# coding=utf-8
# Copyright 2024 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 StableLM model."""
import math
from typing import List, Optional, Tuple, Union
import torch
import torch.utils.checkpoint
from torch import nn
from ...activations import ACT2FN
from ...cache_utils import Cache, DynamicCache, StaticCache
from ...generation import GenerationMixin
from ...modeling_attn_mask_utils import AttentionMaskConverter
from ...modeling_flash_attention_utils import flash_attn_supports_top_left_mask, is_flash_attn_available
from ...modeling_outputs import (
BaseModelOutputWithPast,
CausalLMOutputWithPast,
SequenceClassifierOutputWithPast,
TokenClassifierOutput,
)
from ...modeling_rope_utils import ROPE_INIT_FUNCTIONS, dynamic_rope_update
from ...modeling_utils import PreTrainedModel
from ...utils import (
add_code_sample_docstrings,
add_start_docstrings,
add_start_docstrings_to_model_forward,
can_return_tuple,
is_torch_flex_attn_available,
logging,
replace_return_docstrings,
)
from ...utils.deprecation import deprecate_kwarg
from .configuration_stablelm import StableLmConfig
if is_torch_flex_attn_available():
from torch.nn.attention.flex_attention import BlockMask
from ...integrations.flex_attention import make_flex_block_causal_mask
if is_flash_attn_available():
from ...modeling_flash_attention_utils import _flash_attention_forward
logger = logging.get_logger(__name__)
_CHECKPOINT_FOR_DOC = "stabilityai/stablelm-3b-4e1t"
_CONFIG_FOR_DOC = "StableLmConfig"
# Copied from transformers.models.llama.modeling_llama.LlamaRotaryEmbedding with Llama->StableLm
class StableLmRotaryEmbedding(nn.Module):
def __init__(self, config: StableLmConfig, device=None):
super().__init__()
# BC: "rope_type" was originally "type"
if hasattr(config, "rope_scaling") and config.rope_scaling is not None:
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.mistral.modeling_mistral.MistralMLP with Mistral->StableLm
class StableLmMLP(nn.Module):
def __init__(self, config):
super().__init__()
self.config = config
self.hidden_size = config.hidden_size
self.intermediate_size = config.intermediate_size
self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
self.act_fn = ACT2FN[config.hidden_act]
def forward(self, x):
down_proj = self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
return down_proj
class StableLmLayerNormPerHead(nn.Module):
def __init__(self, dim, num_heads, eps=1e-5, bias=False):
super().__init__()
self.dim = dim
self.num_heads = num_heads
self.norms = nn.ModuleList([nn.LayerNorm(dim, eps=eps, bias=bias) for _ in range(self.num_heads)])
def forward(self, hidden_states: torch.Tensor):
# Split along the num_heads axis to get per-head inputs
# [batch_size, num_heads, seq_len, head_dim] -> [batch_size, 1, seq_len, head_dim] * num_heads
states_per_heads = torch.split(hidden_states, 1, dim=1)
# Normalize and merge the heads back together
return torch.cat([norm(hidden_states) for norm, hidden_states in zip(self.norms, states_per_heads)], dim=1)
# Copied from transformers.models.llama.modeling_llama.repeat_kv
def repeat_kv(hidden_states: torch.Tensor, n_rep: int) -> torch.Tensor:
"""
This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch,
num_key_value_heads, seqlen, head_dim) to (batch, num_attention_heads, seqlen, head_dim)
"""
batch, num_key_value_heads, slen, head_dim = hidden_states.shape
if n_rep == 1:
return hidden_states
hidden_states = hidden_states[:, :, None, :, :].expand(batch, num_key_value_heads, n_rep, slen, head_dim)
return hidden_states.reshape(batch, num_key_value_heads * n_rep, slen, head_dim)
class StableLmAttention(nn.Module):
"""Multi-headed attention from 'Attention Is All You Need' paper"""
def __init__(self, config: StableLmConfig, 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.num_key_value_heads = config.num_key_value_heads
self.num_key_value_groups = self.num_heads // self.num_key_value_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.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=config.use_qkv_bias)
self.k_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=config.use_qkv_bias)
self.v_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=config.use_qkv_bias)
self.o_proj = nn.Linear(self.hidden_size, self.hidden_size, bias=False)
self.qk_layernorm = config.qk_layernorm
if self.qk_layernorm:
self.q_layernorm = StableLmLayerNormPerHead(self.head_dim, self.num_heads, eps=config.layer_norm_eps)
self.k_layernorm = StableLmLayerNormPerHead(
self.head_dim, self.num_key_value_heads, eps=config.layer_norm_eps
)
self.attention_dropout = nn.Dropout(config.attention_dropout)
self.rotary_emb = StableLmRotaryEmbedding(config=self.config)
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
) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
bsz, q_len, _ = hidden_states.size()
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)
if self.qk_layernorm:
query_states = self.q_layernorm(query_states)
key_states = self.k_layernorm(key_states)
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)
# 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)
attn_weights = torch.matmul(query_states, key_states.transpose(2, 3)) / math.sqrt(self.head_dim)
if attention_mask is not None: # no matter the length, we just slice it
causal_mask = attention_mask[:, :, :, : key_states.shape[-2]]
attn_weights += causal_mask
# upcast attention to fp32
attn_weights = nn.functional.softmax(attn_weights, dtype=torch.float32, dim=-1).to(query_states.dtype)
attn_weights = self.attention_dropout(attn_weights)
attn_output = torch.matmul(attn_weights, value_states)
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()
attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)
attn_output = self.o_proj(attn_output)
if not output_attentions:
attn_weights = None
return attn_output, attn_weights, past_key_value
class StableLmSdpaAttention(StableLmAttention):
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
) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
if output_attentions:
# TODO: Improve this warning with e.g. `model.config.attn_implementation = "manual"` once this is implemented.
logger.warning_once(
"StableLmModel is using StableLmSdpaAttention, but `torch.nn.functional.scaled_dot_product_attention` does not support `output_attentions=True`. Falling back to the manual attention implementation, "
'but specifying the manual implementation will be required from Transformers version v5.0.0 onwards. This warning can be removed using the argument `attn_implementation="eager"` when loading the model.'
)
return super().forward(
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,
)
bsz, q_len, _ = hidden_states.size()
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)
if self.qk_layernorm:
query_states = self.q_layernorm(query_states)
key_states = self.k_layernorm(key_states)
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)
# 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)
causal_mask = attention_mask
if attention_mask is not None: # no matter the length, we just slice it
causal_mask = attention_mask[:, :, :, : key_states.shape[-2]]
# SDPA with memory-efficient backend is currently (torch==2.1.2) bugged with non-contiguous inputs with custom attn_mask,
# Reference: https://github.com/pytorch/pytorch/issues/112577.
if query_states.device.type == "cuda" and attention_mask is not None:
query_states = query_states.contiguous()
key_states = key_states.contiguous()
value_states = value_states.contiguous()
# We dispatch to SDPA's Flash Attention or Efficient kernels via this `is_causal` if statement instead of an inline conditional assignment
# in SDPA to support both torch.compile's dynamic shapes and full graph options. An inline conditional prevents dynamic shapes from compiling.
# The q_len > 1 is necessary to match with AttentionMaskConverter.to_causal_4d that does not create a causal mask in case q_len == 1.
is_causal = True if causal_mask is None and q_len > 1 else False
attn_output = torch.nn.functional.scaled_dot_product_attention(
query_states,
key_states,
value_states,
attn_mask=causal_mask,
dropout_p=self.attention_dropout.p if self.training else 0.0,
is_causal=is_causal,
)
attn_output = attn_output.transpose(1, 2).contiguous()
attn_output = attn_output.view(bsz, q_len, self.hidden_size)
attn_output = self.o_proj(attn_output)
return attn_output, None, past_key_value
class StableLmFlashAttention2(StableLmAttention):
"""
StableLM flash attention module. This module inherits from `StableLmAttention` as the weights of the module stays
untouched. The only required change would be on the forward pass where it needs to correctly call the public API of
flash attention and deal with padding tokens in case the input contains any of them.
"""
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
# TODO: Should be removed once Flash Attention for RoCm is bumped to 2.1.
# flash_attn<2.1 generates top-left aligned causal mask, while what is needed here is bottom-right alignment, that was made default for flash_attn>=2.1. This attribute is used to handle this difference. Reference: https://github.com/Dao-AILab/flash-attention/releases/tag/v2.1.0.
# Beware that with flash_attn<2.1, using q_seqlen != k_seqlen (except for the case q_seqlen == 1) produces a wrong mask (top-left).
self._flash_attn_uses_top_left_mask = flash_attn_supports_top_left_mask()
def forward(
self,
hidden_states: torch.Tensor,
attention_mask: Optional[torch.LongTensor] = 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,
) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
# StableLmFlashAttention2 attention does not support output_attentions
output_attentions = False
bsz, q_len, _ = hidden_states.size()
query_states = self.q_proj(hidden_states)
key_states = self.k_proj(hidden_states)
value_states = self.v_proj(hidden_states)
# Flash attention requires the input to have the shape
# batch_size x seq_length x head_dim x hidden_dim
# therefore we just need to keep the original shape
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)
if self.qk_layernorm:
query_states = self.q_layernorm(query_states)
key_states = self.k_layernorm(key_states)
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 :],
)
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:
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)
# TODO: These transpose are quite inefficient but Flash Attention requires the layout [batch_size, sequence_length, num_heads, head_dim]. We would need to refactor the KV cache
# to be able to avoid many of these transpose/reshape/view.
query_states = query_states.transpose(1, 2)
key_states = key_states.transpose(1, 2)
value_states = value_states.transpose(1, 2)
dropout_rate = self.attention_dropout.p if self.training else 0.0
attn_output = _flash_attention_forward(
query_states,
key_states,
value_states,
attention_mask,
q_len,
position_ids=position_ids,
dropout=dropout_rate,
use_top_left_mask=self._flash_attn_uses_top_left_mask,
is_causal=self.is_causal,
)
attn_output = attn_output.reshape(bsz, q_len, self.hidden_size).contiguous()
attn_output = self.o_proj(attn_output)
if not output_attentions:
attn_weights = None
return attn_output, attn_weights, past_key_value
ATTENTION_CLASSES = {
"eager": StableLmAttention,
"sdpa": StableLmSdpaAttention,
"flash_attention_2": StableLmFlashAttention2,
}
class StableLmDecoderLayer(nn.Module):
def __init__(self, config: StableLmConfig, layer_idx: int):
super().__init__()
self.use_parallel_residual = config.use_parallel_residual
self.hidden_size = config.hidden_size
self.self_attn = ATTENTION_CLASSES[config._attn_implementation](config, layer_idx=layer_idx)
self.mlp = StableLmMLP(config)
self.input_layernorm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
self.post_attention_layernorm = None
if not self.use_parallel_residual:
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
) -> 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
self_attn_output, self_attn_weights, present_key_value = 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,
)
# copied from transformers.models.gpt_neox.modeling_gpt_neox.GPTNeoXLayer.forward
if self.use_parallel_residual:
# x = x + attn(ln1(x)) + mlp(ln1(x))
# Fully Connected
mlp_output = self.mlp(hidden_states)
mlp_output = self.dropout(mlp_output)
hidden_states = residual + self_attn_output + mlp_output
else:
# x = x + attn(ln1(x))
# x = x + mlp(ln2(x))
residual = residual + self_attn_output
# Fully Connected
mlp_output = self.mlp(self.post_attention_layernorm(residual))
mlp_output = self.dropout(mlp_output)
hidden_states = residual + mlp_output
outputs = (hidden_states,)
if output_attentions:
outputs += (self_attn_weights,)
if use_cache:
outputs += (present_key_value,)
return outputs
STABLELM_START_DOCSTRING = r"""
This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the
library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
etc.)
This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
and behavior.
Parameters:
config ([`StableLmConfig`]):
Model configuration class with all the parameters of the model. Initializing with a config file does not
load the weights associated with the model, only the configuration. Check out the
[`~PreTrainedModel.from_pretrained`] method to load the model weights.
"""
@add_start_docstrings(
"The bare StableLm Model outputting raw hidden-states without any specific head on top.",
STABLELM_START_DOCSTRING,
)
class StableLmPreTrainedModel(PreTrainedModel):
config_class = StableLmConfig
base_model_prefix = "model"
supports_gradient_checkpointing = True
_no_split_modules = ["StableLmDecoderLayer"]
_skip_keys_device_placement = "past_key_values"
_supports_flash_attn_2 = True
_supports_cache_class = True
_supports_sdpa = True
_supports_quantized_cache = True
_supports_static_cache = 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_()
STABLELM_INPUTS_DOCSTRING = r"""
Args:
input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
it.
Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
[`PreTrainedTokenizer.__call__`] for details.
[What are input IDs?](../glossary#input-ids)
attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
- 1 for tokens that are **not masked**,
- 0 for tokens that are **masked**.
[What are attention masks?](../glossary#attention-mask)
Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
[`PreTrainedTokenizer.__call__`] for details.
If `past_key_values` is used, optionally only the last `decoder_input_ids` have to be input (see
`past_key_values`).
If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`]
and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more
information on the default strategy.
- 1 indicates the head is **not masked**,
- 0 indicates the head is **masked**.
position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *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_values (`Cache` or `tuple(tuple(torch.FloatTensor))`, *optional*):
Pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention
blocks) that can be used to speed up sequential decoding. This typically consists in the `past_key_values`
returned by the model at a previous stage of decoding, when `use_cache=True` or `config.use_cache=True`.
Two formats are allowed:
- a [`~cache_utils.Cache`] instance, see our
[kv cache guide](https://huggingface.co/docs/transformers/en/kv_cache);
- Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of
shape `(batch_size, num_heads, sequence_length, embed_size_per_head)`). This is also known as the legacy
cache format.
The model will output the same cache format that is fed as input. If no `past_key_values` are passed, the
legacy cache format will be returned.
If `past_key_values` are used, the user can optionally input only the last `input_ids` (those that don't
have their past key value states given to this model) of shape `(batch_size, 1)` instead of all `input_ids`
of shape `(batch_size, sequence_length)`.
inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
is useful if you want more control over how to convert `input_ids` indices into associated vectors than the
model's internal embedding lookup matrix.
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`).
output_attentions (`bool`, *optional*):
Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
tensors for more detail.
output_hidden_states (`bool`, *optional*):
Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
more detail.
return_dict (`bool`, *optional*):
Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
cache_position (`torch.LongTensor` of shape `(sequence_length)`, *optional*):
Indices depicting the position of the input sequence tokens in the sequence. Contrarily to `position_ids`,
this tensor is not affected by padding. It is used to update the cache in the correct position and to infer
the complete sequence length.
"""
@add_start_docstrings(
"The bare StableLm Model outputting raw hidden-states without any specific head on top.",
STABLELM_START_DOCSTRING,
)
class StableLmModel(StableLmPreTrainedModel):
"""
Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`StableLmDecoderLayer`]
Args:
config: StableLmConfig
"""
def __init__(self, config: StableLmConfig):
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(
[StableLmDecoderLayer(config, layer_idx) for layer_idx in range(config.num_hidden_layers)]
)
self.norm = nn.LayerNorm(config.hidden_size, eps=config.layer_norm_eps)
self.rotary_emb = StableLmRotaryEmbedding(config=config)
self._attn_implementation = config._attn_implementation
self.gradient_checkpointing = False
# Initialize weights and apply final processing
self.post_init()
def get_input_embeddings(self):
return self.embed_tokens
def set_input_embeddings(self, value):
self.embed_tokens = value
@can_return_tuple
@add_start_docstrings_to_model_forward(STABLELM_INPUTS_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[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,
cache_position: Optional[torch.LongTensor] = None,
) -> 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
# kept for BC (non `Cache` `past_key_values` inputs)
return_legacy_cache = False
if use_cache and not isinstance(past_key_values, Cache):
return_legacy_cache = True
if past_key_values is None:
past_key_values = DynamicCache()
else:
past_key_values = DynamicCache.from_legacy_cache(past_key_values)
logger.warning_once(
"We detected that you are passing `past_key_values` as a tuple of tuples. This is deprecated and "
"will be removed in v4.47. Please convert your cache or use an appropriate `Cache` class "
"(https://huggingface.co/docs/transformers/kv_cache#legacy-cache-format)"
)
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
next_decoder_cache = None
for decoder_layer in self.layers:
if output_hidden_states:
all_hidden_states += (hidden_states,)
if self.gradient_checkpointing and self.training:
layer_outputs = self._gradient_checkpointing_func(
decoder_layer.__call__,
hidden_states,
causal_mask,
position_ids,
past_key_values,
output_attentions,
use_cache,
cache_position,
position_embeddings,
)
else:
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,
)
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 return_legacy_cache:
next_cache = next_cache.to_legacy_cache()
return BaseModelOutputWithPast(
last_hidden_state=hidden_states,
past_key_values=next_cache,
hidden_states=all_hidden_states,
attentions=all_self_attns,
)
# Copied from transformers.models.llama.modeling_llama.LlamaModel._update_causal_mask
def _update_causal_mask(
self,
attention_mask: torch.Tensor,
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)
if isinstance(attention_mask, BlockMask):
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_static_cache = isinstance(past_key_values, StaticCache)
# 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_static_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, device = input_tensor.dtype, input_tensor.device
sequence_length = input_tensor.shape[1]
if using_static_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,
device=device,
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"]
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.llama.modeling_llama.LlamaModel._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,
device: torch.device,
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.
device (`torch.device`):
The device to place the 4D attention mask on.
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=device
)
if sequence_length != 1:
causal_mask = torch.triu(causal_mask, diagonal=1)
causal_mask *= torch.arange(target_length, device=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
# Copied from transformers.models.persimmon.modeling_persimmon.PersimmonForCausalLM with PERSIMMON->STABLELM,Persimmon->StableLm
class StableLmForCausalLM(StableLmPreTrainedModel, GenerationMixin):
_tied_weights_keys = ["lm_head.weight"]
# Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.__init__ with LLAMA->STABLELM,Llama->StableLm
def __init__(self, config):
super().__init__(config)
self.model = StableLmModel(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.get_input_embeddings
def get_input_embeddings(self):
return self.model.embed_tokens
# Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.set_input_embeddings
def set_input_embeddings(self, value):
self.model.embed_tokens = value
# Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.get_output_embeddings
def get_output_embeddings(self):
return self.lm_head
# Copied from transformers.models.llama.modeling_llama.LlamaForCausalLM.set_output_embeddings
def set_output_embeddings(self, new_embeddings):
self.lm_head = new_embeddings
# 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
@deprecate_kwarg("num_logits_to_keep", version="4.50", new_name="logits_to_keep")
@add_start_docstrings_to_model_forward(STABLELM_INPUTS_DOCSTRING)
@replace_return_docstrings(output_type=CausalLMOutputWithPast, config_class=_CONFIG_FOR_DOC)
# Ignore copy
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[List[torch.FloatTensor]] = 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]`.
logits_to_keep (`int` or `torch.Tensor`, *optional*):
If an `int`, compute logits for the last `logits_to_keep` tokens. If `0`, calculate logits for all
`input_ids` (special case). Only last token logits are needed for generation, and calculating them only for that
token can save memory, which becomes pretty significant for long sequences or large vocabulary size.
If a `torch.Tensor`, must be 1D corresponding to the indices to keep in the sequence length dimension.
This is useful when using packed tensor format (single dimension for batch and sequence length).
Returns:
Example:
```python
>>> from transformers import AutoTokenizer, StableLmForCausalLM
>>> model = StableLmForCausalLM.from_pretrained("stabilityai/stablelm-3b-4e1t")
>>> tokenizer = AutoTokenizer.from_pretrained("stabilityai/stablelm-3b-4e1t")
>>> prompt = "The weather is always wonderful in"
>>> 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]
'The weather is always wonderful in the summer in the city of San Diego. The city is located on the coast of the Pacific Ocean and is surrounded by'
```"""
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
)
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,
)
hidden_states = outputs.last_hidden_state
# No upscaling to float was ever done for StableLm
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,
)
@add_start_docstrings(
"""
The StableLm transformer with a sequence classification head on top (linear layer).
[`StableLmForSequenceClassification`] uses the last token in order to do the classification, as other causal
models (e.g. GPT-2) do.
Since it does classification on the last token, it requires to know the position of the last token. If a
`pad_token_id` is defined in the configuration, it finds the last token that is not a padding token in each row. If
no `pad_token_id` is defined, it simply takes the last value in each row of the batch. Since it cannot guess the
padding tokens when `inputs_embeds` are passed instead of `input_ids`, it does the same (take the last value in
each row of the batch).
""",
STABLELM_START_DOCSTRING,
)
# Copied from transformers.models.llama.modeling_llama.LlamaForSequenceClassification with LLAMA->STABLELM,Llama->StableLm
class StableLmForSequenceClassification(StableLmPreTrainedModel):
def __init__(self, config):
super().__init__(config)
self.num_labels = config.num_labels
self.model = StableLmModel(config)
self.score = nn.Linear(config.hidden_size, self.num_labels, bias=False)
# Initialize weights and apply final processing
self.post_init()
def get_input_embeddings(self):
return self.model.embed_tokens
def set_input_embeddings(self, value):
self.model.embed_tokens = value
@can_return_tuple
@add_start_docstrings_to_model_forward(STABLELM_INPUTS_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,
) -> SequenceClassifierOutputWithPast:
r"""
labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
`config.num_labels > 1` a classification loss is computed (Cross-Entropy).
"""
transformer_outputs: BaseModelOutputWithPast = self.model(
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,
)
hidden_states = transformer_outputs.last_hidden_state
logits = self.score(hidden_states)
if input_ids is not None:
batch_size = input_ids.shape[0]
else:
batch_size = inputs_embeds.shape[0]
if self.config.pad_token_id is None and batch_size != 1:
raise ValueError("Cannot handle batch sizes > 1 if no padding token is defined.")
if self.config.pad_token_id is None:
last_non_pad_token = -1
elif input_ids is not None:
# To handle both left- and right- padding, we take the rightmost token that is not equal to pad_token_id
non_pad_mask = (input_ids != self.config.pad_token_id).to(logits.device, torch.int32)
token_indices = torch.arange(input_ids.shape[-1], device=logits.device, dtype=torch.int32)
last_non_pad_token = (token_indices * non_pad_mask).argmax(-1)
else:
last_non_pad_token = -1
logger.warning_once(
f"{self.__class__.__name__} will not detect padding tokens in `inputs_embeds`. Results may be "
"unexpected if using padding tokens in conjunction with `inputs_embeds.`"
)
pooled_logits = logits[torch.arange(batch_size, device=logits.device), last_non_pad_token]
loss = None
if labels is not None:
loss = self.loss_function(logits=logits, labels=labels, pooled_logits=pooled_logits, config=self.config)
return SequenceClassifierOutputWithPast(
loss=loss,
logits=pooled_logits,
past_key_values=transformer_outputs.past_key_values,
hidden_states=transformer_outputs.hidden_states,
attentions=transformer_outputs.attentions,
)
@add_start_docstrings(
"""
The StableLm Model transformer with a token classification head on top (a linear layer on top of the hidden-states
output) e.g. for Named-Entity-Recognition (NER) tasks.
""",
STABLELM_START_DOCSTRING,
)
# Copied from transformers.models.llama.modeling_llama.LlamaForTokenClassification with Llama->StableLm, LLAMA->STABLELM
class StableLmForTokenClassification(StableLmPreTrainedModel):
def __init__(self, config):
super().__init__(config)
self.num_labels = config.num_labels
self.model = StableLmModel(config)
if getattr(config, "classifier_dropout", None) is not None:
classifier_dropout = config.classifier_dropout
elif getattr(config, "hidden_dropout", None) is not None:
classifier_dropout = config.hidden_dropout
else:
classifier_dropout = 0.1
self.dropout = nn.Dropout(classifier_dropout)
self.score = nn.Linear(config.hidden_size, config.num_labels)
# Initialize weights and apply final processing
self.post_init()
def get_input_embeddings(self):
return self.model.embed_tokens
def set_input_embeddings(self, value):
self.model.embed_tokens = value
@can_return_tuple
@add_start_docstrings_to_model_forward(STABLELM_INPUTS_DOCSTRING)
@add_code_sample_docstrings(
checkpoint=_CHECKPOINT_FOR_DOC,
output_type=TokenClassifierOutput,
config_class=_CONFIG_FOR_DOC,
)
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,
) -> TokenClassifierOutput:
r"""
labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
`config.num_labels > 1` a classification loss is computed (Cross-Entropy).
"""
outputs: BaseModelOutputWithPast = self.model(
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,
)
sequence_output = outputs.last_hidden_state
sequence_output = self.dropout(sequence_output)
logits = self.score(sequence_output)
loss = None
if labels is not None:
loss = self.loss_function(logits, labels, self.config)
return TokenClassifierOutput(
loss=loss,
logits=logits,
hidden_states=outputs.hidden_states,
attentions=outputs.attentions,
)
__all__ = [
"StableLmForCausalLM",
"StableLmModel",
"StableLmPreTrainedModel",
"StableLmForSequenceClassification",
"StableLmForTokenClassification",
]