Yucheng Lu commited on
Commit
b71cf75
1 Parent(s): 65b5205
added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "<pad>": 32000
3
+ }
config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "auto_map": {
6
+ "AutoModelForCausalLM": "togethercomputer/LLaMA-2-7B-32K--modeling_flash_llama.LlamaForCausalLM"
7
+ },
8
+ "bos_token_id": 1,
9
+ "eos_token_id": 2,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 4096,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 11008,
14
+ "max_position_embeddings": 32768,
15
+ "model_type": "llama",
16
+ "num_attention_heads": 32,
17
+ "num_hidden_layers": 32,
18
+ "num_key_value_heads": 32,
19
+ "pad_token_id": 0,
20
+ "pretraining_tp": 1,
21
+ "rms_norm_eps": 1e-05,
22
+ "rope_scaling": {
23
+ "factor": 8.0,
24
+ "type": "linear"
25
+ },
26
+ "tie_word_embeddings": false,
27
+ "torch_dtype": "float32",
28
+ "transformers_version": "4.31.0",
29
+ "use_cache": true,
30
+ "vocab_size": 32000
31
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.31.0"
7
+ }
modeling_flash_llama.py ADDED
@@ -0,0 +1,1011 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # coding=utf-8
2
+ # Copyright 2022 EleutherAI and the HuggingFace Inc. team. All rights reserved.
3
+ #
4
+ # This code is based on EleutherAI's GPT-NeoX library and the GPT-NeoX
5
+ # and OPT implementations in this library. It has been modified from its
6
+ # original forms to accommodate minor architectural differences compared
7
+ # to GPT-NeoX and OPT used by the Meta AI team that trained the model.
8
+ #
9
+ # Licensed under the Apache License, Version 2.0 (the "License");
10
+ # you may not use this file except in compliance with the License.
11
+ # You may obtain a copy of the License at
12
+ #
13
+ # http://www.apache.org/licenses/LICENSE-2.0
14
+ #
15
+ # Unless required by applicable law or agreed to in writing, software
16
+ # distributed under the License is distributed on an "AS IS" BASIS,
17
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
18
+ # See the License for the specific language governing permissions and
19
+ # limitations under the License.
20
+ """ PyTorch LLaMA model."""
21
+ import math
22
+ from typing import List, Optional, Tuple, Union
23
+
24
+ import torch
25
+ import torch.nn.functional as F
26
+ import torch.utils.checkpoint
27
+ from torch import nn
28
+ from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
29
+
30
+ from transformers.activations import ACT2FN
31
+ from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast, SequenceClassifierOutputWithPast
32
+ from transformers.modeling_utils import PreTrainedModel
33
+ from transformers.utils import add_start_docstrings, add_start_docstrings_to_model_forward, logging, replace_return_docstrings
34
+ from transformers.models.llama.configuration_llama import LlamaConfig
35
+
36
+
37
+ try:
38
+ from flash_attn.flash_attn_interface import (
39
+ flash_attn_func,
40
+ flash_attn_kvpacked_func,
41
+ flash_attn_qkvpacked_func,
42
+ flash_attn_varlen_kvpacked_func,
43
+ )
44
+ from flash_attn.bert_padding import unpad_input, pad_input
45
+ flash_attn_v2_installed = True
46
+ print('>>>> Flash Attention installed')
47
+ except ImportError:
48
+ flash_attn_v2_installed = False
49
+ raise ImportError('Please install Flash Attention: `pip install flash-attn --no-build-isolation`')
50
+
51
+ try:
52
+ from flash_attn.layers.rotary import apply_rotary_emb_func
53
+ flash_rope_installed = True
54
+ print('>>>> Flash RoPE installed')
55
+ except ImportError:
56
+ flash_rope_installed = False
57
+ raise ImportError('Please install RoPE kernels: `pip install git+https://github.com/HazyResearch/flash-attention.git#subdirectory=csrc/rotary`')
58
+
59
+
60
+ logger = logging.get_logger(__name__)
61
+
62
+ _CONFIG_FOR_DOC = "LlamaConfig"
63
+
64
+
65
+ # @torch.jit.script
66
+ def rmsnorm_func(hidden_states, weight, variance_epsilon):
67
+ input_dtype = hidden_states.dtype
68
+ hidden_states = hidden_states.to(torch.float32)
69
+ variance = hidden_states.pow(2).mean(-1, keepdim=True)
70
+ hidden_states = hidden_states * torch.rsqrt(variance + variance_epsilon)
71
+ return (weight * hidden_states).to(input_dtype)
72
+
73
+
74
+ class LlamaRMSNorm(nn.Module):
75
+ def __init__(self, hidden_size, eps=1e-6):
76
+ """
77
+ LlamaRMSNorm is equivalent to T5LayerNorm
78
+ """
79
+ super().__init__()
80
+ self.weight = nn.Parameter(torch.ones(hidden_size))
81
+ self.register_buffer(
82
+ "variance_epsilon",
83
+ torch.tensor(eps),
84
+ persistent=False,
85
+ )
86
+
87
+ def forward(self, hidden_states):
88
+ return rmsnorm_func(hidden_states, self.weight, self.variance_epsilon)
89
+
90
+
91
+ class FlashRotaryEmbedding(torch.nn.Module):
92
+ """
93
+ The rotary position embeddings from RoFormer_ (Su et. al).
94
+ A crucial insight from the method is that the query and keys are
95
+ transformed by rotation matrices which depend on the relative positions.
96
+
97
+ Other implementations are available in the Rotary Transformer repo_ and in
98
+ GPT-NeoX_, GPT-NeoX was an inspiration
99
+
100
+ .. _RoFormer: https://arxiv.org/abs/2104.09864
101
+ .. _repo: https://github.com/ZhuiyiTechnology/roformer
102
+ .. _GPT-NeoX: https://github.com/EleutherAI/gpt-neox
103
+
104
+ If scale_base is not None, this implements XPos (Sun et al., https://arxiv.org/abs/2212.10554).
105
+ A recommended value for scale_base is 512: https://github.com/HazyResearch/flash-attention/issues/96
106
+ Reference: https://github.com/sunyt32/torchscale/blob/main/torchscale/component/xpos_relative_position.py
107
+ """
108
+
109
+ def __init__(self, dim: int, base=10000.0, interleaved=False, scale_base=None,
110
+ scaling_factor=1.0, pos_idx_in_fp32=True, device=None):
111
+ """
112
+ interleaved: if True, rotate pairs of even and odd dimensions (GPT-J style) instead
113
+ of 1st half and 2nd half (GPT-NeoX style).
114
+ pos_idx_in_fp32: if True, the position indices [0.0, ..., seqlen - 1] are in fp32,
115
+ otherwise they might be in lower precision.
116
+ This option was added because previously (before 2023-07-02), when we construct
117
+ the position indices, we use the dtype of self.inv_freq. In most cases this would
118
+ be fp32, but if the model is trained in pure bf16 (not mixed precision), then
119
+ self.inv_freq would be bf16, and the position indices are also in bf16.
120
+ Because of the limited precision of bf16 (e.g. 1995.0 is rounded to 2000.0), the
121
+ embeddings for some positions will coincide.
122
+ To maintain compatibility with models previously trained in pure bf16,
123
+ we add this option.
124
+ scaling_factor: RotaryEmbedding extended with linear scaling.
125
+ """
126
+ super().__init__()
127
+ self.dim = dim
128
+ self.base = float(base)
129
+ self.pos_idx_in_fp32 = pos_idx_in_fp32
130
+ # Generate and save the inverse frequency buffer (non trainable)
131
+ inv_freq = self._compute_inv_freq(device)
132
+ self.register_buffer("inv_freq", inv_freq, persistent=False)
133
+ self.interleaved = interleaved
134
+ self.scale_base = scale_base
135
+ self.scaling_factor = scaling_factor
136
+ scale = ((torch.arange(0, dim, 2, device=device, dtype=torch.float32) + 0.4 * dim)
137
+ / (1.4 * dim) if scale_base is not None else None)
138
+ self.register_buffer("scale", scale)
139
+
140
+ self._seq_len_cached = 0
141
+ self._cos_cached = None
142
+ self._sin_cached = None
143
+ self._cos_k_cached = None
144
+ self._sin_k_cached = None
145
+
146
+ def _compute_inv_freq(self, device=None):
147
+ return 1 / (self.base ** (torch.arange(0, self.dim, 2, device=device,
148
+ dtype=torch.float32) / self.dim))
149
+
150
+
151
+ def _update_cos_sin_cache(self, seqlen, device=None, dtype=None):
152
+ # Reset the tables if the sequence length has changed,
153
+ # if we're on a new device (possibly due to tracing for instance),
154
+ # or if we're switching from inference mode to training
155
+ if (seqlen > self._seq_len_cached or self._cos_cached.device != device
156
+ or self._cos_cached.dtype != dtype
157
+ or (self.training and self._cos_cached.is_inference())):
158
+ self._seq_len_cached = seqlen
159
+ # We want fp32 here, not self.inv_freq.dtype, since the model could be loaded in bf16
160
+ # And the output of arange can be quite large, so bf16 would lose a lot of precision.
161
+ # However, for compatibility reason, we add an option to use the dtype of self.inv_freq.
162
+ if self.pos_idx_in_fp32:
163
+ t = torch.arange(seqlen, device=device, dtype=torch.float32)
164
+ t /= self.scaling_factor
165
+ # We want fp32 here as well since inv_freq will be multiplied with t, and the output
166
+ # will be large. Having it in bf16 will lose a lot of precision and cause the
167
+ # cos & sin output to change significantly.
168
+ # We want to recompute self.inv_freq if it was not loaded in fp32
169
+ if self.inv_freq.dtype != torch.float32:
170
+ inv_freq = self.inv_freq.to(torch.float32)
171
+ else:
172
+ inv_freq = self.inv_freq
173
+ else:
174
+ t = torch.arange(seqlen, device=device, dtype=self.inv_freq.dtype)
175
+ t /= self.scaling_factor
176
+ inv_freq = self.inv_freq
177
+ # Don't do einsum, it converts fp32 to fp16 under AMP
178
+ # freqs = torch.einsum("i,j->ij", t, self.inv_freq)
179
+ freqs = torch.outer(t, inv_freq)
180
+ if self.scale is None:
181
+ self._cos_cached = torch.cos(freqs).to(dtype)
182
+ self._sin_cached = torch.sin(freqs).to(dtype)
183
+ else:
184
+ power = ((torch.arange(seqlen, dtype=self.scale.dtype, device=self.scale.device)
185
+ - seqlen // 2) / self.scale_base)
186
+ scale = self.scale.to(device=power.device) ** power.unsqueeze(-1)
187
+ # We want the multiplication by scale to happen in fp32
188
+ self._cos_cached = (torch.cos(freqs) * scale).to(dtype)
189
+ self._sin_cached = (torch.sin(freqs) * scale).to(dtype)
190
+ self._cos_k_cached = (torch.cos(freqs) / scale).to(dtype)
191
+ self._sin_k_cached = (torch.sin(freqs) / scale).to(dtype)
192
+
193
+ def forward(self, q: torch.Tensor, k: torch.Tensor, seqlen_offset: int = 0) -> Tuple[torch.Tensor, torch.Tensor]:
194
+ """
195
+ q: (batch, seqlen, nheads, headdim)
196
+ k: (batch, seqlen, nheads, headdim)
197
+ seqlen_offset: can be used in generation where the qkv being passed in is only the last
198
+ token in the batch.
199
+ """
200
+ self._update_cos_sin_cache(q.shape[1] + seqlen_offset, device=q.device, dtype=q.dtype)
201
+ if self.scale is None:
202
+ return apply_rotary_emb_func(
203
+ q, self._cos_cached[seqlen_offset:], self._sin_cached[seqlen_offset:],
204
+ self.interleaved, True # inplace=True
205
+ ), apply_rotary_emb_func(
206
+ k, self._cos_cached[seqlen_offset:], self._sin_cached[seqlen_offset:],
207
+ self.interleaved, True # inplace=True
208
+ )
209
+ else:
210
+ assert False
211
+
212
+ class LlamaMLP(nn.Module):
213
+ def __init__(self, config):
214
+ super().__init__()
215
+ self.config = config
216
+ self.hidden_size = config.hidden_size
217
+ self.intermediate_size = config.intermediate_size
218
+ self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
219
+ self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=False)
220
+ self.down_proj = nn.Linear(self.intermediate_size, self.hidden_size, bias=False)
221
+ self.act_fn = ACT2FN[config.hidden_act]
222
+
223
+ def forward(self, x):
224
+ if self.config.pretraining_tp > 1:
225
+ slice = self.intermediate_size // self.config.pretraining_tp
226
+ gate_proj_slices = self.gate_proj.weight.split(slice, dim=0)
227
+ up_proj_slices = self.up_proj.weight.split(slice, dim=0)
228
+ down_proj_slices = self.down_proj.weight.split(slice, dim=1)
229
+
230
+ gate_proj = torch.cat(
231
+ [F.linear(x, gate_proj_slices[i]) for i in range(self.config.pretraining_tp)], dim=-1
232
+ )
233
+ up_proj = torch.cat([F.linear(x, up_proj_slices[i]) for i in range(self.config.pretraining_tp)], dim=-1)
234
+
235
+ intermediate_states = (self.act_fn(gate_proj) * up_proj).split(slice, dim=2)
236
+ down_proj = [
237
+ F.linear(intermediate_states[i], down_proj_slices[i]) for i in range(self.config.pretraining_tp)
238
+ ]
239
+ down_proj = sum(down_proj)
240
+ else:
241
+ down_proj = self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x))
242
+
243
+ return down_proj
244
+
245
+ @torch.jit.script
246
+ def repeat_kv(hidden_states: torch.Tensor, n_rep: int) -> torch.Tensor:
247
+ """
248
+ This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch,
249
+ num_key_value_heads, seqlen, head_dim) to (batch, num_attention_heads, seqlen, head_dim)
250
+ """
251
+ batch, slen, _, num_key_value_heads, head_dim = hidden_states.shape
252
+ if n_rep == 1:
253
+ return hidden_states
254
+ hidden_states = hidden_states[:, :, :, :, None, :].expand(batch, slen, 2, num_key_value_heads, n_rep, head_dim)
255
+ return hidden_states.reshape(batch, slen, 2, num_key_value_heads * n_rep, head_dim)
256
+
257
+
258
+ class LlamaAttention(nn.Module):
259
+ """Multi-headed attention from 'Attention Is All You Need' paper"""
260
+
261
+ def __init__(self, config: LlamaConfig):
262
+ super().__init__()
263
+ self.config = config
264
+ self.hidden_size = config.hidden_size
265
+ self.num_heads = config.num_attention_heads
266
+ self.head_dim = self.hidden_size // self.num_heads
267
+ self.num_key_value_heads = config.num_key_value_heads
268
+ self.num_key_value_groups = self.num_heads // self.num_key_value_heads
269
+ self.max_position_embeddings = config.max_position_embeddings
270
+
271
+ if (self.head_dim * self.num_heads) != self.hidden_size:
272
+ raise ValueError(
273
+ f"hidden_size must be divisible by num_heads (got `hidden_size`: {self.hidden_size}"
274
+ f" and `num_heads`: {self.num_heads})."
275
+ )
276
+ self.q_proj = nn.Linear(self.hidden_size, self.num_heads * self.head_dim, bias=False)
277
+ self.k_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
278
+ self.v_proj = nn.Linear(self.hidden_size, self.num_key_value_heads * self.head_dim, bias=False)
279
+ self.o_proj = nn.Linear(self.num_heads * self.head_dim, self.hidden_size, bias=False)
280
+
281
+ self.register_buffer(
282
+ "norm_factor",
283
+ torch.sqrt(torch.tensor(self.head_dim, dtype=torch.float32)).to(torch.get_default_dtype()),
284
+ persistent=False,
285
+ )
286
+
287
+ if self.config.rope_scaling is None:
288
+ scaling_factor = 1
289
+ else:
290
+ scaling_type = self.config.rope_scaling["type"]
291
+ scaling_factor = self.config.rope_scaling["factor"]
292
+ assert scaling_type == 'linear'
293
+
294
+ self.rotary_emb = FlashRotaryEmbedding(
295
+ self.head_dim, base=10000, interleaved=False, scaling_factor=scaling_factor,
296
+ )
297
+
298
+ def _shape(self, tensor: torch.Tensor, seq_len: int, bsz: int):
299
+ return tensor.view(bsz, seq_len, self.num_heads, self.head_dim).transpose(1, 2).contiguous()
300
+
301
+ def forward(
302
+ self,
303
+ hidden_states: torch.Tensor,
304
+ attention_mask: Optional[torch.Tensor] = None,
305
+ position_ids: Optional[torch.LongTensor] = None,
306
+ past_key_value: Optional[Tuple[torch.Tensor]] = None,
307
+ output_attentions: bool = False,
308
+ use_cache: bool = False,
309
+ is_padded_inputs: Optional[bool] = False,
310
+ ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
311
+ bsz, q_len, h_size = hidden_states.size()
312
+
313
+ has_layer_past = past_key_value is not None
314
+
315
+ if has_layer_past:
316
+ past_kv = past_key_value[0]
317
+ past_len = past_key_value[1]
318
+ else:
319
+ past_len = 0
320
+
321
+ if self.config.pretraining_tp > 1:
322
+ key_value_slicing = (self.num_key_value_heads * self.head_dim) // self.config.pretraining_tp
323
+ query_slices = self.q_proj.weight.split(
324
+ (self.num_heads * self.head_dim) // self.config.pretraining_tp, dim=0
325
+ )
326
+ key_slices = self.k_proj.weight.split(key_value_slicing, dim=0)
327
+ value_slices = self.v_proj.weight.split(key_value_slicing, dim=0)
328
+
329
+ q = [F.linear(hidden_states, query_slices[i]) for i in range(self.config.pretraining_tp)]
330
+ q = torch.cat(q, dim=-1)
331
+
332
+ k = [F.linear(hidden_states, key_slices[i]) for i in range(self.config.pretraining_tp)]
333
+ k = torch.cat(k, dim=-1)
334
+
335
+ v = [F.linear(hidden_states, value_slices[i]) for i in range(self.config.pretraining_tp)]
336
+ v = torch.cat(v, dim=-1)
337
+
338
+ else:
339
+ q = self.q_proj(hidden_states)
340
+ k = self.k_proj(hidden_states)
341
+ v = self.v_proj(hidden_states)
342
+
343
+ q = q.view(bsz, q_len, self.num_heads, self.head_dim)
344
+ k = k.view(bsz, q_len, self.num_key_value_heads, self.head_dim)
345
+ v = v.view(bsz, q_len, self.num_key_value_heads, self.head_dim)
346
+
347
+ q, k = self.rotary_emb(q, k, past_len)
348
+
349
+ kv = torch.stack([k, v], 2)
350
+ kv = repeat_kv(kv, self.num_key_value_groups)
351
+
352
+ # Cache QKV values
353
+ if has_layer_past:
354
+ new_len = past_len+q.size(1)
355
+ if new_len > past_kv.size(1):
356
+ past_kv = torch.cat([past_kv, torch.empty(bsz, 256, 2, kv.size(3), kv.size(4), dtype=kv.dtype, device=kv.device)], 1)
357
+ past_kv[:, past_len:new_len] = kv
358
+ kv = past_kv[:, :new_len]
359
+ else:
360
+ past_kv = kv
361
+
362
+ past_key_value = (past_kv, past_len+q.size(1)) if use_cache else None
363
+
364
+ if is_padded_inputs:
365
+
366
+ # varlen, ignore padding tokens, efficient for large batch with many paddings
367
+
368
+ assert attention_mask is not None
369
+
370
+ unpadded_kv, indices_k, cu_seqlens_k, max_seqlen_k = unpad_input(kv, attention_mask)
371
+ unpadded_q, indices_q, cu_seqlens_q, max_seqlen_q = unpad_input(q, attention_mask[:, -q.size(1):])
372
+ attn_outputs = flash_attn_varlen_kvpacked_func(
373
+ unpadded_q, unpadded_kv, cu_seqlens_q, cu_seqlens_k,
374
+ max_seqlen_q, max_seqlen_k,
375
+ dropout_p=0.0, softmax_scale=1.0/self.norm_factor,
376
+ causal=(not has_layer_past), return_attn_probs=output_attentions
377
+ )
378
+
379
+ attn_output = attn_outputs[0] if output_attentions else attn_outputs
380
+ attn_output = pad_input(
381
+ attn_output, indices_q, bsz, max_seqlen_q
382
+ ).reshape(bsz, q_len, h_size)
383
+ attn_weights = attn_outputs[2] if output_attentions else None
384
+
385
+ else:
386
+
387
+ # no padding tokens, more efficient
388
+
389
+ attn_outputs = flash_attn_kvpacked_func(
390
+ q, kv, dropout_p=0.0, softmax_scale=1.0/self.norm_factor, causal=(not has_layer_past), return_attn_probs=output_attentions)
391
+
392
+ attn_output = attn_outputs[0] if output_attentions else attn_outputs
393
+ attn_output = attn_output.reshape(bsz, q_len, h_size)
394
+ attn_weights = attn_outputs[2] if output_attentions else None
395
+
396
+ if self.config.pretraining_tp > 1:
397
+ attn_output = attn_output.split(self.hidden_size // self.config.pretraining_tp, dim=2)
398
+ o_proj_slices = self.o_proj.weight.split(self.hidden_size // self.config.pretraining_tp, dim=1)
399
+ attn_output = sum([F.linear(attn_output[i], o_proj_slices[i]) for i in range(self.config.pretraining_tp)])
400
+ else:
401
+ attn_output = self.o_proj(attn_output)
402
+
403
+ if not output_attentions:
404
+ attn_weights = None
405
+
406
+ return attn_output, attn_weights, past_key_value
407
+
408
+
409
+ class LlamaDecoderLayer(nn.Module):
410
+ def __init__(self, config: LlamaConfig):
411
+ super().__init__()
412
+ self.hidden_size = config.hidden_size
413
+ self.self_attn = LlamaAttention(config=config)
414
+ self.mlp = LlamaMLP(config)
415
+ self.input_layernorm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
416
+ self.post_attention_layernorm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
417
+
418
+ def forward(
419
+ self,
420
+ hidden_states: torch.Tensor,
421
+ attention_mask: Optional[torch.Tensor] = None,
422
+ position_ids: Optional[torch.LongTensor] = None,
423
+ past_key_value: Optional[Tuple[torch.Tensor]] = None,
424
+ is_padded_inputs: Optional[bool] = False,
425
+ output_attentions: Optional[bool] = False,
426
+ use_cache: Optional[bool] = False,
427
+ ) -> Tuple[torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]]:
428
+ """
429
+ Args:
430
+ hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)`
431
+ attention_mask (`torch.FloatTensor`, *optional*): attention mask of size
432
+ `(batch, 1, tgt_len, src_len)` where padding elements are indicated by very large negative values.
433
+ output_attentions (`bool`, *optional*):
434
+ Whether or not to return the attentions tensors of all attention layers. See `attentions` under
435
+ returned tensors for more detail.
436
+ use_cache (`bool`, *optional*):
437
+ If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding
438
+ (see `past_key_values`).
439
+ past_key_value (`Tuple(torch.FloatTensor)`, *optional*): cached past key and value projection states
440
+ """
441
+
442
+ residual = hidden_states
443
+
444
+ hidden_states = self.input_layernorm(hidden_states)
445
+
446
+ # Self Attention
447
+ hidden_states, self_attn_weights, present_key_value = self.self_attn(
448
+ hidden_states=hidden_states,
449
+ attention_mask=attention_mask,
450
+ position_ids=position_ids,
451
+ past_key_value=past_key_value,
452
+ output_attentions=output_attentions,
453
+ use_cache=use_cache,
454
+ is_padded_inputs=is_padded_inputs,
455
+ )
456
+ hidden_states = residual + hidden_states
457
+
458
+ # Fully Connected
459
+ residual = hidden_states
460
+ hidden_states = self.post_attention_layernorm(hidden_states)
461
+ hidden_states = self.mlp(hidden_states)
462
+ hidden_states = residual + hidden_states
463
+
464
+ outputs = (hidden_states,)
465
+
466
+ if output_attentions:
467
+ outputs += (self_attn_weights,)
468
+
469
+ if use_cache:
470
+ outputs += (present_key_value,)
471
+
472
+ return outputs
473
+
474
+
475
+ LLAMA_START_DOCSTRING = r"""
476
+ This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the
477
+ library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
478
+ etc.)
479
+
480
+ This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
481
+ Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
482
+ and behavior.
483
+
484
+ Parameters:
485
+ config ([`LlamaConfig`]):
486
+ Model configuration class with all the parameters of the model. Initializing with a config file does not
487
+ load the weights associated with the model, only the configuration. Check out the
488
+ [`~PreTrainedModel.from_pretrained`] method to load the model weights.
489
+ """
490
+
491
+
492
+ @add_start_docstrings(
493
+ "The bare LLaMA Model outputting raw hidden-states without any specific head on top.",
494
+ LLAMA_START_DOCSTRING,
495
+ )
496
+ class LlamaPreTrainedModel(PreTrainedModel):
497
+ config_class = LlamaConfig
498
+ base_model_prefix = "model"
499
+ supports_gradient_checkpointing = True
500
+ _no_split_modules = ["LlamaDecoderLayer"]
501
+ _skip_keys_device_placement = "past_key_values"
502
+
503
+ def _init_weights(self, module):
504
+ std = self.config.initializer_range
505
+ if isinstance(module, nn.Linear):
506
+ module.weight.data.normal_(mean=0.0, std=std)
507
+ if module.bias is not None:
508
+ module.bias.data.zero_()
509
+ elif isinstance(module, nn.Embedding):
510
+ module.weight.data.normal_(mean=0.0, std=std)
511
+ if module.padding_idx is not None:
512
+ module.weight.data[module.padding_idx].zero_()
513
+
514
+ def _set_gradient_checkpointing(self, module, value=False):
515
+ if isinstance(module, LlamaModel):
516
+ module.gradient_checkpointing = value
517
+
518
+
519
+ LLAMA_INPUTS_DOCSTRING = r"""
520
+ Args:
521
+ input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
522
+ Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
523
+ it.
524
+
525
+ Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
526
+ [`PreTrainedTokenizer.__call__`] for details.
527
+
528
+ [What are input IDs?](../glossary#input-ids)
529
+ attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
530
+ Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
531
+
532
+ - 1 for tokens that are **not masked**,
533
+ - 0 for tokens that are **masked**.
534
+
535
+ [What are attention masks?](../glossary#attention-mask)
536
+
537
+ Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
538
+ [`PreTrainedTokenizer.__call__`] for details.
539
+
540
+ If `past_key_values` is used, optionally only the last `decoder_input_ids` have to be input (see
541
+ `past_key_values`).
542
+
543
+ If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`]
544
+ and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more
545
+ information on the default strategy.
546
+
547
+ - 1 indicates the head is **not masked**,
548
+ - 0 indicates the head is **masked**.
549
+ position_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
550
+ Indices of positions of each input sequence tokens in the position embeddings. Selected in the range `[0,
551
+ config.n_positions - 1]`.
552
+
553
+ [What are position IDs?](../glossary#position-ids)
554
+ past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
555
+ Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of shape
556
+ `(batch_size, num_heads, sequence_length, embed_size_per_head)`) and 2 additional tensors of shape
557
+ `(batch_size, num_heads, encoder_sequence_length, embed_size_per_head)`.
558
+
559
+ Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention
560
+ blocks) that can be used (see `past_key_values` input) to speed up sequential decoding.
561
+
562
+ If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those that
563
+ don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of all
564
+ `decoder_input_ids` of shape `(batch_size, sequence_length)`.
565
+ inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
566
+ Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
567
+ is useful if you want more control over how to convert `input_ids` indices into associated vectors than the
568
+ model's internal embedding lookup matrix.
569
+ use_cache (`bool`, *optional*):
570
+ If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see
571
+ `past_key_values`).
572
+ output_attentions (`bool`, *optional*):
573
+ Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
574
+ tensors for more detail.
575
+ output_hidden_states (`bool`, *optional*):
576
+ Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
577
+ more detail.
578
+ return_dict (`bool`, *optional*):
579
+ Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
580
+ """
581
+
582
+
583
+ @add_start_docstrings(
584
+ "The bare LLaMA Model outputting raw hidden-states without any specific head on top.",
585
+ LLAMA_START_DOCSTRING,
586
+ )
587
+ class LlamaModel(LlamaPreTrainedModel):
588
+ """
589
+ Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`LlamaDecoderLayer`]
590
+
591
+ Args:
592
+ config: LlamaConfig
593
+ """
594
+
595
+ def __init__(self, config: LlamaConfig):
596
+ super().__init__(config)
597
+ self.padding_idx = config.pad_token_id
598
+ self.vocab_size = config.vocab_size
599
+
600
+ self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
601
+ self.layers = nn.ModuleList([LlamaDecoderLayer(config) for _ in range(config.num_hidden_layers)])
602
+ self.norm = LlamaRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
603
+
604
+ self.gradient_checkpointing = False
605
+ # Initialize weights and apply final processing
606
+ self.post_init()
607
+
608
+ def get_input_embeddings(self):
609
+ return self.embed_tokens
610
+
611
+ def set_input_embeddings(self, value):
612
+ self.embed_tokens = value
613
+
614
+ @add_start_docstrings_to_model_forward(LLAMA_INPUTS_DOCSTRING)
615
+ def forward(
616
+ self,
617
+ input_ids: torch.LongTensor = None,
618
+ attention_mask: Optional[torch.Tensor] = None,
619
+ position_ids: Optional[torch.LongTensor] = None,
620
+ past_key_values: Optional[List[torch.FloatTensor]] = None,
621
+ inputs_embeds: Optional[torch.FloatTensor] = None,
622
+ use_cache: Optional[bool] = None,
623
+ output_attentions: Optional[bool] = None,
624
+ output_hidden_states: Optional[bool] = None,
625
+ return_dict: Optional[bool] = None,
626
+ is_padded_inputs: Optional[bool] = False,
627
+ ) -> Union[Tuple, BaseModelOutputWithPast]:
628
+ output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
629
+ output_hidden_states = (
630
+ output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
631
+ )
632
+ use_cache = use_cache if use_cache is not None else self.config.use_cache
633
+
634
+ return_dict = return_dict if return_dict is not None else self.config.use_return_dict
635
+
636
+ # retrieve input_ids and inputs_embeds
637
+ if input_ids is not None and inputs_embeds is not None:
638
+ raise ValueError("You cannot specify both decoder_input_ids and decoder_inputs_embeds at the same time")
639
+ elif input_ids is not None:
640
+ batch_size, seq_length = input_ids.shape
641
+ elif inputs_embeds is not None:
642
+ batch_size, seq_length, _ = inputs_embeds.shape
643
+ else:
644
+ raise ValueError("You have to specify either decoder_input_ids or decoder_inputs_embeds")
645
+
646
+ seq_length_with_past = seq_length
647
+ past_key_values_length = 0
648
+
649
+ if past_key_values is not None:
650
+ past_key_values_length = past_key_values[0][0].shape[2]
651
+ seq_length_with_past = seq_length_with_past + past_key_values_length
652
+
653
+ position_ids = None
654
+
655
+ if inputs_embeds is None:
656
+ inputs_embeds = self.embed_tokens(input_ids)
657
+
658
+ hidden_states = inputs_embeds
659
+
660
+ if self.gradient_checkpointing and self.training:
661
+ if use_cache:
662
+ logger.warning_once(
663
+ "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
664
+ )
665
+ use_cache = False
666
+
667
+ # decoder layers
668
+ all_hidden_states = () if output_hidden_states else None
669
+ all_self_attns = () if output_attentions else None
670
+ next_decoder_cache = () if use_cache else None
671
+
672
+ for idx, decoder_layer in enumerate(self.layers):
673
+ if output_hidden_states:
674
+ all_hidden_states += (hidden_states,)
675
+
676
+ past_key_value = past_key_values[idx] if past_key_values is not None else None
677
+
678
+ if self.gradient_checkpointing and self.training:
679
+
680
+ def create_custom_forward(module):
681
+ def custom_forward(*inputs):
682
+ # None for past_key_value
683
+ return module(*inputs, output_attentions, None)
684
+
685
+ return custom_forward
686
+
687
+ layer_outputs = torch.utils.checkpoint.checkpoint(
688
+ create_custom_forward(decoder_layer),
689
+ hidden_states,
690
+ attention_mask,
691
+ position_ids,
692
+ None,
693
+ is_padded_inputs
694
+ )
695
+ else:
696
+ layer_outputs = decoder_layer(
697
+ hidden_states,
698
+ attention_mask=attention_mask,
699
+ position_ids=position_ids,
700
+ past_key_value=past_key_value,
701
+ output_attentions=output_attentions,
702
+ use_cache=use_cache,
703
+ is_padded_inputs=is_padded_inputs,
704
+ )
705
+
706
+ hidden_states = layer_outputs[0]
707
+
708
+ if use_cache:
709
+ next_decoder_cache += (layer_outputs[2 if output_attentions else 1],)
710
+
711
+ if output_attentions:
712
+ all_self_attns += (layer_outputs[1],)
713
+
714
+ hidden_states = self.norm(hidden_states)
715
+
716
+ # add hidden states from the last decoder layer
717
+ if output_hidden_states:
718
+ all_hidden_states += (hidden_states,)
719
+
720
+ next_cache = next_decoder_cache if use_cache else None
721
+ if not return_dict:
722
+ return tuple(v for v in [hidden_states, next_cache, all_hidden_states, all_self_attns] if v is not None)
723
+ return BaseModelOutputWithPast(
724
+ last_hidden_state=hidden_states,
725
+ past_key_values=next_cache,
726
+ hidden_states=all_hidden_states,
727
+ attentions=all_self_attns,
728
+ )
729
+
730
+
731
+ class LlamaForCausalLM(LlamaPreTrainedModel):
732
+ _tied_weights_keys = ["lm_head.weight"]
733
+
734
+ def __init__(self, config):
735
+ super().__init__(config)
736
+ self.model = LlamaModel(config)
737
+ self.vocab_size = config.vocab_size
738
+ self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
739
+
740
+ # Initialize weights and apply final processing
741
+ self.post_init()
742
+
743
+ def get_input_embeddings(self):
744
+ return self.model.embed_tokens
745
+
746
+ def set_input_embeddings(self, value):
747
+ self.model.embed_tokens = value
748
+
749
+ def get_output_embeddings(self):
750
+ return self.lm_head
751
+
752
+ def set_output_embeddings(self, new_embeddings):
753
+ self.lm_head = new_embeddings
754
+
755
+ def set_decoder(self, decoder):
756
+ self.model = decoder
757
+
758
+ def get_decoder(self):
759
+ return self.model
760
+
761
+ @add_start_docstrings_to_model_forward(LLAMA_INPUTS_DOCSTRING)
762
+ @replace_return_docstrings(output_type=CausalLMOutputWithPast, config_class=_CONFIG_FOR_DOC)
763
+ def forward(
764
+ self,
765
+ input_ids: torch.LongTensor = None,
766
+ attention_mask: Optional[torch.Tensor] = None,
767
+ position_ids: Optional[torch.LongTensor] = None,
768
+ past_key_values: Optional[List[torch.FloatTensor]] = None,
769
+ inputs_embeds: Optional[torch.FloatTensor] = None,
770
+ labels: Optional[torch.LongTensor] = None,
771
+ use_cache: Optional[bool] = None,
772
+ output_attentions: Optional[bool] = None,
773
+ output_hidden_states: Optional[bool] = None,
774
+ return_dict: Optional[bool] = None,
775
+ is_padded_inputs: Optional[bool] = None,
776
+ ) -> Union[Tuple, CausalLMOutputWithPast]:
777
+ r"""
778
+ Args:
779
+ labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
780
+ Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
781
+ config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
782
+ (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.
783
+
784
+ Returns:
785
+
786
+ Example:
787
+
788
+ ```python
789
+ >>> from transformers import AutoTokenizer, LlamaForCausalLM
790
+
791
+ >>> model = LlamaForCausalLM.from_pretrained(PATH_TO_CONVERTED_WEIGHTS)
792
+ >>> tokenizer = AutoTokenizer.from_pretrained(PATH_TO_CONVERTED_TOKENIZER)
793
+
794
+ >>> prompt = "Hey, are you conscious? Can you talk to me?"
795
+ >>> inputs = tokenizer(prompt, return_tensors="pt")
796
+
797
+ >>> # Generate
798
+ >>> generate_ids = model.generate(inputs.input_ids, max_length=30)
799
+ >>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
800
+ "Hey, are you conscious? Can you talk to me?\nI'm not conscious, but I can talk to you."
801
+ ```"""
802
+
803
+ output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
804
+ output_hidden_states = (
805
+ output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
806
+ )
807
+ return_dict = return_dict if return_dict is not None else self.config.use_return_dict
808
+
809
+ is_padded_inputs = ((attention_mask is not None) and (not attention_mask.all().item()))
810
+
811
+ # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
812
+ outputs = self.model(
813
+ input_ids=input_ids,
814
+ attention_mask=attention_mask,
815
+ position_ids=position_ids,
816
+ past_key_values=past_key_values,
817
+ inputs_embeds=inputs_embeds,
818
+ use_cache=use_cache,
819
+ output_attentions=output_attentions,
820
+ output_hidden_states=output_hidden_states,
821
+ return_dict=return_dict,
822
+ is_padded_inputs=is_padded_inputs,
823
+ )
824
+
825
+ hidden_states = outputs[0]
826
+ if self.config.pretraining_tp > 1:
827
+ lm_head_slices = self.lm_head.weight.split(self.vocab_size // self.config.pretraining_tp, dim=0)
828
+ logits = [F.linear(hidden_states, lm_head_slices[i]) for i in range(self.config.pretraining_tp)]
829
+ logits = torch.cat(logits, dim=-1)
830
+ else:
831
+ logits = self.lm_head(hidden_states)
832
+ logits = logits.float()
833
+
834
+ loss = None
835
+ if labels is not None:
836
+ # Shift so that tokens < n predict n
837
+ shift_logits = logits[..., :-1, :].contiguous()
838
+ shift_labels = labels[..., 1:].contiguous()
839
+ # Flatten the tokens
840
+ loss_fct = CrossEntropyLoss()
841
+ shift_logits = shift_logits.view(-1, self.config.vocab_size)
842
+ shift_labels = shift_labels.view(-1)
843
+ # Enable model parallelism
844
+ shift_labels = shift_labels.to(shift_logits.device)
845
+ loss = loss_fct(shift_logits, shift_labels)
846
+
847
+ if not return_dict:
848
+ output = (logits,) + outputs[1:]
849
+ return (loss,) + output if loss is not None else output
850
+
851
+ return CausalLMOutputWithPast(
852
+ loss=loss,
853
+ logits=logits,
854
+ past_key_values=outputs.past_key_values,
855
+ hidden_states=outputs.hidden_states,
856
+ attentions=outputs.attentions,
857
+ )
858
+
859
+ def prepare_inputs_for_generation(
860
+ self, input_ids, past_key_values=None, attention_mask=None, inputs_embeds=None, **kwargs
861
+ ):
862
+ if past_key_values:
863
+ input_ids = input_ids[:, -1:]
864
+
865
+ position_ids = kwargs.get("position_ids", None)
866
+
867
+ # if `inputs_embeds` are passed, we only want to use them in the 1st generation step
868
+ if inputs_embeds is not None and past_key_values is None:
869
+ model_inputs = {"inputs_embeds": inputs_embeds}
870
+ else:
871
+ model_inputs = {"input_ids": input_ids}
872
+
873
+ model_inputs.update(
874
+ {
875
+ "position_ids": position_ids,
876
+ "past_key_values": past_key_values,
877
+ "use_cache": kwargs.get("use_cache"),
878
+ "attention_mask": attention_mask,
879
+ "is_padded_inputs": ((attention_mask is not None) and (not attention_mask.all().item()))
880
+ }
881
+ )
882
+ return model_inputs
883
+
884
+ @staticmethod
885
+ def _reorder_cache(past_key_values, beam_idx):
886
+ reordered_past = ()
887
+ for layer_past in past_key_values:
888
+ reordered_past += (
889
+ tuple(past_state.index_select(0, beam_idx.to(past_state.device)) for past_state in layer_past),
890
+ )
891
+ return reordered_past
892
+
893
+
894
+ @add_start_docstrings(
895
+ """
896
+ The LLaMa Model transformer with a sequence classification head on top (linear layer).
897
+
898
+ [`LlamaForSequenceClassification`] uses the last token in order to do the classification, as other causal models
899
+ (e.g. GPT-2) do.
900
+
901
+ Since it does classification on the last token, it requires to know the position of the last token. If a
902
+ `pad_token_id` is defined in the configuration, it finds the last token that is not a padding token in each row. If
903
+ no `pad_token_id` is defined, it simply takes the last value in each row of the batch. Since it cannot guess the
904
+ padding tokens when `inputs_embeds` are passed instead of `input_ids`, it does the same (take the last value in
905
+ each row of the batch).
906
+ """,
907
+ LLAMA_START_DOCSTRING,
908
+ )
909
+ class LlamaForSequenceClassification(LlamaPreTrainedModel):
910
+ def __init__(self, config):
911
+ super().__init__(config)
912
+ self.num_labels = config.num_labels
913
+ self.model = LlamaModel(config)
914
+ self.score = nn.Linear(config.hidden_size, self.num_labels, bias=False)
915
+
916
+ # Initialize weights and apply final processing
917
+ self.post_init()
918
+
919
+ def get_input_embeddings(self):
920
+ return self.model.embed_tokens
921
+
922
+ def set_input_embeddings(self, value):
923
+ self.model.embed_tokens = value
924
+
925
+ @add_start_docstrings_to_model_forward(LLAMA_INPUTS_DOCSTRING)
926
+ def forward(
927
+ self,
928
+ input_ids: torch.LongTensor = None,
929
+ attention_mask: Optional[torch.Tensor] = None,
930
+ position_ids: Optional[torch.LongTensor] = None,
931
+ past_key_values: Optional[List[torch.FloatTensor]] = None,
932
+ inputs_embeds: Optional[torch.FloatTensor] = None,
933
+ labels: Optional[torch.LongTensor] = None,
934
+ use_cache: Optional[bool] = None,
935
+ output_attentions: Optional[bool] = None,
936
+ output_hidden_states: Optional[bool] = None,
937
+ return_dict: Optional[bool] = None,
938
+ ) -> Union[Tuple, SequenceClassifierOutputWithPast]:
939
+ r"""
940
+ labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
941
+ Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
942
+ config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
943
+ `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
944
+ """
945
+ return_dict = return_dict if return_dict is not None else self.config.use_return_dict
946
+
947
+ transformer_outputs = self.model(
948
+ input_ids,
949
+ attention_mask=attention_mask,
950
+ position_ids=position_ids,
951
+ past_key_values=past_key_values,
952
+ inputs_embeds=inputs_embeds,
953
+ use_cache=use_cache,
954
+ output_attentions=output_attentions,
955
+ output_hidden_states=output_hidden_states,
956
+ return_dict=return_dict,
957
+ )
958
+ hidden_states = transformer_outputs[0]
959
+ logits = self.score(hidden_states)
960
+
961
+ if input_ids is not None:
962
+ batch_size = input_ids.shape[0]
963
+ else:
964
+ batch_size = inputs_embeds.shape[0]
965
+
966
+ if self.config.pad_token_id is None and batch_size != 1:
967
+ raise ValueError("Cannot handle batch sizes > 1 if no padding token is defined.")
968
+ if self.config.pad_token_id is None:
969
+ sequence_lengths = -1
970
+ else:
971
+ if input_ids is not None:
972
+ sequence_lengths = (torch.ne(input_ids, self.config.pad_token_id).sum(-1) - 1).to(logits.device)
973
+ else:
974
+ sequence_lengths = -1
975
+
976
+ pooled_logits = logits[torch.arange(batch_size, device=logits.device), sequence_lengths]
977
+
978
+ loss = None
979
+ if labels is not None:
980
+ labels = labels.to(logits.device)
981
+ if self.config.problem_type is None:
982
+ if self.num_labels == 1:
983
+ self.config.problem_type = "regression"
984
+ elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int):
985
+ self.config.problem_type = "single_label_classification"
986
+ else:
987
+ self.config.problem_type = "multi_label_classification"
988
+
989
+ if self.config.problem_type == "regression":
990
+ loss_fct = MSELoss()
991
+ if self.num_labels == 1:
992
+ loss = loss_fct(pooled_logits.squeeze(), labels.squeeze())
993
+ else:
994
+ loss = loss_fct(pooled_logits, labels)
995
+ elif self.config.problem_type == "single_label_classification":
996
+ loss_fct = CrossEntropyLoss()
997
+ loss = loss_fct(pooled_logits.view(-1, self.num_labels), labels.view(-1))
998
+ elif self.config.problem_type == "multi_label_classification":
999
+ loss_fct = BCEWithLogitsLoss()
1000
+ loss = loss_fct(pooled_logits, labels)
1001
+ if not return_dict:
1002
+ output = (pooled_logits,) + transformer_outputs[1:]
1003
+ return ((loss,) + output) if loss is not None else output
1004
+
1005
+ return SequenceClassifierOutputWithPast(
1006
+ loss=loss,
1007
+ logits=pooled_logits,
1008
+ past_key_values=transformer_outputs.past_key_values,
1009
+ hidden_states=transformer_outputs.hidden_states,
1010
+ attentions=transformer_outputs.attentions,
1011
+ )
pytorch_model-00001-of-00003.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:23fac2664922e7a64efce701849d30f54d42ea44ce4cc0159404c2a681f9d27f
3
+ size 9877989586
pytorch_model-00002-of-00003.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7c62729c6cf78ef213cd6c33045cd24a621e4598f0638c0b9409666e78ffa10
3
+ size 9894801014
pytorch_model-00003-of-00003.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74746070014809b91024870c7e32538ef424333741b09c761f1f66e1d50b3722
3
+ size 7180990649
pytorch_model.bin.index.json ADDED
@@ -0,0 +1,330 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 26953670656
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "pytorch_model-00003-of-00003.bin",
7
+ "model.embed_tokens.weight": "pytorch_model-00001-of-00003.bin",
8
+ "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
9
+ "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
10
+ "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
11
+ "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
12
+ "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
13
+ "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
14
+ "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
15
+ "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
16
+ "model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
17
+ "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
18
+ "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
19
+ "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
20
+ "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
21
+ "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
22
+ "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
23
+ "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
24
+ "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
25
+ "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
26
+ "model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
27
+ "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
28
+ "model.layers.10.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
29
+ "model.layers.10.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
30
+ "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
31
+ "model.layers.10.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
32
+ "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
33
+ "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
34
+ "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
35
+ "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
36
+ "model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
37
+ "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
38
+ "model.layers.11.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
39
+ "model.layers.11.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
40
+ "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
41
+ "model.layers.11.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
42
+ "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
43
+ "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
44
+ "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
45
+ "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
46
+ "model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
47
+ "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
48
+ "model.layers.12.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
49
+ "model.layers.12.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
50
+ "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
51
+ "model.layers.12.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
52
+ "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
53
+ "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
54
+ "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
55
+ "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
56
+ "model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
57
+ "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
58
+ "model.layers.13.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
59
+ "model.layers.13.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
60
+ "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
61
+ "model.layers.13.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
62
+ "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
63
+ "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
64
+ "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
65
+ "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
66
+ "model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
67
+ "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
68
+ "model.layers.14.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
69
+ "model.layers.14.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
70
+ "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
71
+ "model.layers.14.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
72
+ "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
73
+ "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
74
+ "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
75
+ "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
76
+ "model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
77
+ "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
78
+ "model.layers.15.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
79
+ "model.layers.15.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
80
+ "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
81
+ "model.layers.15.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
82
+ "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
83
+ "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
84
+ "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
85
+ "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
86
+ "model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
87
+ "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
88
+ "model.layers.16.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
89
+ "model.layers.16.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
90
+ "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
91
+ "model.layers.16.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
92
+ "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
93
+ "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
94
+ "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
95
+ "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
96
+ "model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
97
+ "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
98
+ "model.layers.17.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
99
+ "model.layers.17.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
100
+ "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
101
+ "model.layers.17.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
102
+ "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
103
+ "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
104
+ "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
105
+ "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
106
+ "model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
107
+ "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
108
+ "model.layers.18.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
109
+ "model.layers.18.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
110
+ "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
111
+ "model.layers.18.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
112
+ "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
113
+ "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
114
+ "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
115
+ "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
116
+ "model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
117
+ "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
118
+ "model.layers.19.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
119
+ "model.layers.19.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
120
+ "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
121
+ "model.layers.19.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
122
+ "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
123
+ "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
124
+ "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
125
+ "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
126
+ "model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
127
+ "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
128
+ "model.layers.2.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
129
+ "model.layers.2.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
130
+ "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
131
+ "model.layers.2.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
132
+ "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
133
+ "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
134
+ "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
135
+ "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
136
+ "model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
137
+ "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
138
+ "model.layers.20.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
139
+ "model.layers.20.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
140
+ "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
141
+ "model.layers.20.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
142
+ "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
143
+ "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
144
+ "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
145
+ "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
146
+ "model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
147
+ "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
148
+ "model.layers.21.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
149
+ "model.layers.21.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
150
+ "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
151
+ "model.layers.21.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
152
+ "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
153
+ "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
154
+ "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
155
+ "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
156
+ "model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
157
+ "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
158
+ "model.layers.22.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
159
+ "model.layers.22.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
160
+ "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
161
+ "model.layers.22.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
162
+ "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
163
+ "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
164
+ "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
165
+ "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
166
+ "model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
167
+ "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
168
+ "model.layers.23.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
169
+ "model.layers.23.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
170
+ "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
171
+ "model.layers.23.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
172
+ "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
173
+ "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
174
+ "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
175
+ "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
176
+ "model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
177
+ "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
178
+ "model.layers.24.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
179
+ "model.layers.24.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
180
+ "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
181
+ "model.layers.24.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
182
+ "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
183
+ "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
184
+ "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
185
+ "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
186
+ "model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
187
+ "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
188
+ "model.layers.25.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
189
+ "model.layers.25.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
190
+ "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
191
+ "model.layers.25.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
192
+ "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
193
+ "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
194
+ "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
195
+ "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
196
+ "model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
197
+ "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
198
+ "model.layers.26.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
199
+ "model.layers.26.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
200
+ "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
201
+ "model.layers.26.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
202
+ "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
203
+ "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
204
+ "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
205
+ "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
206
+ "model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
207
+ "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
208
+ "model.layers.27.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
209
+ "model.layers.27.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
210
+ "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
211
+ "model.layers.27.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
212
+ "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
213
+ "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
214
+ "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
215
+ "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
216
+ "model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
217
+ "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
218
+ "model.layers.28.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
219
+ "model.layers.28.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
220
+ "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
221
+ "model.layers.28.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
222
+ "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
223
+ "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
224
+ "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
225
+ "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
226
+ "model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
227
+ "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
228
+ "model.layers.29.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
229
+ "model.layers.29.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
230
+ "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
231
+ "model.layers.29.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
232
+ "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
233
+ "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
234
+ "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
235
+ "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
236
+ "model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
237
+ "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
238
+ "model.layers.3.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
239
+ "model.layers.3.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
240
+ "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
241
+ "model.layers.3.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
242
+ "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
243
+ "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
244
+ "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
245
+ "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
246
+ "model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
247
+ "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
248
+ "model.layers.30.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
249
+ "model.layers.30.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
250
+ "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
251
+ "model.layers.30.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
252
+ "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
253
+ "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
254
+ "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
255
+ "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
256
+ "model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
257
+ "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
258
+ "model.layers.31.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
259
+ "model.layers.31.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
260
+ "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
261
+ "model.layers.31.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
262
+ "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
263
+ "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
264
+ "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
265
+ "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
266
+ "model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
267
+ "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
268
+ "model.layers.4.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
269
+ "model.layers.4.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
270
+ "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
271
+ "model.layers.4.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
272
+ "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
273
+ "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
274
+ "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
275
+ "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
276
+ "model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
277
+ "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
278
+ "model.layers.5.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
279
+ "model.layers.5.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
280
+ "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
281
+ "model.layers.5.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
282
+ "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
283
+ "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
284
+ "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
285
+ "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
286
+ "model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
287
+ "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
288
+ "model.layers.6.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
289
+ "model.layers.6.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
290
+ "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
291
+ "model.layers.6.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
292
+ "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
293
+ "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
294
+ "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
295
+ "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
296
+ "model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
297
+ "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
298
+ "model.layers.7.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
299
+ "model.layers.7.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
300
+ "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
301
+ "model.layers.7.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
302
+ "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
303
+ "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
304
+ "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
305
+ "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
306
+ "model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
307
+ "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
308
+ "model.layers.8.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
309
+ "model.layers.8.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
310
+ "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
311
+ "model.layers.8.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
312
+ "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
313
+ "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
314
+ "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
315
+ "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
316
+ "model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
317
+ "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
318
+ "model.layers.9.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
319
+ "model.layers.9.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
320
+ "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
321
+ "model.layers.9.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
322
+ "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
323
+ "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
324
+ "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
325
+ "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
326
+ "model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
327
+ "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
328
+ "model.norm.weight": "pytorch_model-00003-of-00003.bin"
329
+ }
330
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "content": "<s>",
4
+ "lstrip": false,
5
+ "normalized": true,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "</s>",
11
+ "lstrip": false,
12
+ "normalized": true,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": "<unk>",
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ }
24
+ }
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
3
+ size 499723
tokenizer_config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": {
3
+ "__type": "AddedToken",
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": true,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ "clean_up_tokenization_spaces": false,
11
+ "eos_token": {
12
+ "__type": "AddedToken",
13
+ "content": "</s>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false
18
+ },
19
+ "legacy": false,
20
+ "model_max_length": 32768,
21
+ "pad_token": null,
22
+ "tokenizer_class": "LlamaTokenizer",
23
+ "unk_token": {
24
+ "__type": "AddedToken",
25
+ "content": "<unk>",
26
+ "lstrip": false,
27
+ "normalized": true,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }