Fix attention for Nvidia V100s compatibility (no FlashAttention). Based on work of puru22 for Falcon-40B
Browse files- .gitignore +211 -0
- modelling_RW.py +139 -111
.gitignore
ADDED
@@ -0,0 +1,211 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Created by https://www.toptal.com/developers/gitignore/api/python,macos
|
2 |
+
# Edit at https://www.toptal.com/developers/gitignore?templates=python,macos
|
3 |
+
|
4 |
+
### macOS ###
|
5 |
+
# General
|
6 |
+
.DS_Store
|
7 |
+
.AppleDouble
|
8 |
+
.LSOverride
|
9 |
+
|
10 |
+
# Icon must end with two \r
|
11 |
+
Icon
|
12 |
+
|
13 |
+
|
14 |
+
# Thumbnails
|
15 |
+
._*
|
16 |
+
|
17 |
+
# Files that might appear in the root of a volume
|
18 |
+
.DocumentRevisions-V100
|
19 |
+
.fseventsd
|
20 |
+
.Spotlight-V100
|
21 |
+
.TemporaryItems
|
22 |
+
.Trashes
|
23 |
+
.VolumeIcon.icns
|
24 |
+
.com.apple.timemachine.donotpresent
|
25 |
+
|
26 |
+
# Directories potentially created on remote AFP share
|
27 |
+
.AppleDB
|
28 |
+
.AppleDesktop
|
29 |
+
Network Trash Folder
|
30 |
+
Temporary Items
|
31 |
+
.apdisk
|
32 |
+
|
33 |
+
### macOS Patch ###
|
34 |
+
# iCloud generated files
|
35 |
+
*.icloud
|
36 |
+
|
37 |
+
### Python ###
|
38 |
+
# Byte-compiled / optimized / DLL files
|
39 |
+
__pycache__/
|
40 |
+
*.py[cod]
|
41 |
+
*$py.class
|
42 |
+
|
43 |
+
# C extensions
|
44 |
+
*.so
|
45 |
+
|
46 |
+
# Distribution / packaging
|
47 |
+
.Python
|
48 |
+
build/
|
49 |
+
develop-eggs/
|
50 |
+
dist/
|
51 |
+
downloads/
|
52 |
+
eggs/
|
53 |
+
.eggs/
|
54 |
+
lib/
|
55 |
+
lib64/
|
56 |
+
parts/
|
57 |
+
sdist/
|
58 |
+
var/
|
59 |
+
wheels/
|
60 |
+
share/python-wheels/
|
61 |
+
*.egg-info/
|
62 |
+
.installed.cfg
|
63 |
+
*.egg
|
64 |
+
MANIFEST
|
65 |
+
|
66 |
+
# PyInstaller
|
67 |
+
# Usually these files are written by a python script from a template
|
68 |
+
# before PyInstaller builds the exe, so as to inject date/other infos into it.
|
69 |
+
*.manifest
|
70 |
+
*.spec
|
71 |
+
|
72 |
+
# Installer logs
|
73 |
+
pip-log.txt
|
74 |
+
pip-delete-this-directory.txt
|
75 |
+
|
76 |
+
# Unit test / coverage reports
|
77 |
+
htmlcov/
|
78 |
+
.tox/
|
79 |
+
.nox/
|
80 |
+
.coverage
|
81 |
+
.coverage.*
|
82 |
+
.cache
|
83 |
+
nosetests.xml
|
84 |
+
coverage.xml
|
85 |
+
*.cover
|
86 |
+
*.py,cover
|
87 |
+
.hypothesis/
|
88 |
+
.pytest_cache/
|
89 |
+
cover/
|
90 |
+
|
91 |
+
# Translations
|
92 |
+
*.mo
|
93 |
+
*.pot
|
94 |
+
|
95 |
+
# Django stuff:
|
96 |
+
*.log
|
97 |
+
local_settings.py
|
98 |
+
db.sqlite3
|
99 |
+
db.sqlite3-journal
|
100 |
+
|
101 |
+
# Flask stuff:
|
102 |
+
instance/
|
103 |
+
.webassets-cache
|
104 |
+
|
105 |
+
# Scrapy stuff:
|
106 |
+
.scrapy
|
107 |
+
|
108 |
+
# Sphinx documentation
|
109 |
+
docs/_build/
|
110 |
+
|
111 |
+
# PyBuilder
|
112 |
+
.pybuilder/
|
113 |
+
target/
|
114 |
+
|
115 |
+
# Jupyter Notebook
|
116 |
+
.ipynb_checkpoints
|
117 |
+
|
118 |
+
# IPython
|
119 |
+
profile_default/
|
120 |
+
ipython_config.py
|
121 |
+
|
122 |
+
# pyenv
|
123 |
+
# For a library or package, you might want to ignore these files since the code is
|
124 |
+
# intended to run in multiple environments; otherwise, check them in:
|
125 |
+
# .python-version
|
126 |
+
|
127 |
+
# pipenv
|
128 |
+
# According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
|
129 |
+
# However, in case of collaboration, if having platform-specific dependencies or dependencies
|
130 |
+
# having no cross-platform support, pipenv may install dependencies that don't work, or not
|
131 |
+
# install all needed dependencies.
|
132 |
+
#Pipfile.lock
|
133 |
+
|
134 |
+
# poetry
|
135 |
+
# Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
|
136 |
+
# This is especially recommended for binary packages to ensure reproducibility, and is more
|
137 |
+
# commonly ignored for libraries.
|
138 |
+
# https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
|
139 |
+
#poetry.lock
|
140 |
+
|
141 |
+
# pdm
|
142 |
+
# Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
|
143 |
+
#pdm.lock
|
144 |
+
# pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
|
145 |
+
# in version control.
|
146 |
+
# https://pdm.fming.dev/#use-with-ide
|
147 |
+
.pdm.toml
|
148 |
+
|
149 |
+
# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
|
150 |
+
__pypackages__/
|
151 |
+
|
152 |
+
# Celery stuff
|
153 |
+
celerybeat-schedule
|
154 |
+
celerybeat.pid
|
155 |
+
|
156 |
+
# SageMath parsed files
|
157 |
+
*.sage.py
|
158 |
+
|
159 |
+
# Environments
|
160 |
+
.env
|
161 |
+
.venv
|
162 |
+
env/
|
163 |
+
venv/
|
164 |
+
ENV/
|
165 |
+
env.bak/
|
166 |
+
venv.bak/
|
167 |
+
|
168 |
+
# Spyder project settings
|
169 |
+
.spyderproject
|
170 |
+
.spyproject
|
171 |
+
|
172 |
+
# Rope project settings
|
173 |
+
.ropeproject
|
174 |
+
|
175 |
+
# mkdocs documentation
|
176 |
+
/site
|
177 |
+
|
178 |
+
# mypy
|
179 |
+
.mypy_cache/
|
180 |
+
.dmypy.json
|
181 |
+
dmypy.json
|
182 |
+
|
183 |
+
# Pyre type checker
|
184 |
+
.pyre/
|
185 |
+
|
186 |
+
# pytype static type analyzer
|
187 |
+
.pytype/
|
188 |
+
|
189 |
+
# Cython debug symbols
|
190 |
+
cython_debug/
|
191 |
+
|
192 |
+
# PyCharm
|
193 |
+
# JetBrains specific template is maintained in a separate JetBrains.gitignore that can
|
194 |
+
# be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
|
195 |
+
# and can be added to the global gitignore or merged into this file. For a more nuclear
|
196 |
+
# option (not recommended) you can uncomment the following to ignore the entire idea folder.
|
197 |
+
#.idea/
|
198 |
+
|
199 |
+
### Python Patch ###
|
200 |
+
# Poetry local configuration file - https://python-poetry.org/docs/configuration/#local-configuration
|
201 |
+
poetry.toml
|
202 |
+
|
203 |
+
# ruff
|
204 |
+
.ruff_cache/
|
205 |
+
|
206 |
+
# LSP config files
|
207 |
+
pyrightconfig.json
|
208 |
+
|
209 |
+
# End of https://www.toptal.com/developers/gitignore/api/python,macos
|
210 |
+
|
211 |
+
.idea
|
modelling_RW.py
CHANGED
@@ -25,6 +25,7 @@ from .configuration_RW import RWConfig
|
|
25 |
|
26 |
logger = logging.get_logger(__name__)
|
27 |
|
|
|
28 |
# NOTE(Hesslow): Unfortunately we did not fuse matmul and bias during training, this means that there's one additional quantization to bfloat16 between the operations.
|
29 |
# In order not to degrade the quality of our HF-port, we keep these characteristics in the final model.
|
30 |
class Linear(nn.Linear):
|
@@ -38,9 +39,10 @@ class Linear(nn.Linear):
|
|
38 |
|
39 |
from einops import rearrange
|
40 |
|
|
|
41 |
# rotary pos emb helpers (torch.jit.script does not seem to support staticmethod...)
|
42 |
def rotate_half(x):
|
43 |
-
x1, x2 = x[..., : x.shape[-1] // 2], x[..., x.shape[-1] // 2
|
44 |
return torch.cat((-x2, x1), dim=x1.ndim - 1) # dim=-1 triggers a bug in torch < 1.8.0
|
45 |
|
46 |
|
@@ -51,9 +53,9 @@ class RotaryEmbedding(torch.nn.Module):
|
|
51 |
"""
|
52 |
|
53 |
def __init__(
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
):
|
58 |
super().__init__()
|
59 |
inv_freq = 1.0 / (base ** (torch.arange(0, head_dim, 2).float() / head_dim))
|
@@ -65,10 +67,10 @@ class RotaryEmbedding(torch.nn.Module):
|
|
65 |
self.sin_cached: torch.Tensor | None = None
|
66 |
|
67 |
def cos_sin(
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
) -> torch.Tensor:
|
73 |
if seq_len != self.seq_len_cached:
|
74 |
self.seq_len_cached = seq_len
|
@@ -87,23 +89,31 @@ class RotaryEmbedding(torch.nn.Module):
|
|
87 |
|
88 |
return self.cos_cached, self.sin_cached
|
89 |
|
90 |
-
def forward(self, q, k):
|
91 |
-
|
|
|
|
|
|
|
|
|
92 |
cos, sin = self.cos_sin(seq_len, q.device, q.dtype)
|
93 |
-
|
|
|
|
|
|
|
|
|
94 |
|
95 |
|
96 |
def _make_causal_mask(
|
97 |
-
|
98 |
) -> torch.BoolTensor:
|
99 |
batch_size, target_length = input_ids_shape
|
100 |
mask = torch.empty((target_length, target_length + past_key_values_length), dtype=torch.bool, device=device)
|
101 |
# ONNX doesn't support `torch.Tensor.triu` properly, thus we use this workaround
|
102 |
seq_ids = torch.arange(target_length, device=device)
|
103 |
-
mask[:, past_key_values_length:] = seq_ids[:, None]
|
104 |
|
105 |
if past_key_values_length > 0:
|
106 |
-
mask[:, :past_key_values_length] =
|
107 |
|
108 |
expanded_mask = mask[None, None, :, :].expand(batch_size, 1, target_length, target_length + past_key_values_length)
|
109 |
return expanded_mask
|
@@ -230,14 +240,14 @@ class Attention(nn.Module):
|
|
230 |
return x.reshape(batch_size, seq_length, self.num_heads * self.head_dim)
|
231 |
|
232 |
def forward(
|
233 |
-
|
234 |
-
|
235 |
-
|
236 |
-
|
237 |
-
|
238 |
-
|
239 |
-
|
240 |
-
|
241 |
):
|
242 |
fused_qkv = self.query_key_value(hidden_states) # [batch_size, seq_length, 3 x hidden_size]
|
243 |
|
@@ -256,18 +266,27 @@ class Attention(nn.Module):
|
|
256 |
|
257 |
query_layer, key_layer = self.maybe_rotary(query_layer, key_layer)
|
258 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
259 |
if layer_past is not None:
|
260 |
past_key, past_value = layer_past
|
261 |
# concatenate along seq_length dimension:
|
262 |
# - key: [batch_size * self.num_heads, head_dim, kv_length]
|
263 |
# - value: [batch_size * self.num_heads, kv_length, head_dim]
|
|
|
264 |
key_layer = torch.cat((past_key, key_layer), dim=1)
|
265 |
value_layer = torch.cat((past_value, value_layer), dim=1)
|
266 |
|
267 |
_, kv_length, _ = key_layer.shape
|
268 |
|
269 |
if use_cache is True:
|
270 |
-
|
|
|
271 |
else:
|
272 |
present = None
|
273 |
|
@@ -276,9 +295,14 @@ class Attention(nn.Module):
|
|
276 |
key_layer_ = key_layer.reshape(batch_size, self.num_kv, -1, self.head_dim)
|
277 |
value_layer_ = value_layer.reshape(batch_size, self.num_kv, -1, self.head_dim)
|
278 |
|
279 |
-
|
280 |
-
|
281 |
-
|
|
|
|
|
|
|
|
|
|
|
282 |
|
283 |
x = attn_output.view(batch_size, self.num_heads, q_length, self.head_dim)
|
284 |
x = x.permute(0, 2, 1, 3)
|
@@ -303,7 +327,8 @@ class Attention(nn.Module):
|
|
303 |
attention_scores = attention_scores.to(torch.float32)
|
304 |
# attn_weights = torch.masked_fill(attention_scores, attention_mask, torch.finfo(attention_scores.dtype).min)
|
305 |
attention_probs = F.softmax(
|
306 |
-
(attention_scores + alibi.view(batch_size, self.num_heads, 1,
|
|
|
307 |
dim=-1,
|
308 |
dtype=hidden_states.dtype,
|
309 |
)
|
@@ -368,14 +393,14 @@ class DecoderLayer(nn.Module):
|
|
368 |
self.config = config
|
369 |
|
370 |
def forward(
|
371 |
-
|
372 |
-
|
373 |
-
|
374 |
-
|
375 |
-
|
376 |
-
|
377 |
-
|
378 |
-
|
379 |
):
|
380 |
|
381 |
layernorm_output = self.input_layernorm(hidden_states)
|
@@ -453,7 +478,7 @@ class RWPreTrainedModel(PreTrainedModel):
|
|
453 |
|
454 |
@staticmethod
|
455 |
def _convert_to_standard_cache(
|
456 |
-
|
457 |
) -> Tuple[Tuple[torch.Tensor, torch.Tensor]]:
|
458 |
"""
|
459 |
Standardizes the format of the cache so as to match most implementations, i.e. to tuple(tuple([batch_size,
|
@@ -473,7 +498,7 @@ class RWPreTrainedModel(PreTrainedModel):
|
|
473 |
|
474 |
@staticmethod
|
475 |
def _convert_to_rw_cache(
|
476 |
-
|
477 |
) -> Tuple[Tuple[torch.Tensor, torch.Tensor]]:
|
478 |
batch_size, num_heads, head_dim, seq_length = past_key_value[0][0].shape
|
479 |
batch_size_times_num_heads = batch_size * num_heads
|
@@ -514,7 +539,7 @@ class RWModel(RWPreTrainedModel):
|
|
514 |
return self.word_embeddings
|
515 |
|
516 |
def _prepare_attn_mask(
|
517 |
-
|
518 |
) -> torch.BoolTensor:
|
519 |
# create causal mask
|
520 |
# [batch_size, seq_length] -> [batch_size, 1, tgt_length, src_length]
|
@@ -522,10 +547,10 @@ class RWModel(RWPreTrainedModel):
|
|
522 |
device = attention_mask.device
|
523 |
_, src_length = input_shape
|
524 |
|
525 |
-
if src_length > 1:
|
526 |
-
|
527 |
-
|
528 |
-
|
529 |
|
530 |
# [batch_size, seq_length] -> [batch_size, 1, tgt_length, src_length]
|
531 |
expanded_attn_mask = _expand_mask(attention_mask, tgt_length=src_length)
|
@@ -539,17 +564,17 @@ class RWModel(RWPreTrainedModel):
|
|
539 |
self.word_embeddings = new_embeddings
|
540 |
|
541 |
def forward(
|
542 |
-
|
543 |
-
|
544 |
-
|
545 |
-
|
546 |
-
|
547 |
-
|
548 |
-
|
549 |
-
|
550 |
-
|
551 |
-
|
552 |
-
|
553 |
) -> Union[Tuple[torch.Tensor, ...], BaseModelOutputWithPastAndCrossAttentions]:
|
554 |
if deprecated_arguments.pop("position_ids", False) is not False:
|
555 |
# `position_ids` could have been `torch.Tensor` or `None` so defaulting pop to `False` allows to detect if users were passing explicitly `None`
|
@@ -697,40 +722,43 @@ class RWForCausalLM(RWPreTrainedModel):
|
|
697 |
self.lm_head = new_embeddings
|
698 |
|
699 |
def prepare_inputs_for_generation(
|
700 |
-
|
701 |
-
|
702 |
-
|
703 |
-
|
704 |
-
|
705 |
) -> dict:
|
706 |
# only last token for input_ids if past is not None
|
707 |
-
if
|
708 |
input_ids = input_ids[:, -1].unsqueeze(-1)
|
709 |
-
|
710 |
# the cache may be in the stardard format (e.g. in contrastive search), convert to our's format if needed
|
711 |
-
if
|
712 |
-
|
|
|
|
|
|
|
713 |
|
714 |
return {
|
715 |
"input_ids": input_ids,
|
716 |
-
"past_key_values":
|
717 |
"use_cache": kwargs.get("use_cache"),
|
718 |
"attention_mask": attention_mask,
|
719 |
}
|
720 |
|
721 |
def forward(
|
722 |
-
|
723 |
-
|
724 |
-
|
725 |
-
|
726 |
-
|
727 |
-
|
728 |
-
|
729 |
-
|
730 |
-
|
731 |
-
|
732 |
-
|
733 |
-
|
734 |
) -> Union[Tuple[torch.Tensor], CausalLMOutputWithCrossAttentions]:
|
735 |
r"""
|
736 |
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
|
@@ -790,7 +818,7 @@ class RWForCausalLM(RWPreTrainedModel):
|
|
790 |
)
|
791 |
|
792 |
def _reorder_cache(
|
793 |
-
|
794 |
) -> Tuple[Tuple[torch.Tensor, torch.Tensor], ...]:
|
795 |
"""
|
796 |
This function is used to re-order the `past_key_values` cache if [`~PreTrainedModel.beam_search`] or
|
@@ -828,18 +856,18 @@ class RWForSequenceClassification(RWPreTrainedModel):
|
|
828 |
self.post_init()
|
829 |
|
830 |
def forward(
|
831 |
-
|
832 |
-
|
833 |
-
|
834 |
-
|
835 |
-
|
836 |
-
|
837 |
-
|
838 |
-
|
839 |
-
|
840 |
-
|
841 |
-
|
842 |
-
|
843 |
) -> Union[Tuple[torch.Tensor], SequenceClassifierOutputWithPast]:
|
844 |
r"""
|
845 |
labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
|
@@ -951,18 +979,18 @@ class RWForTokenClassification(RWPreTrainedModel):
|
|
951 |
self.post_init()
|
952 |
|
953 |
def forward(
|
954 |
-
|
955 |
-
|
956 |
-
|
957 |
-
|
958 |
-
|
959 |
-
|
960 |
-
|
961 |
-
|
962 |
-
|
963 |
-
|
964 |
-
|
965 |
-
|
966 |
) -> Union[Tuple[torch.Tensor], TokenClassifierOutput]:
|
967 |
r"""
|
968 |
labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
|
@@ -1028,17 +1056,17 @@ class RWForQuestionAnswering(RWPreTrainedModel):
|
|
1028 |
self.post_init()
|
1029 |
|
1030 |
def forward(
|
1031 |
-
|
1032 |
-
|
1033 |
-
|
1034 |
-
|
1035 |
-
|
1036 |
-
|
1037 |
-
|
1038 |
-
|
1039 |
-
|
1040 |
-
|
1041 |
-
|
1042 |
) -> Union[Tuple, QuestionAnsweringModelOutput]:
|
1043 |
r"""
|
1044 |
start_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
|
|
|
25 |
|
26 |
logger = logging.get_logger(__name__)
|
27 |
|
28 |
+
|
29 |
# NOTE(Hesslow): Unfortunately we did not fuse matmul and bias during training, this means that there's one additional quantization to bfloat16 between the operations.
|
30 |
# In order not to degrade the quality of our HF-port, we keep these characteristics in the final model.
|
31 |
class Linear(nn.Linear):
|
|
|
39 |
|
40 |
from einops import rearrange
|
41 |
|
42 |
+
|
43 |
# rotary pos emb helpers (torch.jit.script does not seem to support staticmethod...)
|
44 |
def rotate_half(x):
|
45 |
+
x1, x2 = x[..., : x.shape[-1] // 2], x[..., x.shape[-1] // 2:]
|
46 |
return torch.cat((-x2, x1), dim=x1.ndim - 1) # dim=-1 triggers a bug in torch < 1.8.0
|
47 |
|
48 |
|
|
|
53 |
"""
|
54 |
|
55 |
def __init__(
|
56 |
+
self,
|
57 |
+
head_dim: int,
|
58 |
+
base=10000,
|
59 |
):
|
60 |
super().__init__()
|
61 |
inv_freq = 1.0 / (base ** (torch.arange(0, head_dim, 2).float() / head_dim))
|
|
|
67 |
self.sin_cached: torch.Tensor | None = None
|
68 |
|
69 |
def cos_sin(
|
70 |
+
self,
|
71 |
+
seq_len: int,
|
72 |
+
device="cuda",
|
73 |
+
dtype=torch.bfloat16,
|
74 |
) -> torch.Tensor:
|
75 |
if seq_len != self.seq_len_cached:
|
76 |
self.seq_len_cached = seq_len
|
|
|
89 |
|
90 |
return self.cos_cached, self.sin_cached
|
91 |
|
92 |
+
def forward(self, q, k, past_seq_length=None):
|
93 |
+
if past_seq_length is None:
|
94 |
+
batch, seq_len, head_dim = q.shape
|
95 |
+
else:
|
96 |
+
batch, input_seq_len, head_dim = q.shape
|
97 |
+
seq_len = input_seq_len + past_seq_length
|
98 |
cos, sin = self.cos_sin(seq_len, q.device, q.dtype)
|
99 |
+
if past_seq_length is not None:
|
100 |
+
return (q * cos[:, past_seq_length:, :]) + (rotate_half(q) * sin[:, past_seq_length:, :]), (
|
101 |
+
k * cos[:, past_seq_length:, :]) + (rotate_half(k) * sin[:, past_seq_length:, :])
|
102 |
+
else:
|
103 |
+
return (q * cos) + (rotate_half(q) * sin), (k * cos) + (rotate_half(k) * sin)
|
104 |
|
105 |
|
106 |
def _make_causal_mask(
|
107 |
+
input_ids_shape: torch.Size, device: torch.device, past_key_values_length: int
|
108 |
) -> torch.BoolTensor:
|
109 |
batch_size, target_length = input_ids_shape
|
110 |
mask = torch.empty((target_length, target_length + past_key_values_length), dtype=torch.bool, device=device)
|
111 |
# ONNX doesn't support `torch.Tensor.triu` properly, thus we use this workaround
|
112 |
seq_ids = torch.arange(target_length, device=device)
|
113 |
+
mask[:, past_key_values_length:] = seq_ids[:, None] >= seq_ids[None, :]
|
114 |
|
115 |
if past_key_values_length > 0:
|
116 |
+
mask[:, :past_key_values_length] = True
|
117 |
|
118 |
expanded_mask = mask[None, None, :, :].expand(batch_size, 1, target_length, target_length + past_key_values_length)
|
119 |
return expanded_mask
|
|
|
240 |
return x.reshape(batch_size, seq_length, self.num_heads * self.head_dim)
|
241 |
|
242 |
def forward(
|
243 |
+
self,
|
244 |
+
hidden_states: torch.Tensor,
|
245 |
+
alibi: torch.Tensor,
|
246 |
+
attention_mask: torch.Tensor,
|
247 |
+
layer_past: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,
|
248 |
+
head_mask: Optional[torch.Tensor] = None,
|
249 |
+
use_cache: bool = False,
|
250 |
+
output_attentions: bool = False,
|
251 |
):
|
252 |
fused_qkv = self.query_key_value(hidden_states) # [batch_size, seq_length, 3 x hidden_size]
|
253 |
|
|
|
266 |
|
267 |
query_layer, key_layer = self.maybe_rotary(query_layer, key_layer)
|
268 |
|
269 |
+
if layer_past is not None:
|
270 |
+
past_key, past_value = layer_past
|
271 |
+
past_kv_length = past_key.shape[2]
|
272 |
+
query_layer, key_layer = self.maybe_rotary(query_layer, key_layer, past_kv_length)
|
273 |
+
else:
|
274 |
+
query_layer, key_layer = self.maybe_rotary(query_layer, key_layer)
|
275 |
+
|
276 |
if layer_past is not None:
|
277 |
past_key, past_value = layer_past
|
278 |
# concatenate along seq_length dimension:
|
279 |
# - key: [batch_size * self.num_heads, head_dim, kv_length]
|
280 |
# - value: [batch_size * self.num_heads, kv_length, head_dim]
|
281 |
+
past_key = past_key.permute(0, 2, 1)
|
282 |
key_layer = torch.cat((past_key, key_layer), dim=1)
|
283 |
value_layer = torch.cat((past_value, value_layer), dim=1)
|
284 |
|
285 |
_, kv_length, _ = key_layer.shape
|
286 |
|
287 |
if use_cache is True:
|
288 |
+
key_layer_permute = key_layer.permute(0, 2, 1)
|
289 |
+
present = (key_layer_permute, value_layer)
|
290 |
else:
|
291 |
present = None
|
292 |
|
|
|
295 |
key_layer_ = key_layer.reshape(batch_size, self.num_kv, -1, self.head_dim)
|
296 |
value_layer_ = value_layer.reshape(batch_size, self.num_kv, -1, self.head_dim)
|
297 |
|
298 |
+
if attention_mask is not None:
|
299 |
+
attn_output = F.scaled_dot_product_attention(
|
300 |
+
query_layer_, key_layer_, value_layer_, attention_mask, 0.0, is_causal=False
|
301 |
+
)
|
302 |
+
else:
|
303 |
+
attn_output = F.scaled_dot_product_attention(
|
304 |
+
query_layer_, key_layer_, value_layer_, None, 0.0, is_causal=True
|
305 |
+
)
|
306 |
|
307 |
x = attn_output.view(batch_size, self.num_heads, q_length, self.head_dim)
|
308 |
x = x.permute(0, 2, 1, 3)
|
|
|
327 |
attention_scores = attention_scores.to(torch.float32)
|
328 |
# attn_weights = torch.masked_fill(attention_scores, attention_mask, torch.finfo(attention_scores.dtype).min)
|
329 |
attention_probs = F.softmax(
|
330 |
+
(attention_scores + alibi.view(batch_size, self.num_heads, 1,
|
331 |
+
-1)) * self.inv_norm_factor + attention_mask_float,
|
332 |
dim=-1,
|
333 |
dtype=hidden_states.dtype,
|
334 |
)
|
|
|
393 |
self.config = config
|
394 |
|
395 |
def forward(
|
396 |
+
self,
|
397 |
+
hidden_states: torch.Tensor,
|
398 |
+
alibi: torch.Tensor,
|
399 |
+
attention_mask: torch.Tensor,
|
400 |
+
layer_past: Optional[Tuple[torch.Tensor, torch.Tensor]] = None,
|
401 |
+
head_mask: Optional[torch.Tensor] = None,
|
402 |
+
use_cache: bool = False,
|
403 |
+
output_attentions: bool = False,
|
404 |
):
|
405 |
|
406 |
layernorm_output = self.input_layernorm(hidden_states)
|
|
|
478 |
|
479 |
@staticmethod
|
480 |
def _convert_to_standard_cache(
|
481 |
+
past_key_value: Tuple[Tuple[torch.Tensor, torch.Tensor]], batch_size: int
|
482 |
) -> Tuple[Tuple[torch.Tensor, torch.Tensor]]:
|
483 |
"""
|
484 |
Standardizes the format of the cache so as to match most implementations, i.e. to tuple(tuple([batch_size,
|
|
|
498 |
|
499 |
@staticmethod
|
500 |
def _convert_to_rw_cache(
|
501 |
+
past_key_value: Tuple[Tuple[torch.Tensor, torch.Tensor]]
|
502 |
) -> Tuple[Tuple[torch.Tensor, torch.Tensor]]:
|
503 |
batch_size, num_heads, head_dim, seq_length = past_key_value[0][0].shape
|
504 |
batch_size_times_num_heads = batch_size * num_heads
|
|
|
539 |
return self.word_embeddings
|
540 |
|
541 |
def _prepare_attn_mask(
|
542 |
+
self, attention_mask: torch.Tensor, input_shape: Tuple[int, int], past_key_values_length: int
|
543 |
) -> torch.BoolTensor:
|
544 |
# create causal mask
|
545 |
# [batch_size, seq_length] -> [batch_size, 1, tgt_length, src_length]
|
|
|
547 |
device = attention_mask.device
|
548 |
_, src_length = input_shape
|
549 |
|
550 |
+
#if src_length > 1:
|
551 |
+
combined_attention_mask = _make_causal_mask(
|
552 |
+
input_shape, device=device, past_key_values_length=past_key_values_length
|
553 |
+
)
|
554 |
|
555 |
# [batch_size, seq_length] -> [batch_size, 1, tgt_length, src_length]
|
556 |
expanded_attn_mask = _expand_mask(attention_mask, tgt_length=src_length)
|
|
|
564 |
self.word_embeddings = new_embeddings
|
565 |
|
566 |
def forward(
|
567 |
+
self,
|
568 |
+
input_ids: Optional[torch.LongTensor] = None,
|
569 |
+
past_key_values: Optional[Tuple[Tuple[torch.Tensor, torch.Tensor], ...]] = None,
|
570 |
+
attention_mask: Optional[torch.Tensor] = None,
|
571 |
+
head_mask: Optional[torch.LongTensor] = None,
|
572 |
+
inputs_embeds: Optional[torch.LongTensor] = None,
|
573 |
+
use_cache: Optional[bool] = None,
|
574 |
+
output_attentions: Optional[bool] = None,
|
575 |
+
output_hidden_states: Optional[bool] = None,
|
576 |
+
return_dict: Optional[bool] = None,
|
577 |
+
**deprecated_arguments,
|
578 |
) -> Union[Tuple[torch.Tensor, ...], BaseModelOutputWithPastAndCrossAttentions]:
|
579 |
if deprecated_arguments.pop("position_ids", False) is not False:
|
580 |
# `position_ids` could have been `torch.Tensor` or `None` so defaulting pop to `False` allows to detect if users were passing explicitly `None`
|
|
|
722 |
self.lm_head = new_embeddings
|
723 |
|
724 |
def prepare_inputs_for_generation(
|
725 |
+
self,
|
726 |
+
input_ids: torch.LongTensor,
|
727 |
+
past: Optional[torch.Tensor] = None,
|
728 |
+
attention_mask: Optional[torch.Tensor] = None,
|
729 |
+
**kwargs,
|
730 |
) -> dict:
|
731 |
# only last token for input_ids if past is not None
|
732 |
+
if kwargs.get("past_key_values", None) :
|
733 |
input_ids = input_ids[:, -1].unsqueeze(-1)
|
734 |
+
past_key_values = kwargs["past_key_values"]
|
735 |
# the cache may be in the stardard format (e.g. in contrastive search), convert to our's format if needed
|
736 |
+
# if kwargs["past_key_values"][0][0].shape[0] == input_ids.shape[0]:
|
737 |
+
# past_key_values = self._convert_to_rw_cache(kwargs["past_key_values"])
|
738 |
+
# past_key_values = kwargs["past_key_values"]
|
739 |
+
else :
|
740 |
+
past_key_values = None
|
741 |
|
742 |
return {
|
743 |
"input_ids": input_ids,
|
744 |
+
"past_key_values": past_key_values,
|
745 |
"use_cache": kwargs.get("use_cache"),
|
746 |
"attention_mask": attention_mask,
|
747 |
}
|
748 |
|
749 |
def forward(
|
750 |
+
self,
|
751 |
+
input_ids: Optional[torch.LongTensor] = None,
|
752 |
+
past_key_values: Optional[Tuple[Tuple[torch.Tensor, torch.Tensor], ...]] = None,
|
753 |
+
attention_mask: Optional[torch.Tensor] = None,
|
754 |
+
head_mask: Optional[torch.Tensor] = None,
|
755 |
+
inputs_embeds: Optional[torch.Tensor] = None,
|
756 |
+
labels: Optional[torch.Tensor] = None,
|
757 |
+
use_cache: Optional[bool] = None,
|
758 |
+
output_attentions: Optional[bool] = None,
|
759 |
+
output_hidden_states: Optional[bool] = None,
|
760 |
+
return_dict: Optional[bool] = None,
|
761 |
+
**deprecated_arguments,
|
762 |
) -> Union[Tuple[torch.Tensor], CausalLMOutputWithCrossAttentions]:
|
763 |
r"""
|
764 |
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
|
|
|
818 |
)
|
819 |
|
820 |
def _reorder_cache(
|
821 |
+
self, past: Tuple[Tuple[torch.Tensor, torch.Tensor], ...], beam_idx: torch.LongTensor
|
822 |
) -> Tuple[Tuple[torch.Tensor, torch.Tensor], ...]:
|
823 |
"""
|
824 |
This function is used to re-order the `past_key_values` cache if [`~PreTrainedModel.beam_search`] or
|
|
|
856 |
self.post_init()
|
857 |
|
858 |
def forward(
|
859 |
+
self,
|
860 |
+
input_ids: Optional[torch.LongTensor] = None,
|
861 |
+
past_key_values: Optional[Tuple[Tuple[torch.Tensor, torch.Tensor], ...]] = None,
|
862 |
+
attention_mask: Optional[torch.Tensor] = None,
|
863 |
+
head_mask: Optional[torch.Tensor] = None,
|
864 |
+
inputs_embeds: Optional[torch.Tensor] = None,
|
865 |
+
labels: Optional[torch.Tensor] = None,
|
866 |
+
use_cache: Optional[bool] = None,
|
867 |
+
output_attentions: Optional[bool] = None,
|
868 |
+
output_hidden_states: Optional[bool] = None,
|
869 |
+
return_dict: Optional[bool] = None,
|
870 |
+
**deprecated_arguments,
|
871 |
) -> Union[Tuple[torch.Tensor], SequenceClassifierOutputWithPast]:
|
872 |
r"""
|
873 |
labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
|
|
|
979 |
self.post_init()
|
980 |
|
981 |
def forward(
|
982 |
+
self,
|
983 |
+
input_ids: Optional[torch.LongTensor] = None,
|
984 |
+
past_key_values: Optional[Tuple[Tuple[torch.Tensor, torch.Tensor], ...]] = None,
|
985 |
+
attention_mask: Optional[torch.Tensor] = None,
|
986 |
+
head_mask: Optional[torch.Tensor] = None,
|
987 |
+
inputs_embeds: Optional[torch.Tensor] = None,
|
988 |
+
labels: Optional[torch.Tensor] = None,
|
989 |
+
use_cache: Optional[bool] = None,
|
990 |
+
output_attentions: Optional[bool] = None,
|
991 |
+
output_hidden_states: Optional[bool] = None,
|
992 |
+
return_dict: Optional[bool] = None,
|
993 |
+
**deprecated_arguments,
|
994 |
) -> Union[Tuple[torch.Tensor], TokenClassifierOutput]:
|
995 |
r"""
|
996 |
labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
|
|
|
1056 |
self.post_init()
|
1057 |
|
1058 |
def forward(
|
1059 |
+
self,
|
1060 |
+
input_ids: Optional[torch.LongTensor] = None,
|
1061 |
+
attention_mask: Optional[torch.FloatTensor] = None,
|
1062 |
+
position_ids: Optional[torch.LongTensor] = None,
|
1063 |
+
head_mask: Optional[torch.FloatTensor] = None,
|
1064 |
+
inputs_embeds: Optional[torch.FloatTensor] = None,
|
1065 |
+
start_positions: Optional[torch.LongTensor] = None,
|
1066 |
+
end_positions: Optional[torch.LongTensor] = None,
|
1067 |
+
output_attentions: Optional[bool] = None,
|
1068 |
+
output_hidden_states: Optional[bool] = None,
|
1069 |
+
return_dict: Optional[bool] = None,
|
1070 |
) -> Union[Tuple, QuestionAnsweringModelOutput]:
|
1071 |
r"""
|
1072 |
start_positions (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
|