
|
causal_mask = attention_mask[:, :, :, : key.shape[-2]]
|
|
attn_weights = attn_weights + causal_mask
|
|
|
|
attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query.dtype)
|
|
attn_weights = nn.functional.dropout(attn_weights, p=dropout, training=module.training)
|
|
attn_output = torch.matmul(attn_weights, value)
|
|
attn_output = attn_output.transpose(1, 2).contiguous()
|
|
|
|
return attn_output, attn_weights
|
|
|
|
|
|
class PersimmonAttention(nn.Module):
|
|
"""Multi-headed attention from 'Attention Is All You Need' paper"""
|
|
|
|
def __init__(self, config: PersimmonConfig, layer_idx: Optional[int] = None):
|
|
super().__init__()
|
|
self.config = config
|
|
self.layer_idx = layer_idx
|
|
if layer_idx is None:
|
|
logger.warning_once(
|
|
f"Instantiating {self.__class__.__name__} without passing a `layer_idx` is not recommended and will "
|
|
"lead to errors during the forward call if caching is used. Please make sure to provide a `layer_idx` "
|
|
"when creating this class."
|
|
)
|
|
|
|
self.hidden_size = config.hidden_size
|
|
self.num_heads = config.num_attention_heads
|
|
self.head_dim = self.hidden_size // self.num_heads
|
|
self.rope_theta = config.rope_theta
|
|
self.rotary_ndims = int(self.head_dim * config.partial_rotary_factor)
|
|
self.is_causal = True
|
|
|
|
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.query_key_value = nn.Linear(self.hidden_size, 3 * self.hidden_size, bias=True)
|
|
self.dense = nn.Linear(self.num_heads * self.head_dim, self.hidden_size, bias=True)
|
|
self.qk_layernorm = config.qk_layernorm
|
|
self.scaling = self.head_dim**-0.5
|
|
|
|
if self.qk_layernorm:
|
|
self.q_layernorm = nn.LayerNorm(
|
|
config.hidden_size // self.num_heads, eps=config.layer_norm_eps, elementwise_affine=True
|
|
)
|
|
self.k_layernorm = nn.LayerNorm(
|
|
config.hidden_size // self.num_heads, eps=config.layer_norm_eps, elementwise_affine=True
|
|
)
|
|
self.attention_dropout = nn.Dropout(config.attention_dropout)
|
|
self.rotary_emb = PersimmonRotaryEmbedding(config=self.config)
|
|
|
|
def _split_heads(self, fused_qkv: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
|
"""
|
|
Split the last dimension into (num_heads, head_dim) without making any copies, results share same memory
|
|
storage as `fused_qkv`
|
|
|
|
Args:
|
|
fused_qkv (`torch.tensor`): [batch_size, seq_length, num_heads * 3 * head_dim]
|
|
|
|
Returns:
|
|
query: [batch_size, seq_length, num_heads, head_dim] key: [batch_size, seq_length, num_heads, head_dim]
|
|
value: [batch_size, seq_length, num_heads, head_dim]
|
|
"""
|
|
batch_size, seq_length, three_times_hidden_size = fused_qkv.shape
|
|
fused_qkv = fused_qkv.view(batch_size, seq_length, self.num_heads, 3, self.head_dim)
|
|
return fused_qkv[..., 0, :], fused_qkv[..., 1, :], fused_qkv[..., 2, :]
|
|
|
|
def forward(
|
|
self,
|
|
hidden_states: torch.Tensor,
|
|
attention_mask: Optional[torch.Tensor] = None,
|
|
position_ids: Optional[torch.LongTensor] = None,
|
|
past_key_value: Optional[Cache] = None,
|
|
output_attentions: bool = False,
|
|
use_cache: bool = False,
|
|
cache_position: Optional[torch.LongTensor] = None,
|
|
position_embeddings: Optional[tuple[torch.Tensor, torch.Tensor]] = None, # necessary, but kept here for BC
|
|
**kwargs: Unpack[FlashAttentionKwargs],
|
|
) -> tuple[torch.Tensor, Optional[torch.Tensor], Optional[tuple[torch.Tensor]]]:
|
|
bsz, q_len, _ = hidden_states.size()
|
|
|
|
# [batch_size, seq_length, 3 x hidden_size]
|
|
fused_qkv = self.query_key_value(hidden_states)
|
|
|
|
# 3 x [batch_size, seq_length, num_heads, head_dim]
|
|
(query_states, key_states, value_states) = self._split_heads(fused_qkv)
|
|
|
|
if self.qk_layernorm:
|
|
query_states = self.q_layernorm(query_states)
|
|
key_states = self.k_layernorm(key_states)
|
|
|
|
# [batch_size, num_heads, seq_length, head_dim] -> [batch_size, seq_length, num_heads, head_dim]
|
|
query_states = query_states.transpose(1, 2)
|
|
value_states = value_states.transpose(1, 2)
|
|
key_states = key_states.transpose(1, 2)
|
|
|
|
cos, sin = position_embeddings
|
|
|
|
# Partial rotary embedding
|
|
query_rot, query_pass = (
|
|
query_states[..., : self.rotary_ndims],
|
|
query_states[..., self.rotary_ndims :],
|
|
)
|
|
key_rot, key_pass = (
|
|
key_states[..., : self.rotary_ndims],
|
|
key_states[..., self.rotary_ndims :],
|
|
)
|
|
# [batch_size, seq_length, num_heads, head_dim // config.partial_rotary_factor]
|
|
query_rot, key_rot = apply_rotary_pos_emb(query_rot, key_rot, cos, sin)
|
|
|
|
# [batch_size, seq_length, num_heads, head_dim]
|
|
query_states = torch.cat((query_rot, query_pass), dim=-1)
|
|
key_states = torch.cat((key_rot, key_pass), dim=-1)
|
|
|
|
if past_key_value is not None:
|
|
# Specific to RoPE models with partial rotation
|
|
cache_kwargs = {
|
|
"sin": sin,
|
|
"cos": cos,
|
|
"partial_rotation_size": self.rotary_ndims,
|
|
"cache_position": cache_position,
|
|
}
|
|
key_states, value_states = past_key_value.update(key_states, value_states, self.layer_idx, cache_kwargs)
|
|
|
|
attention_interface: Callable = eager_attention_forward
|
|
if self.config._attn_implementation != "eager":
|
|
attention_interface = ALL_ATTENTION_FUNCTIONS[self.config._attn_implementation]
|
|
|
|
attn_output, attn_weights = attention_interface(
|
|
self,
|
|
query_states,
|
|
key_states,
|
|
value_states,
|
|
attention_mask,
|
|
dropout=0.0 if not self.training else self.config.attention_dropout,
|
|
scaling=self.scaling,
|
|
**kwargs,
|
|
)
|
|
|
|
attn_output = attn_output.reshape(bsz, q_len, -1)
|
|
attn_output = self.dense(attn_output)
|
|
|
|
if not output_attentions:
|
|
attn_weights = None
|
|
|
|
return attn_output, attn_weights
|
|
|
|
|
|
class PersimmonDecoderLayer(GradientCheckpointingLayer):
|
|
def __init__(self, config: PersimmonConfig, layer_idx: int):
|
|
super().__init__()
|
|
self.hidden_size = config.hidden_size
|
|
self.self_attn = PersimmonAttention(config=config, layer_idx=layer_idx)
|
|
self.mlp = PersimmonMLP(config)
|
|
self.input_layernorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
|
|
self.post_attention_layernorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
|
|
self.dropout = nn.Dropout(config.hidden_dropout)
|
|
|
|
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: Optional[bool] = False,
|
|
use_cache: Optional[bool] = False,
|
|
cache_position: Optional[torch.LongTensor] = None,
|
|
position_embeddings: Optional[tuple[torch.Tensor, torch.Tensor]] = None, # necessary, but kept here for BC
|
|
**kwargs: Unpack[FlashAttentionKwargs],
|
|
) -> tuple[torch.FloatTensor, Optional[tuple[torch.FloatTensor, torch.FloatTensor]]]:
|
|
"""
|
|
Args:
|
|
hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)`
|
|
attention_mask (`torch.FloatTensor`, *optional*): attention mask of size
|
|
`(batch, 1, tgt_len, src_len)` where padding elements are indicated by very large negative values.
|
|
position_ids (`torch.LongTensor` of shape `({0})`, *optional*):
|
|
Indices of positions of each input sequence tokens in the position embeddings. Selected in the range
|
|
`[0, config.n_positions - 1]`.
|
|
[What are position IDs?](../glossary#position-ids)
|
|
past_key_value (`Tuple(torch.FloatTensor)`, *optional*):
|
|
cached past key and value projection states
|
|
output_attentions (`bool`, *optional*):
|
|
Whether or not to return the attentions tensors of all attention layers. See `attentions` under
|
|
returned tensors for more detail.
|
|
use_cache (`bool`, *optional*):
|
|
If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding
|
|
(see `past_key_values`).
|
|
cache_position (`torch.LongTensor` of shape `(sequence_length)`, *optional*):
|
|
Indices depicting the position of the input sequence tokens in the sequence
|
|
position_embeddings (`tuple[torch.FloatTensor, torch.FloatTensor]`, *optional*):
|
|
Tuple containing the cosine and sine positional embeddings of shape `(batch_size, seq_len, head_dim)`,
|
|
with `head_dim` being the embedding dimension of each attention head.
|
|
"""
|
|
|
|
residual = hidden_states
|
|
|
|
hidden_states = self.input_layernorm(hidden_states)
|
|
|
|
# Self Attention
|
|
hidden_states, self_attn_weights = self.self_attn(
|
|
hidden_states=hidden_states,
|
|
attention_mask=attention_mask,
|
|
position_ids=position_ids,
|
|
past_key_value=past_key_value,
|
|
output_attentions=output_attentions,
|
|
use_cache=use_cache,
|
|
cache_position=cache_position,
|
|
position_embeddings=position_embeddings,
|
|
**kwargs,
|
|
)
|
|
hidden_states = residual + hidden_states
|
|
|
|
# Fully Connected
|
|
residual = hidden_states
|
|
hidden_states = self.post_attention_layernorm(hidden_states)
|
|
hidden_states = self.mlp(hidden_states)
|
|
|
|
hidden_states = self.dropout(hidden_states)
|
|
hidden_states = hidden_states + residual
|
|
|
|
outputs = (hidden_states,)
|
|
|
|
if output_attentions:
|
|
outputs += (self_attn_weights,)
|
|
|
|
return outputs
|
|
|
|
|
|
@auto_docstring
|
|
class PersimmonPreTrainedModel(PreTrainedModel):
|
|
config: PersimmonConfig
|
|
base_model_prefix = "model"
|
|
supports_gradient_checkpointing = True
|
|
_no_split_modules = ["PersimmonDecoderLayer"]
|
|
_skip_keys_device_placement = "past_key_values"
|
|
|
|
_can_compile_fullgraph = True
|
|
_supports_sdpa = True
|
|
_supports_flash_attn = True
|
|
_supports_attention_backend = True
|
|
|
|
def _init_weights(self, module):
|
|
std = self.config.initializer_range
|
|
if isinstance(module, nn.Linear):
|
|
module.weight.data.normal_(mean=0.0, std=std)
|
|
if module.bias is not None:
|
|
module.bias.data.zero_()
|
|
elif isinstance(module, nn.Embedding):
|
|
module.weight.data.normal_(mean=0.0, std=std)
|
|
if module.padding_idx is not None:
|
|
module.weight.data[module.padding_idx].zero_()
|
|
elif isinstance(module, nn.LayerNorm):
|
|
module.weight.data.fill_(1.0)
|
|
module.bias.data.zero_()
|
|
|
|
|
|
@auto_docstring
|
|
class PersimmonModel(PersimmonPreTrainedModel):
|
|
"""
|
|
Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`PersimmonDecoderLayer`]
|
|
|
|
Args:
|
|
config: PersimmonConfig
|
|
"""
|
|
|
|
def __init__(self, config: PersimmonConfig):
|
|
super().__init__(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(
|
|
[PersimmonDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
|
|
)
|
|
self.final_layernorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
|
|
|
|
self.rotary_emb = PersimmonRotaryEmbedding(config=config)
|
|
|
|
self.gradient_checkpointing = False
|
|
# Initialize weights and apply final processing
|
|
self.post_init()
|
|
|
|
@can_return_tuple
|
|
@auto_docstring
|
|
def forward(
|
|
self,
|
|
input_ids: Optional[torch.LongTensor] = None,
|
|
attention_mask: Optional[torch.Tensor] = None,
|
|
position_ids: Optional[torch.LongTensor] = None,
|
|
past_key_values: Optional[Cache] = None,
|
|
inputs_embeds: Optional[torch.FloatTensor] = None,
|
|
use_cache: Optional[bool] = None,
|
|
output_attentions: Optional[bool] = None,
|
|
output_hidden_states: Optional[bool] = None,
|
|
cache_position: Optional[torch.LongTensor] = None,
|
|
**kwargs: Unpack[FlashAttentionKwargs],
|
|
) -> 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
|
|
|
|
if (input_ids is None) ^ (inputs_embeds is not None):
|
|
raise ValueError("You must specify exactly one of input_ids or 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
|
|
|
|
# TODO (joao): remove this exception in v4.56 -- it exists for users that try to pass a legacy cache
|
|
if not isinstance(past_key_values, (type(None), Cache)):
|
|
raise ValueError("The `past_key_values` should be either a `Cache` object or `None`.")
|
|
|
|
if use_cache and past_key_values is None:
|
|
past_key_values = DynamicCache()
|
|
|
|
if inputs_embeds is None:
|
|
inputs_embeds = self.embed_tokens(input_ids)
|
|
|
|
if cache_position is None:
|
|
past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
|
|
cache_position = torch.arange(
|
|
past_seen_tokens, past_seen_tokens + inputs_embeds.shape[1], device=inputs_embeds.device
|
|
)
|
|
if position_ids is None:
|
|
position_ids = cache_position.unsqueeze(0)
|
|
|
|
causal_mask = self._update_causal_mask(
|
|
attention_mask, inputs_embeds, cache_position, past_key_values, output_attentions
|
|
)
|
|
|
|
hidden_states = inputs_embeds
|
|
|
|
# create position embeddings to be shared across the decoder layers
|
|
position_embeddings = self.rotary_emb(hidden_states, position_ids)
|
|
|
|
# decoder layers
|
|
all_hidden_states = () if output_hidden_states else None
|
|
all_self_attns = () if output_attentions else None
|
|
|
|
for decoder_layer in self.layers:
|
|
if output_hidden_states:
|
|
all_hidden_states += (hidden_states,)
|
|
|
|
layer_outputs = decoder_layer(
|
|
hidden_states,
|
|
attention_mask=causal_mask,
|
|
position_ids=position_ids,
|
|
past_key_value=past_key_values,
|
|
output_attentions=output_attentions,
|
|
use_cache=use_cache,
|
|
cache_position=cache_position,
|
|
position_embeddings=position_embeddings,
|
|
**kwargs,
|
|
)
|
|
|
|
hidden_states = layer_outputs[0]
|
|
|
|
if output_attentions:
|
|
all_self_attns += (layer_outputs[1],)
|
|
|
|
hidden_states = self.final_layernorm(hidden_states)
|
|
|
|
# add hidden states from the last decoder layer
|
|
if output_hidden_states:
|
|
all_hidden_states += (hidden_states,)
|
|
|
|
return BaseModelOutputWithPast(
|
|
last_hidden_state=hidden_states,
|
|
past_key_values=past_key_values,
|
|
hidden_states=all_hidden_states,
|
|
attentions=all_self_attns,
|
|
)
|
|
|
|
# Copied from transformers.models.gptj.modeling_gptj.GPTJModel._update_causal_mask
|
|
def _update_causal_mask(
|
|
self,
|
|
attention_mask: Union[torch.Tensor, "BlockMask"],
|
|
input_tensor: torch.Tensor,
|
|
cache_position: torch.Tensor,
|
|
past_key_values: Cache,
|
|
output_attentions: bool = False,
|
|
):
|
|
if self.config._attn_implementation == "flash_attention_2":
|
|
if attention_mask is not None and (attention_mask == 0.0).any():
|
|
return attention_mask
|
|
return None
|
|
if self.config._attn_implementation == "flex_attention":
|
|
if isinstance(attention_mask, torch.Tensor):
|
|
attention_mask = make_flex_block_causal_mask(attention_mask)
|
|
return attention_mask
|
|
|
|
# For SDPA, when possible, we will rely on its `is_causal` argument instead of its `attn_mask` argument, in
|
|
# order to dispatch on Flash Attention 2. This feature is not compatible with static cache, as SDPA will fail
|
|
# to infer the attention mask.
|
|
past_seen_tokens = past_key_values.get_seq_length() if past_key_values is not None else 0
|
|
using_compilable_cache = past_key_values.is_compileable if past_key_values is not None else False
|
|
|
|
# When output attentions is True, sdpa implementation's forward method calls the eager implementation's forward
|
|
if self.config._attn_implementation == "sdpa" and not using_compilable_cache and not output_attentions:
|
|
if AttentionMaskConverter._ignore_causal_mask_sdpa(
|
|
attention_mask,
|
|
inputs_embeds=input_tensor,
|
|
past_key_values_length=past_seen_tokens,
|
|
is_training=self.training,
|
|
):
|
|
return None
|
|
|
|
dtype = input_tensor.dtype
|
|
sequence_length = input_tensor.shape[1]
|
|
if using_compilable_cache:
|
|
target_length = past_key_values.get_max_cache_shape()
|
|
else:
|
|
target_length = (
|
|
attention_mask.shape[-1]
|
|
if isinstance(attention_mask, torch.Tensor)
|
|
else past_seen_tokens + sequence_length + 1
|
|
)
|
|
|
|
# In case the provided `attention` mask is 2D, we generate a causal mask here (4D).
|
|
causal_mask = self._prepare_4d_causal_attention_mask_with_cache_position(
|
|
attention_mask,
|
|
sequence_length=sequence_length,
|
|
target_length=target_length,
|
|
dtype=dtype,
|
|
cache_position=cache_position,
|
|
batch_size=input_tensor.shape[0],
|
|
)
|
|
|
|
if (
|
|
self.config._attn_implementation == "sdpa"
|
|
and attention_mask is not None
|
|
and attention_mask.device.type in ["cuda", "xpu", "npu"]
|
|
and not output_attentions
|
|
):
|
|
# Attend to all tokens in fully masked rows in the causal_mask, for example the relevant first rows when
|
|
# using left padding. This is required by F.scaled_dot_product_attention memory-efficient attention path.
|
|
# Details: https://github.com/pytorch/pytorch/issues/110213
|
|
min_dtype = torch.finfo(dtype).min
|
|
causal_mask = AttentionMaskConverter._unmask_unattended(causal_mask, min_dtype)
|
|
|
|
return causal_mask
|
|
|
|
@staticmethod
|
|
# Copied from transformers.models.gptj.modeling_gptj.GPTJModel._prepare_4d_causal_attention_mask_with_cache_position
|
|
def _prepare_4d_causal_attention_mask_with_cache_position(
|
|
attention_mask: torch.Tensor,
|
|
sequence_length: int,
|
|
target_length: int,
|
|
dtype: torch.dtype,
|
|
cache_position: torch.Tensor,
|
|
batch_size: int,
|
|
**kwargs,
|
|
):
|
|
"""
|
|
Creates a causal 4D mask of shape `(batch_size, 1, query_length, key_value_length)` from a 2D mask of shape
|
|
`(batch_size, key_value_length)`, or if the input `attention_mask` is already 4D, do nothing.
|
|
|
|
Args:
|
|
attention_mask (`torch.Tensor`):
|
|
A 2D attention mask of shape `(batch_size, key_value_length)` or a 4D attention mask of shape
|
|
`(batch_size, 1, query_length, key_value_length)`.
|
|
sequence_length (`int`):
|
|
The sequence length being processed.
|
|
target_length (`int`):
|
|
The target length: when generating with static cache, the mask should be as long as the static cache,
|
|
to account for the 0 padding, the part of the cache that is not filled yet.
|
|
dtype (`torch.dtype`):
|
|
The dtype to use for the 4D attention mask.
|
|
cache_position (`torch.Tensor`):
|
|
Indices depicting the position of the input sequence tokens in the sequence.
|
|
batch_size (`torch.Tensor`):
|
|
Batch size.
|
|
"""
|
|
if attention_mask is not None and attention_mask.dim() == 4:
|
|
# In this case we assume that the mask comes already in inverted form and requires no inversion or slicing.
|
|
causal_mask = attention_mask
|
|
else:
|
|
min_dtype = torch.finfo(dtype).min
|
|
causal_mask = torch.full(
|
|
(sequence_length, target_length), fill_value=min_dtype, dtype=dtype, device=cache_position.device
|
|
)
|
|
if sequence_length != 1:
|
|
causal_mask = torch.triu(causal_mask, diagonal=1)
|
|
causal_mask *= torch.arange(target_length, device=cache_position.device) > cache_position.reshape(-1, 1)
|
|
causal_mask = causal_mask[None, None, :, :].expand(batch_size, 1, -1, -1)
|
|
if attention_mask is not None:
|
|
causal_mask = causal_mask.clone() # copy to contiguous memory for in-place edit
|
|
mask_length = attention_mask.shape[-1]
|
|
padding_mask = causal_mask[:, :, :, :mask_length] + attention_mask[:, None, None, :].to(
|
|
causal_mask.device
|
|
)
|
|
padding_mask = padding_mask == 0
|
|
causal_mask[:, :, :, :mask_length] = causal_mask[:, :, :, :mask_length].masked_fill(
|
|
padding_mask, min_dtype
|
|
)
|
|
|
|
return causal_mask
|
|
|
|
|
|
class PersimmonForCausalLM(PersimmonPreTrainedModel, GenerationMixin):
|
|
_tied_weights_keys = ["lm_head.weight"]
|
|
|
|
# Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.__init__ with LLAMA->PERSIMMON,Llama->Persimmon
|
|
def __init__(self, config):
|
|
super().__init__(config)
|
|
self.model = PersimmonModel(config)
|
|
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()
|
|
|
|
# Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.set_decoder
|
|
def set_decoder(self, decoder):
|
|
self.model = decoder
|
|
|
|
# Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.get_decoder
|
|
def get_decoder(self):
|
|
return self.model
|
|
|
|
@can_return_tuple
|
|
@auto_docstring
|
|
def forward(
|
|
self,
|
|
input_ids: Optional[torch.LongTensor] = None,
|
|
attention_mask: Optional[torch.Tensor] = None,
|
|
position_ids: Optional[torch.LongTensor] = None,
|
|
past_key_values: Optional[Cache] = None,
|
|
inputs_embeds: 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,
|
|
cache_position: Optional[torch.LongTensor] = None,
|
|
logits_to_keep: Union[int, torch.Tensor] = 0,
|
|
**kwargs,
|
|
) -> CausalLMOutputWithPast:
|
|
r"""
|
|
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
|
|
Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
|
|
config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
|
|
(masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.
|
|
|
|
Example:
|
|
|
|
```python
|
|
>>> from transformers import AutoTokenizer, PersimmonForCausalLM
|
|
|
|
>>> model = PersimmonForCausalLM.from_pretrained("adept/persimmon-8b-base")
|
|
>>> tokenizer = AutoTokenizer.from_pretrained("adept/persimmon-8b-base")
|
|
|
|
>>> prompt = "human: Hey, what should I eat for dinner?"
|
|
>>> inputs = tokenizer(prompt, return_tensors="pt")
|
|
|
|
>>> # Generate
|
|
>>> generate_ids = model.generate(inputs.input_ids, max_length=30)
|
|
>>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
|
|
'human: Hey, what should I eat for dinner?\n\ncat: 🐱\n\nhuman: 😐\n\n'
|
|
```"""
|
|
|
|
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
|
|
)
|
|
|
|
# decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
|
|
outputs: BaseModelOutputWithPast = self.model(
|
|
input_ids=input_ids,
|
|
attention_mask=attention_mask,
|
|
position_ids=position_ids,
|
|
past_key_values=past_key_values,
|
|
inputs_embeds=inputs_embeds,
|
|
use_cache=use_cache,
|
|
output_attentions=output_attentions,
|
|
output_hidden_states=output_hidden_states,
|
|
cache_position=cache_position,
|
|
**kwargs,
|
|
)
|
|
|
|
hidden_states = outputs.last_hidden_state
|
|
# No upscaling to float was ever done for Persimmon
|
|
slice_indices = slice(-logits_to_keep, None) if isinstance(logits_to_keep, int) else logits_to_keep
|
|
logits = self.lm_head(hidden_states[:, slice_indices, :])
|
|
|
|
loss = None
|
|
if labels is not None:
|
|
loss = self.loss_function(
|
|
logits,
|
|
labels,
|
|
vocab_size=self.config.vocab_size,
|
|
**kwargs,
|
|
)
|
|
|
|
return CausalLMOutputWithPast(
|
|
loss=loss,
|
|
logits=logits,
|
|
past_key_values=outputs.past_key_values,
|
|
hidden_states=outputs.hidden_states,
|
|
attentions=outputs.attentions,
|
|
)
|
|
|
|
|
|
class PersimmonForSequenceClassification(GenericForSequenceClassification, PersimmonPreTrainedModel): ...
|
|
|
|
|
|
class PersimmonForTokenClassification(GenericForTokenClassification, PersimmonPreTrainedModel): ...
|
|
|
|
|
|
__all__ = [
|
|
"PersimmonForCausalLM",
|
|
"PersimmonModel",
|
|
"PersimmonPreTrainedModel",
|
|
"PersimmonForSequenceClassification",
|
|
"PersimmonForTokenClassification",
|
|
]
|