11104491 commited on
Commit
631e8b1
1 Parent(s): f9131f1

First model version

Browse files
config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "BlueLMForCausalLM"
4
+ ],
5
+ "auto_map": {
6
+ "AutoConfig": "configuration_bluelm.BlueLMConfig",
7
+ "AutoModelForCausalLM": "modeling_bluelm.BlueLMForCausalLM"
8
+ },
9
+ "bos_token_id": 1,
10
+ "eos_token_id": 2,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 4096,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 11008,
15
+ "max_position_embeddings": 2048,
16
+ "model_type": "BlueLM",
17
+ "num_attention_heads": 32,
18
+ "num_hidden_layers": 32,
19
+ "num_key_value_heads": 32,
20
+ "pad_token_id": 3,
21
+ "pretraining_tp": 1,
22
+ "rms_norm_eps": 1e-06,
23
+ "rope_scaling": null,
24
+ "rope_theta": 10000.0,
25
+ "tie_word_embeddings": false,
26
+ "torch_dtype": "bfloat16",
27
+ "transformers_version": "4.30.1",
28
+ "use_cache": true,
29
+ "use_stable_embedding": true,
30
+ "vocab_size": 100000
31
+ }
configuration_bluelm.py ADDED
@@ -0,0 +1,163 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+
21
+ """ BlueLM model configuration"""
22
+
23
+ from transformers.configuration_utils import PretrainedConfig
24
+
25
+ BlueLM_PRETRAINED_CONFIG_ARCHIVE_MAP = {}
26
+
27
+
28
+ class BlueLMConfig(PretrainedConfig):
29
+ r"""
30
+ This is the configuration class to store the configuration of a [`BlueLMModel`]. It is used to instantiate an BlueLM
31
+ model according to the specified arguments, defining the model architecture. Instantiating a configuration with the
32
+ defaults will yield a similar configuration to that of the BlueLM-7B.
33
+
34
+ Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
35
+ documentation from [`PretrainedConfig`] for more information.
36
+
37
+
38
+ Args:
39
+ vocab_size (`int`, *optional*, defaults to 32000):
40
+ Vocabulary size of the BlueLM model. Defines the number of different tokens that can be represented by the
41
+ `inputs_ids` passed when calling [`BlueLMModel`]
42
+ hidden_size (`int`, *optional*, defaults to 4096):
43
+ Dimension of the hidden representations.
44
+ intermediate_size (`int`, *optional*, defaults to 11008):
45
+ Dimension of the MLP representations.
46
+ num_hidden_layers (`int`, *optional*, defaults to 32):
47
+ Number of hidden layers in the Transformer encoder.
48
+ num_attention_heads (`int`, *optional*, defaults to 32):
49
+ Number of attention heads for each attention layer in the Transformer encoder.
50
+ num_key_value_heads (`int`, *optional*):
51
+ This is the number of key_value heads that should be used to implement Grouped Query Attention. If
52
+ `num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if
53
+ `num_key_value_heads=1 the model will use Multi Query Attention (MQA) otherwise GQA is used. When
54
+ converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed
55
+ by meanpooling all the original heads within that group. For more details checkout [this
56
+ paper](https://arxiv.org/pdf/2305.13245.pdf). If it is not specified, will default to
57
+ `num_attention_heads`.
58
+ pretraining_tp (`int`, *optional*, defaults to `1`):
59
+ Experimental feature. Tensor parallelism rank used during pretraining. Please refer to [this
60
+ document](https://huggingface.co/docs/transformers/parallelism) to understand more about it. This value is
61
+ necessary to ensure exact reproducibility of the pretraining results. Please refer to [this
62
+ issue](https://github.com/pytorch/pytorch/issues/76232).
63
+ hidden_act (`str` or `function`, *optional*, defaults to `"silu"`):
64
+ The non-linear activation function (function or string) in the decoder.
65
+ max_position_embeddings (`int`, *optional*, defaults to 2048):
66
+ The maximum sequence length that this model might ever be used with.
67
+ initializer_range (`float`, *optional*, defaults to 0.02):
68
+ The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
69
+ rms_norm_eps (`float`, *optional*, defaults to 1e-12):
70
+ The epsilon used by the rms normalization layers.
71
+ use_cache (`bool`, *optional*, defaults to `True`):
72
+ Whether or not the model should return the last key/values attentions (not used by all models). Only
73
+ relevant if `config.is_decoder=True`.
74
+ tie_word_embeddings(`bool`, *optional*, defaults to `False`):
75
+ Whether to tie weight embeddings
76
+ rope_theta (`float`, *optional*, defaults to 10000.0):
77
+ The base period of the RoPE embeddings.
78
+ rope_scaling (`Dict`, *optional*):
79
+ Dictionary containing the scaling configuration for the RoPE embeddings. Currently supports two scaling
80
+ strategies: linear and dynamic. Their scaling factor must be an float greater than 1. The expected format
81
+ is `{"type": strategy name, "factor": scaling factor}`. When using this flag, don't update
82
+ `max_position_embeddings` to the expected new maximum. See the following thread for more information on how
83
+ these scaling strategies behave:
84
+ https://www.reddit.com/r/LocalLLaMA/comments/14mrgpr/dynamically_scaled_rope_further_increases/. This is an
85
+ experimental feature, subject to breaking API changes in future versions.
86
+
87
+ """
88
+
89
+ model_type = "BlueLM"
90
+ keys_to_ignore_at_inference = ["past_key_values"]
91
+
92
+ def __init__(
93
+ self,
94
+ vocab_size=100096,
95
+ hidden_size=4096,
96
+ intermediate_size=11008,
97
+ num_hidden_layers=32,
98
+ num_attention_heads=32,
99
+ num_key_value_heads=None,
100
+ hidden_act="silu",
101
+ max_position_embeddings=2048,
102
+ initializer_range=0.02,
103
+ rms_norm_eps=1e-6,
104
+ use_cache=True,
105
+ pad_token_id=None,
106
+ bos_token_id=1,
107
+ eos_token_id=2,
108
+ pretraining_tp=1,
109
+ tie_word_embeddings=False,
110
+ rope_theta=10000.0,
111
+ rope_scaling=None,
112
+ use_stable_embedding=True,
113
+ **kwargs,
114
+ ):
115
+ self.vocab_size = vocab_size
116
+ self.max_position_embeddings = max_position_embeddings
117
+ self.hidden_size = hidden_size
118
+ self.intermediate_size = intermediate_size
119
+ self.num_hidden_layers = num_hidden_layers
120
+ self.num_attention_heads = num_attention_heads
121
+ self.use_stable_embedding = use_stable_embedding
122
+ # for backward compatibility
123
+ if num_key_value_heads is None:
124
+ num_key_value_heads = num_attention_heads
125
+
126
+ self.num_key_value_heads = num_key_value_heads
127
+ self.hidden_act = hidden_act
128
+ self.initializer_range = initializer_range
129
+ self.rms_norm_eps = rms_norm_eps
130
+ self.pretraining_tp = pretraining_tp
131
+ self.use_cache = use_cache
132
+ self.rope_theta = rope_theta
133
+ self.rope_scaling = rope_scaling
134
+ self._rope_scaling_validation()
135
+
136
+ super().__init__(
137
+ pad_token_id=pad_token_id,
138
+ bos_token_id=bos_token_id,
139
+ eos_token_id=eos_token_id,
140
+ tie_word_embeddings=tie_word_embeddings,
141
+ **kwargs,
142
+ )
143
+
144
+ def _rope_scaling_validation(self):
145
+ """
146
+ Validate the `rope_scaling` configuration.
147
+ """
148
+ if self.rope_scaling is None:
149
+ return
150
+
151
+ if not isinstance(self.rope_scaling, dict) or len(self.rope_scaling) != 2:
152
+ raise ValueError(
153
+ "`rope_scaling` must be a dictionary with with two fields, `type` and `factor`, "
154
+ f"got {self.rope_scaling}"
155
+ )
156
+ rope_scaling_type = self.rope_scaling.get("type", None)
157
+ rope_scaling_factor = self.rope_scaling.get("factor", None)
158
+ if rope_scaling_type is None or rope_scaling_type not in ["linear", "dynamic"]:
159
+ raise ValueError(
160
+ f"`rope_scaling`'s type field must be one of ['linear', 'dynamic'], got {rope_scaling_type}"
161
+ )
162
+ if rope_scaling_factor is None or not isinstance(rope_scaling_factor, float) or rope_scaling_factor <= 1.0:
163
+ raise ValueError(f"`rope_scaling`'s factor field must be an float > 1, got {rope_scaling_factor}")
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": 3,
6
+ "transformers_version": "4.30.1"
7
+ }
modeling_bluelm.py ADDED
@@ -0,0 +1,1002 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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 BlueLM model."""
21
+ import math
22
+ from typing import List, Optional, Tuple, Union
23
+
24
+ import torch
25
+ import torch.utils.checkpoint
26
+ from torch import nn
27
+ from torch.nn import BCEWithLogitsLoss, CrossEntropyLoss, MSELoss
28
+
29
+ from transformers.activations import ACT2FN
30
+ from transformers.modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast, SequenceClassifierOutputWithPast
31
+ from transformers.modeling_utils import PreTrainedModel
32
+ from transformers.utils import add_start_docstrings, add_start_docstrings_to_model_forward, logging, replace_return_docstrings
33
+ from .configuration_bluelm import BlueLMConfig
34
+
35
+
36
+ try:
37
+ from xformers import ops as xops
38
+ except ImportError:
39
+ xops = None
40
+ # print("xformers is not installed correctly.")
41
+
42
+ try:
43
+ from apex.normalization import MixedFusedRMSNorm
44
+ except ImportError:
45
+ MixedFusedRMSNorm = None
46
+ # print("Please install nvidia apex from source (https://github.com/NVIDIA/apex#linux) or use ngc container.")
47
+
48
+
49
+ logger = logging.get_logger(__name__)
50
+
51
+ _CONFIG_FOR_DOC = "BlueLmConfig"
52
+
53
+
54
+ def _make_causal_mask(input_ids_shape: torch.Size, dtype: torch.dtype, past_key_values_length: int = 0):
55
+ """
56
+ Make causal mask used for bi-directional self-attention.
57
+ """
58
+ bsz, tgt_len = input_ids_shape
59
+ mask = torch.full((tgt_len, tgt_len), torch.tensor(torch.finfo(dtype).min))
60
+ mask_cond = torch.arange(mask.size(-1))
61
+ mask.masked_fill_(mask_cond < (mask_cond + 1).view(mask.size(-1), 1), 0)
62
+ mask = mask.to(dtype)
63
+
64
+ if past_key_values_length > 0:
65
+ mask = torch.cat([torch.zeros(tgt_len, past_key_values_length, dtype=dtype), mask], dim=-1)
66
+ return mask[None, None, :, :].expand(bsz, 1, tgt_len, tgt_len + past_key_values_length)
67
+
68
+
69
+ def _expand_mask(mask: torch.Tensor, dtype: torch.dtype, tgt_len: Optional[int] = None):
70
+ """
71
+ Expands attention_mask from `[bsz, seq_len]` to `[bsz, 1, tgt_seq_len, src_seq_len]`.
72
+ """
73
+ bsz, src_len = mask.size()
74
+ tgt_len = tgt_len if tgt_len is not None else src_len
75
+
76
+ expanded_mask = mask[:, None, None, :].expand(bsz, 1, tgt_len, src_len).to(dtype)
77
+
78
+ inverted_mask = 1.0 - expanded_mask
79
+
80
+ return inverted_mask.masked_fill(inverted_mask.to(torch.bool), torch.finfo(dtype).min)
81
+
82
+
83
+ class BlueLMRMSNorm(nn.Module):
84
+ def __init__(self, hidden_size, eps=1e-6):
85
+ """
86
+ BlueLMRMSNorm is equivalent to T5LayerNorm
87
+ """
88
+ super().__init__()
89
+ self.weight = nn.Parameter(torch.ones(hidden_size))
90
+ self.variance_epsilon = eps
91
+
92
+ def forward(self, hidden_states):
93
+ variance = hidden_states.to(torch.float32).pow(2).mean(-1, keepdim=True)
94
+ hidden_states = hidden_states * torch.rsqrt(variance + self.variance_epsilon)
95
+
96
+ # convert into half-precision if necessary
97
+ if self.weight.dtype in [torch.float16, torch.bfloat16]:
98
+ hidden_states = hidden_states.to(self.weight.dtype)
99
+
100
+ return self.weight * hidden_states
101
+
102
+
103
+ class BlueLMRotaryEmbedding(torch.nn.Module):
104
+ def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None, k=16, b=0.3):
105
+ super().__init__()
106
+ # hard code bluedLM-long support 32k window size only
107
+ max_position_embeddings = 2048 * k
108
+ a = math.log(k) / ((dim / 2) ** b)
109
+ inv_freq = 1.0 / (base ** (torch.arange(0, dim, 2).float().to(device) / dim)) \
110
+ / torch.exp(a * torch.arange(1, dim / 2 + 1).float() ** b)
111
+
112
+ self.register_buffer("inv_freq", inv_freq, persistent=False)
113
+ assert self.inv_freq.dtype == torch.float32 # inv_freq must be float32 for ensuring numeric precision
114
+
115
+ # Build here to make `torch.jit.trace` work.
116
+ self.max_seq_len_cached = max_position_embeddings
117
+ t = torch.arange(self.max_seq_len_cached, device=self.inv_freq.device, dtype=self.inv_freq.dtype)
118
+ freqs = torch.einsum("i,j->ij", t, self.inv_freq)
119
+ # Different from paper, but it uses a different permutation in order to obtain the same calculation
120
+ emb = torch.cat((freqs, freqs), dim=-1)
121
+ self.register_buffer("cos_cached", emb.cos()[None, :, None, :], persistent=False)
122
+ self.register_buffer("sin_cached", emb.sin()[None, :, None, :], persistent=False)
123
+
124
+ def forward(self, x, seq_len=None):
125
+ # x: [bs, num_attention_heads, seq_len, head_size]
126
+ # This `if` block is unlikely to be run after we build sin/cos in `__init__`. Keep the logic here just in case.
127
+ if seq_len > self.max_seq_len_cached:
128
+ self.max_seq_len_cached = seq_len
129
+ t = torch.arange(self.max_seq_len_cached, device=x.device, dtype=self.inv_freq.dtype)
130
+ freqs = torch.einsum("i,j->ij", t, self.inv_freq)
131
+ # Different from paper, but it uses a different permutation in order to obtain the same calculation
132
+ emb = torch.cat((freqs, freqs), dim=-1).to(x.device)
133
+ self.register_buffer("cos_cached", emb.cos()[None, :, None, :], persistent=False)
134
+ self.register_buffer("sin_cached", emb.sin()[None, :, None, :], persistent=False)
135
+ return (
136
+ self.cos_cached[:, :seq_len, ...].to(dtype=x.dtype),
137
+ self.sin_cached[:, :seq_len, ...].to(dtype=x.dtype),
138
+ )
139
+
140
+
141
+ def rotate_half(x):
142
+ """Rotates half the hidden dims of the input."""
143
+ x1 = x[..., : x.shape[-1] // 2]
144
+ x2 = x[..., x.shape[-1] // 2 :]
145
+ return torch.cat((-x2, x1), dim=-1)
146
+
147
+
148
+ def apply_rotary_pos_emb(q, k, cos, sin, offset: int = 0):
149
+ cos = cos[:, offset : q.shape[1] + offset, ...]
150
+ sin = sin[:, offset : q.shape[1] + offset, ...]
151
+ q_embed = (q * cos) + (rotate_half(q) * sin)
152
+ k_embed = (k * cos) + (rotate_half(k) * sin)
153
+ return q_embed, k_embed
154
+
155
+
156
+ class BlueLMMLP(nn.Module):
157
+ def __init__(
158
+ self,
159
+ hidden_size: int,
160
+ intermediate_size: int,
161
+ hidden_act: str,
162
+ dropout_prob: float,
163
+ ):
164
+ super().__init__()
165
+ self.gate_proj = nn.Linear(hidden_size, intermediate_size, bias=False)
166
+ self.down_proj = nn.Linear(intermediate_size, hidden_size, bias=False)
167
+ self.up_proj = nn.Linear(hidden_size, intermediate_size, bias=False)
168
+ self.act_fn = ACT2FN[hidden_act]
169
+ self.dropout = nn.Dropout(dropout_prob)
170
+
171
+ def forward(self, x):
172
+ return self.dropout(self.down_proj(self.act_fn(self.gate_proj(x)) * self.up_proj(x)))
173
+
174
+
175
+ class BlueLMAttention(nn.Module):
176
+ """Multi-headed attention from 'Attention Is All You Need' paper"""
177
+
178
+ def __init__(
179
+ self,
180
+ hidden_size: int,
181
+ num_heads: int,
182
+ dropout_prob: float,
183
+ ):
184
+ super().__init__()
185
+ self.hidden_size = hidden_size
186
+ self.num_heads = num_heads
187
+ self.head_dim = hidden_size // num_heads
188
+ self.dropout_prob = dropout_prob
189
+
190
+ if (self.head_dim * num_heads) != self.hidden_size:
191
+ raise ValueError(
192
+ f"hidden_size must be divisible by num_heads (got `hidden_size`: {self.hidden_size}"
193
+ f" and `num_heads`: {num_heads})."
194
+ )
195
+ self.q_proj = nn.Linear(
196
+ hidden_size,
197
+ num_heads * self.head_dim,
198
+ bias=False,
199
+ )
200
+ self.k_proj = nn.Linear(
201
+ hidden_size,
202
+ num_heads * self.head_dim,
203
+ bias=False,
204
+ )
205
+ self.v_proj = nn.Linear(
206
+ hidden_size,
207
+ num_heads * self.head_dim,
208
+ bias=False,
209
+ )
210
+ self.o_proj = nn.Linear(
211
+ num_heads * self.head_dim,
212
+ hidden_size,
213
+ bias=False,
214
+ )
215
+ self.rotary_emb = BlueLMRotaryEmbedding(self.head_dim)
216
+ if xops is not None:
217
+ self.causal_mask = xops.LowerTriangularMask()
218
+
219
+ def _shape(self, tensor: torch.Tensor, seq_len: int, bsz: int):
220
+ return tensor.view(bsz, seq_len, self.num_heads, self.head_dim).contiguous()
221
+
222
+ def forward(
223
+ self,
224
+ hidden_states: torch.Tensor,
225
+ past_key_value: Optional[Tuple[torch.Tensor]] = None,
226
+ attention_mask: Optional[torch.Tensor] = None,
227
+ output_attentions: bool = False,
228
+ use_cache: bool = False,
229
+ ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
230
+ """Input shape: Batch x Time x Channel"""
231
+
232
+ bsz, q_len, _ = hidden_states.size()
233
+
234
+ query_states = self.q_proj(hidden_states).view(bsz, q_len, self.num_heads, self.head_dim)
235
+ key_states = self.k_proj(hidden_states).view(bsz, q_len, self.num_heads, self.head_dim)
236
+ value_states = self.v_proj(hidden_states).view(bsz, q_len, self.num_heads, self.head_dim)
237
+
238
+ kv_seq_len = key_states.shape[1]
239
+ offset = 0
240
+ if past_key_value is not None:
241
+ offset = past_key_value[0].shape[1]
242
+ kv_seq_len += offset
243
+ cos, sin = self.rotary_emb(value_states, seq_len=kv_seq_len)
244
+ query_states, key_states = apply_rotary_pos_emb(query_states, key_states, cos, sin, offset=offset)
245
+ # [bsz, t, nh, hd]
246
+
247
+ if past_key_value is not None:
248
+ # reuse k, v, self_attention
249
+ key_states = torch.cat([past_key_value[0], key_states], dim=1)
250
+ value_states = torch.cat([past_key_value[1], value_states], dim=1)
251
+
252
+ past_key_value = (key_states, value_states) if use_cache else None
253
+
254
+ if xops is not None and self.training:
255
+ attn_weights = None
256
+ attn_output = xops.memory_efficient_attention(
257
+ query_states, key_states, value_states, attn_bias=self.causal_mask, p=self.dropout_prob,
258
+ op=xops.fmha.MemoryEfficientAttentionFlashAttentionOp
259
+ )
260
+ else:
261
+ # [bsz, t, nh, hd]
262
+ attn_weights = torch.einsum("bqnh,bknh->bnqk", query_states, key_states) / math.sqrt(self.head_dim)
263
+
264
+ if attn_weights.size() != (bsz, self.num_heads, q_len, kv_seq_len):
265
+ raise ValueError(
266
+ f"Attention weights should be of size {(bsz * self.num_heads, q_len, kv_seq_len)}, but is"
267
+ f" {attn_weights.size()}"
268
+ )
269
+
270
+ if attention_mask is not None:
271
+ if attention_mask.size() != (bsz, 1, q_len, kv_seq_len):
272
+ raise ValueError(
273
+ f"Attention mask should be of size {(bsz, 1, q_len, kv_seq_len)}, but is {attention_mask.size()}"
274
+ )
275
+ attn_weights = attn_weights + attention_mask
276
+ attn_weights = torch.max(attn_weights, torch.tensor(torch.finfo(attn_weights.dtype).min))
277
+
278
+ # upcast attention to fp32
279
+ attn_weights = nn.functional.softmax(attn_weights, dim=-1, dtype=torch.float32).to(query_states.dtype)
280
+ attn_output = torch.einsum("bnqk,bknh->bqnh", attn_weights, value_states)
281
+
282
+ if attn_output.size() != (bsz, q_len, self.num_heads, self.head_dim):
283
+ raise ValueError(
284
+ f"`attn_output` should be of size {(bsz, q_len, self.num_heads, self.head_dim)}, but is"
285
+ f" {attn_output.size()}"
286
+ )
287
+
288
+ attn_output = attn_output.reshape(bsz, q_len, self.hidden_size)
289
+
290
+ attn_output = self.o_proj(attn_output)
291
+
292
+ if not output_attentions:
293
+ attn_weights = None
294
+
295
+ return attn_output, attn_weights, past_key_value
296
+
297
+
298
+ class BlueLMDecoderLayer(nn.Module):
299
+ def __init__(self, config: BlueLMConfig):
300
+ super().__init__()
301
+ self.hidden_size = config.hidden_size
302
+ self.self_attn = BlueLMAttention(
303
+ hidden_size=self.hidden_size,
304
+ num_heads=config.num_attention_heads,
305
+ dropout_prob=0,
306
+ )
307
+ self.mlp = BlueLMMLP(
308
+ hidden_size=self.hidden_size,
309
+ intermediate_size=config.intermediate_size,
310
+ hidden_act=config.hidden_act,
311
+ dropout_prob=0,
312
+ )
313
+ if MixedFusedRMSNorm is None:
314
+ self.input_layernorm = BlueLMRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
315
+ self.post_attention_layernorm = BlueLMRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
316
+ else:
317
+ self.input_layernorm = MixedFusedRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
318
+ self.post_attention_layernorm = MixedFusedRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
319
+
320
+ def forward(
321
+ self,
322
+ hidden_states: torch.Tensor,
323
+ attention_mask: Optional[torch.Tensor] = None,
324
+ output_attentions: Optional[bool] = False,
325
+ use_cache: Optional[bool] = False,
326
+ past_key_value: Optional[Tuple[torch.Tensor]] = None,
327
+ ) -> Tuple[torch.FloatTensor, Optional[Tuple[torch.FloatTensor, torch.FloatTensor]]]:
328
+ """
329
+ Args:
330
+ hidden_states (`torch.FloatTensor`): input to the layer of shape `(batch, seq_len, embed_dim)`
331
+ attention_mask (`torch.FloatTensor`, *optional*): attention mask of size
332
+ `(batch, 1, tgt_len, src_len)` where padding elements are indicated by very large negative values.
333
+ output_attentions (`bool`, *optional*):
334
+ Whether or not to return the attentions tensors of all attention layers. See `attentions` under
335
+ returned tensors for more detail.
336
+ use_cache (`bool`, *optional*):
337
+ If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding
338
+ (see `past_key_values`).
339
+ past_key_value (`Tuple(torch.FloatTensor)`, *optional*): cached past key and value projection states
340
+ """
341
+
342
+ residual = hidden_states
343
+
344
+ hidden_states = self.input_layernorm(hidden_states)
345
+
346
+ # Self Attention
347
+ hidden_states, self_attn_weights, present_key_value = self.self_attn(
348
+ hidden_states=hidden_states,
349
+ past_key_value=past_key_value,
350
+ attention_mask=attention_mask,
351
+ output_attentions=output_attentions,
352
+ use_cache=use_cache,
353
+ )
354
+ hidden_states = residual + hidden_states
355
+
356
+ # Fully Connected
357
+ residual = hidden_states
358
+ hidden_states = self.post_attention_layernorm(hidden_states)
359
+ hidden_states = self.mlp(hidden_states)
360
+ hidden_states = residual + hidden_states
361
+
362
+ outputs = (hidden_states,)
363
+
364
+ if output_attentions:
365
+ outputs += (self_attn_weights,)
366
+
367
+ if use_cache:
368
+ outputs += (present_key_value,)
369
+
370
+ return outputs
371
+
372
+
373
+ BlueLM_START_DOCSTRING = r"""
374
+ This model inherits from [`PreTrainedModel`]. Check the superclass documentation for the generic methods the
375
+ library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
376
+ etc.)
377
+
378
+ This model is also a PyTorch [torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) subclass.
379
+ Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage
380
+ and behavior.
381
+
382
+ Parameters:
383
+ config ([`BlueLMConfig`]):
384
+ Model configuration class with all the parameters of the model. Initializing with a config file does not
385
+ load the weights associated with the model, only the configuration. Check out the
386
+ [`~PreTrainedModel.from_pretrained`] method to load the model weights.
387
+ """
388
+
389
+
390
+ @add_start_docstrings(
391
+ "The bare BlueLM Model outputting raw hidden-states without any specific head on top.",
392
+ BlueLM_START_DOCSTRING,
393
+ )
394
+ class BlueLMPreTrainedModel(PreTrainedModel):
395
+ config_class = BlueLMConfig
396
+ base_model_prefix = "model"
397
+ supports_gradient_checkpointing = True
398
+ _no_split_modules = ["BlueLMDecoderLayer"]
399
+ _keys_to_ignore_on_load_unexpected = [r"decoder\.version"]
400
+
401
+ def _init_weights(self, module):
402
+ std = self.config.initializer_range
403
+ if isinstance(module, nn.Linear):
404
+ # module.weight.data.normal_(mean=0.0, std=std)
405
+ torch.nn.init.xavier_normal_(module.weight.data)
406
+ if module.bias is not None:
407
+ module.bias.data.zero_()
408
+ elif isinstance(module, nn.Embedding):
409
+ if self.config.use_stable_embedding:
410
+ torch.nn.init.xavier_normal_(module.weight.data)
411
+ else:
412
+ module.weight.data.normal_(mean=0.0, std=std)
413
+ if module.padding_idx is not None:
414
+ module.weight.data[module.padding_idx].zero_()
415
+
416
+ def _set_gradient_checkpointing(self, module, value=False):
417
+ if isinstance(module, BlueLMModel):
418
+ module.gradient_checkpointing = value
419
+
420
+
421
+ BlueLM_INPUTS_DOCSTRING = r"""
422
+ Args:
423
+ input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
424
+ Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you provide
425
+ it.
426
+
427
+ Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
428
+ [`PreTrainedTokenizer.__call__`] for details.
429
+
430
+ [What are input IDs?](../glossary#input-ids)
431
+ attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
432
+ Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
433
+
434
+ - 1 for tokens that are **not masked**,
435
+ - 0 for tokens that are **masked**.
436
+
437
+ [What are attention masks?](../glossary#attention-mask)
438
+
439
+ Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
440
+ [`PreTrainedTokenizer.__call__`] for details.
441
+
442
+ If `past_key_values` is used, optionally only the last `decoder_input_ids` have to be input (see
443
+ `past_key_values`).
444
+
445
+ If you want to change padding behavior, you should read [`modeling_opt._prepare_decoder_attention_mask`]
446
+ and modify to your needs. See diagram 1 in [the paper](https://arxiv.org/abs/1910.13461) for more
447
+ information on the default strategy.
448
+
449
+ - 1 indicates the head is **not masked**,
450
+ - 0 indicates the head is **masked**.
451
+
452
+ past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
453
+ Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of shape
454
+ `(batch_size, num_heads, sequence_length, embed_size_per_head)`) and 2 additional tensors of shape
455
+ `(batch_size, num_heads, encoder_sequence_length, embed_size_per_head)`.
456
+
457
+ Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention
458
+ blocks) that can be used (see `past_key_values` input) to speed up sequential decoding.
459
+
460
+ If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those that
461
+ don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of all
462
+ `decoder_input_ids` of shape `(batch_size, sequence_length)`.
463
+ inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
464
+ Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation. This
465
+ is useful if you want more control over how to convert `input_ids` indices into associated vectors than the
466
+ model's internal embedding lookup matrix.
467
+ use_cache (`bool`, *optional*):
468
+ If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding (see
469
+ `past_key_values`).
470
+ output_attentions (`bool`, *optional*):
471
+ Whether or not to return the attentions tensors of all attention layers. See `attentions` under returned
472
+ tensors for more detail.
473
+ output_hidden_states (`bool`, *optional*):
474
+ Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors for
475
+ more detail.
476
+ return_dict (`bool`, *optional*):
477
+ Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
478
+ """
479
+
480
+
481
+ @add_start_docstrings(
482
+ "The bare BlueLM Model outputting raw hidden-states without any specific head on top.",
483
+ BlueLM_START_DOCSTRING,
484
+ )
485
+ class BlueLMModel(BlueLMPreTrainedModel):
486
+ """
487
+ Transformer decoder consisting of *config.num_hidden_layers* layers. Each layer is a [`BlueLMDecoderLayer`]
488
+
489
+ Args:
490
+ config: BlueLMConfig
491
+ """
492
+
493
+ def __init__(self, config: BlueLMConfig):
494
+ super().__init__(config)
495
+ self.padding_idx = config.pad_token_id
496
+ self.vocab_size = config.vocab_size
497
+
498
+ self.embed_tokens = nn.Embedding(config.vocab_size, config.hidden_size, self.padding_idx)
499
+ if config.use_stable_embedding:
500
+ self.embed_layer_norm = nn.LayerNorm(config.hidden_size,eps=1e-06)
501
+ else:
502
+ self.embed_layer_norm = None
503
+ self.layers = nn.ModuleList([BlueLMDecoderLayer(config) for _ in range(config.num_hidden_layers)])
504
+ if MixedFusedRMSNorm is None:
505
+ self.norm = BlueLMRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
506
+ else:
507
+ self.norm = MixedFusedRMSNorm(config.hidden_size, eps=config.rms_norm_eps)
508
+
509
+ self.gradient_checkpointing = False
510
+ # Initialize weights and apply final processing
511
+ self.post_init()
512
+
513
+ def get_input_embeddings(self):
514
+ return self.embed_tokens
515
+
516
+ def set_input_embeddings(self, value):
517
+ self.embed_tokens = value
518
+
519
+ # Copied from transformers.models.bart.modeling_bart.BartDecoder._prepare_decoder_attention_mask
520
+ def _prepare_decoder_attention_mask(self, attention_mask, input_shape, inputs_embeds, past_key_values_length):
521
+ # create causal mask
522
+ # [bsz, seq_len] -> [bsz, 1, tgt_seq_len, src_seq_len]
523
+ combined_attention_mask = None
524
+ if input_shape[-1] > 1:
525
+ combined_attention_mask = _make_causal_mask(
526
+ input_shape, inputs_embeds.dtype, past_key_values_length=past_key_values_length
527
+ ).to(inputs_embeds.device)
528
+
529
+ if attention_mask is not None:
530
+ # [bsz, seq_len] -> [bsz, 1, tgt_seq_len, src_seq_len]
531
+ expanded_attn_mask = _expand_mask(attention_mask, inputs_embeds.dtype, tgt_len=input_shape[-1]).to(
532
+ inputs_embeds.device
533
+ )
534
+ combined_attention_mask = (
535
+ expanded_attn_mask if combined_attention_mask is None else expanded_attn_mask + combined_attention_mask
536
+ )
537
+
538
+ return combined_attention_mask
539
+
540
+ def forward(
541
+ self,
542
+ input_ids: torch.LongTensor = None,
543
+ attention_mask: Optional[torch.Tensor] = None,
544
+ past_key_values: Optional[List[torch.FloatTensor]] = None,
545
+ inputs_embeds: Optional[torch.FloatTensor] = None,
546
+ use_cache: Optional[bool] = None,
547
+ output_attentions: Optional[bool] = None,
548
+ output_hidden_states: Optional[bool] = None,
549
+ return_dict: Optional[bool] = None,
550
+ ) -> Union[Tuple, BaseModelOutputWithPast]:
551
+ r"""
552
+ Args:
553
+ input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
554
+ Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you
555
+ provide it.
556
+
557
+ Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
558
+ [`PreTrainedTokenizer.__call__`] for details.
559
+
560
+ [What are input IDs?](../glossary#input-ids)
561
+ attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
562
+ Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
563
+
564
+ - 1 for tokens that are **not masked**,
565
+ - 0 for tokens that are **masked**.
566
+
567
+ [What are attention masks?](../glossary#attention-mask)
568
+ past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
569
+ Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of
570
+ shape `(batch_size, num_heads, sequence_length, embed_size_per_head)`) and 2 additional tensors of
571
+
572
+ Contains pre-computed hidden-states (key and values in the self-attention blocks and in the
573
+ cross-attention blocks) that can be used (see `past_key_values` input) to speed up sequential decoding.
574
+
575
+ If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those
576
+ that don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of
577
+ all `decoder_input_ids` of shape `(batch_size, sequence_length)`.
578
+ use_cache (`bool`, *optional*):
579
+ If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding
580
+ (see `past_key_values`).
581
+ inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
582
+ Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation.
583
+ This is useful if you want more control over how to convert `input_ids` indices into associated vectors
584
+ than the model's internal embedding lookup matrix.
585
+ output_attentions (`bool`, *optional*):
586
+ Whether or not to return the attentions tensors of all attention layers. See `attentions` under
587
+ returned tensors for more detail.
588
+ output_hidden_states (`bool`, *optional*):
589
+ Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
590
+ for more detail.
591
+ return_dict (`bool`, *optional*):
592
+ Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
593
+ """
594
+ output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
595
+ output_hidden_states = (
596
+ output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
597
+ )
598
+ use_cache = use_cache if use_cache is not None else self.config.use_cache
599
+
600
+ return_dict = return_dict if return_dict is not None else self.config.use_return_dict
601
+
602
+ # retrieve input_ids and inputs_embeds
603
+ if input_ids is not None and inputs_embeds is not None:
604
+ raise ValueError("You cannot specify both decoder_input_ids and decoder_inputs_embeds at the same time")
605
+ elif input_ids is not None:
606
+ batch_size, seq_length = input_ids.shape
607
+ elif inputs_embeds is not None:
608
+ batch_size, seq_length, _ = inputs_embeds.shape
609
+ else:
610
+ raise ValueError("You have to specify either decoder_input_ids or decoder_inputs_embeds")
611
+ seq_length_with_past = seq_length
612
+ past_key_values_length = 0
613
+ if past_key_values is not None:
614
+ past_key_values_length = past_key_values[0][0].shape[2]
615
+ seq_length_with_past = seq_length_with_past + past_key_values_length
616
+ if inputs_embeds is None:
617
+ inputs_embeds = self.embed_tokens(input_ids)
618
+ if self.embed_layer_norm:
619
+ inputs_embeds = self.embed_layer_norm(inputs_embeds)
620
+ # embed positions
621
+ if xops is not None and self.training:
622
+ attention_mask = None
623
+ else:
624
+ if attention_mask is None:
625
+ attention_mask = torch.ones(
626
+ (batch_size, seq_length_with_past), dtype=torch.bool, device=inputs_embeds.device
627
+ )
628
+ attention_mask = self._prepare_decoder_attention_mask(
629
+ attention_mask, (batch_size, seq_length), inputs_embeds, past_key_values_length
630
+ )
631
+
632
+ hidden_states = inputs_embeds
633
+
634
+ if self.gradient_checkpointing and self.training:
635
+ if use_cache:
636
+ logger.warning_once(
637
+ "`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`..."
638
+ )
639
+ use_cache = False
640
+
641
+ # decoder layers
642
+ all_hidden_states = () if output_hidden_states else None
643
+ all_self_attns = () if output_attentions else None
644
+ next_decoder_cache = () if use_cache else None
645
+
646
+ for idx, decoder_layer in enumerate(self.layers):
647
+ if output_hidden_states:
648
+ all_hidden_states += (hidden_states,)
649
+
650
+ past_key_value = past_key_values[idx] if past_key_values is not None else None
651
+
652
+ if self.gradient_checkpointing and self.training:
653
+
654
+ def create_custom_forward(module):
655
+ def custom_forward(*inputs):
656
+ # None for past_key_value
657
+ return module(*inputs, output_attentions, None)
658
+
659
+ return custom_forward
660
+
661
+ layer_outputs = torch.utils.checkpoint.checkpoint(
662
+ create_custom_forward(decoder_layer),
663
+ hidden_states,
664
+ attention_mask,
665
+ None,
666
+ )
667
+ else:
668
+ layer_outputs = decoder_layer(
669
+ hidden_states,
670
+ attention_mask=attention_mask,
671
+ past_key_value=past_key_value,
672
+ output_attentions=output_attentions,
673
+ use_cache=use_cache,
674
+ )
675
+
676
+ hidden_states = layer_outputs[0]
677
+
678
+ if use_cache:
679
+ next_decoder_cache += (layer_outputs[2 if output_attentions else 1],)
680
+
681
+ if output_attentions:
682
+ all_self_attns += (layer_outputs[1],)
683
+
684
+ hidden_states = self.norm(hidden_states)
685
+
686
+ # add hidden states from the last decoder layer
687
+ if output_hidden_states:
688
+ all_hidden_states += (hidden_states,)
689
+
690
+ next_cache = next_decoder_cache if use_cache else None
691
+ if not return_dict:
692
+ return tuple(v for v in [hidden_states, next_cache, all_hidden_states, all_self_attns] if v is not None)
693
+ return BaseModelOutputWithPast(
694
+ last_hidden_state=hidden_states,
695
+ past_key_values=next_cache,
696
+ hidden_states=all_hidden_states,
697
+ attentions=all_self_attns,
698
+ )
699
+
700
+
701
+ class BlueLMForCausalLM(BlueLMPreTrainedModel):
702
+ _keys_to_ignore_on_load_missing = [r"lm_head.weight"]
703
+
704
+ def __init__(self, config):
705
+ super().__init__(config)
706
+ self.model = BlueLMModel(config)
707
+ self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
708
+
709
+ # Initialize weights and apply final processing
710
+ self.post_init()
711
+
712
+ def get_input_embeddings(self):
713
+ return self.model.embed_tokens
714
+
715
+ def set_input_embeddings(self, value):
716
+ self.model.embed_tokens = value
717
+
718
+ def get_output_embeddings(self):
719
+ return self.lm_head
720
+
721
+ def set_output_embeddings(self, new_embeddings):
722
+ self.lm_head = new_embeddings
723
+
724
+ def set_decoder(self, decoder):
725
+ self.model = decoder
726
+
727
+ def get_decoder(self):
728
+ return self.model
729
+
730
+ @replace_return_docstrings(output_type=CausalLMOutputWithPast, config_class=_CONFIG_FOR_DOC)
731
+ def forward(
732
+ self,
733
+ input_ids: torch.LongTensor = None,
734
+ attention_mask: Optional[torch.Tensor] = None,
735
+ past_key_values: Optional[List[torch.FloatTensor]] = None,
736
+ inputs_embeds: Optional[torch.FloatTensor] = None,
737
+ labels: Optional[torch.LongTensor] = None,
738
+ use_cache: Optional[bool] = None,
739
+ output_attentions: Optional[bool] = None,
740
+ output_hidden_states: Optional[bool] = None,
741
+ return_dict: Optional[bool] = None,
742
+ ) -> Union[Tuple, CausalLMOutputWithPast]:
743
+ r"""
744
+ Args:
745
+ input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
746
+ Indices of input sequence tokens in the vocabulary. Padding will be ignored by default should you
747
+ provide it.
748
+
749
+ Indices can be obtained using [`AutoTokenizer`]. See [`PreTrainedTokenizer.encode`] and
750
+ [`PreTrainedTokenizer.__call__`] for details.
751
+
752
+ [What are input IDs?](../glossary#input-ids)
753
+ attention_mask (`torch.Tensor` of shape `(batch_size, sequence_length)`, *optional*):
754
+ Mask to avoid performing attention on padding token indices. Mask values selected in `[0, 1]`:
755
+
756
+ - 1 for tokens that are **not masked**,
757
+ - 0 for tokens that are **masked**.
758
+
759
+ [What are attention masks?](../glossary#attention-mask)
760
+ past_key_values (`tuple(tuple(torch.FloatTensor))`, *optional*, returned when `use_cache=True` is passed or when `config.use_cache=True`):
761
+ Tuple of `tuple(torch.FloatTensor)` of length `config.n_layers`, with each tuple having 2 tensors of
762
+ shape `(batch_size, num_heads, sequence_length, embed_size_per_head)`) and 2 additional tensors of
763
+ shape `(batch_size, num_heads, encoder_sequence_length, embed_size_per_head)`. The two additional
764
+ tensors are only required when the model is used as a decoder in a Sequence to Sequence model.
765
+
766
+ Contains pre-computed hidden-states (key and values in the self-attention blocks and in the
767
+ cross-attention blocks) that can be used (see `past_key_values` input) to speed up sequential decoding.
768
+
769
+ If `past_key_values` are used, the user can optionally input only the last `decoder_input_ids` (those
770
+ that don't have their past key value states given to this model) of shape `(batch_size, 1)` instead of
771
+ all `decoder_input_ids` of shape `(batch_size, sequence_length)`.
772
+ inputs_embeds (`torch.FloatTensor` of shape `(batch_size, sequence_length, hidden_size)`, *optional*):
773
+ Optionally, instead of passing `input_ids` you can choose to directly pass an embedded representation.
774
+ This is useful if you want more control over how to convert `input_ids` indices into associated vectors
775
+ than the model's internal embedding lookup matrix.
776
+ labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
777
+ Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
778
+ config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
779
+ (masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.
780
+ use_cache (`bool`, *optional*):
781
+ If set to `True`, `past_key_values` key value states are returned and can be used to speed up decoding
782
+ (see `past_key_values`).
783
+ output_attentions (`bool`, *optional*):
784
+ Whether or not to return the attentions tensors of all attention layers. See `attentions` under
785
+ returned tensors for more detail.
786
+ output_hidden_states (`bool`, *optional*):
787
+ Whether or not to return the hidden states of all layers. See `hidden_states` under returned tensors
788
+ for more detail.
789
+ return_dict (`bool`, *optional*):
790
+ Whether or not to return a [`~utils.ModelOutput`] instead of a plain tuple.
791
+
792
+ Returns:
793
+
794
+ Example:
795
+
796
+ ```python
797
+ >>> from transformers import AutoTokenizer, BlueLMForCausalLM
798
+
799
+ >>> model = BlueLMForCausalLM.from_pretrained(PATH_TO_CONVERTED_WEIGHTS)
800
+ >>> tokenizer = AutoTokenizer.from_pretrained(PATH_TO_CONVERTED_TOKENIZER)
801
+
802
+ >>> prompt = "Hey, are you consciours? Can you talk to me?"
803
+ >>> inputs = tokenizer(prompt, return_tensors="pt")
804
+
805
+ >>> # Generate
806
+ >>> generate_ids = model.generate(inputs.input_ids, max_length=30)
807
+ >>> tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
808
+ "Hey, are you consciours? Can you talk to me?\nI'm not consciours, but I can talk to you."
809
+ ```"""
810
+
811
+ output_attentions = output_attentions if output_attentions is not None else self.config.output_attentions
812
+ output_hidden_states = (
813
+ output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
814
+ )
815
+ return_dict = return_dict if return_dict is not None else self.config.use_return_dict
816
+
817
+ # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
818
+ outputs = self.model(
819
+ input_ids=input_ids,
820
+ attention_mask=attention_mask,
821
+ past_key_values=past_key_values,
822
+ inputs_embeds=inputs_embeds,
823
+ use_cache=use_cache,
824
+ output_attentions=output_attentions,
825
+ output_hidden_states=output_hidden_states,
826
+ return_dict=return_dict,
827
+ )
828
+
829
+ hidden_states = outputs[0]
830
+ logits = self.lm_head(hidden_states)
831
+
832
+ loss = None
833
+ if labels is not None:
834
+ # Shift so that tokens < n predict n
835
+ shift_logits = logits[..., :-1, :].contiguous()
836
+ shift_labels = labels[..., 1:].contiguous()
837
+ # Flatten the tokens
838
+ loss_fct = CrossEntropyLoss()
839
+ shift_logits = shift_logits.view(-1, self.config.vocab_size)
840
+ shift_labels = shift_labels.view(-1)
841
+ # Enable model/pipeline parallelism
842
+ shift_labels = shift_labels.to(shift_logits.device)
843
+ loss = loss_fct(shift_logits, shift_labels)
844
+
845
+ if not return_dict:
846
+ output = (logits,) + outputs[1:]
847
+ return (loss,) + output if loss is not None else output
848
+
849
+ return CausalLMOutputWithPast(
850
+ loss=loss,
851
+ logits=logits,
852
+ past_key_values=outputs.past_key_values,
853
+ hidden_states=outputs.hidden_states,
854
+ attentions=outputs.attentions,
855
+ )
856
+
857
+ def prepare_inputs_for_generation(
858
+ self, input_ids, past_key_values=None, attention_mask=None, inputs_embeds=None, **kwargs
859
+ ):
860
+ if past_key_values:
861
+ input_ids = input_ids[:, -1:]
862
+
863
+ # if `inputs_embeds` are passed, we only want to use them in the 1st generation step
864
+ if inputs_embeds is not None and past_key_values is None:
865
+ model_inputs = {"inputs_embeds": inputs_embeds}
866
+ else:
867
+ model_inputs = {"input_ids": input_ids}
868
+
869
+ model_inputs.update(
870
+ {
871
+ "past_key_values": past_key_values,
872
+ "use_cache": kwargs.get("use_cache"),
873
+ "attention_mask": attention_mask,
874
+ }
875
+ )
876
+ return model_inputs
877
+
878
+ @staticmethod
879
+ def _reorder_cache(past_key_values, beam_idx):
880
+ reordered_past = ()
881
+ for layer_past in past_key_values:
882
+ reordered_past += (tuple(past_state.index_select(0, beam_idx) for past_state in layer_past),)
883
+ return reordered_past
884
+
885
+
886
+ @add_start_docstrings(
887
+ """
888
+ The BlueLM Model transformer with a sequence classification head on top (linear layer).
889
+
890
+ [`BlueLMForSequenceClassification`] uses the last token in order to do the classification, as other causal models
891
+ (e.g. GPT-2) do.
892
+
893
+ Since it does classification on the last token, it requires to know the position of the last token. If a
894
+ `pad_token_id` is defined in the configuration, it finds the last token that is not a padding token in each row. If
895
+ no `pad_token_id` is defined, it simply takes the last value in each row of the batch. Since it cannot guess the
896
+ padding tokens when `inputs_embeds` are passed instead of `input_ids`, it does the same (take the last value in
897
+ each row of the batch).
898
+ """,
899
+ BlueLM_START_DOCSTRING,
900
+ )
901
+ class BlueLMForSequenceClassification(BlueLMPreTrainedModel):
902
+ _keys_to_ignore_on_load_missing = [r"lm_head.weight"]
903
+
904
+ def __init__(self, config):
905
+ super().__init__(config)
906
+ self.num_labels = config.num_labels
907
+ self.model = BlueLMModel(config)
908
+ self.score = nn.Linear(config.hidden_size, self.num_labels, bias=False)
909
+
910
+ # Initialize weights and apply final processing
911
+ self.post_init()
912
+
913
+ def get_input_embeddings(self):
914
+ return self.model.embed_tokens
915
+
916
+ def set_input_embeddings(self, value):
917
+ self.model.embed_tokens = value
918
+
919
+ @add_start_docstrings_to_model_forward(BlueLM_INPUTS_DOCSTRING)
920
+ def forward(
921
+ self,
922
+ input_ids: torch.LongTensor = None,
923
+ attention_mask: Optional[torch.Tensor] = None,
924
+ past_key_values: Optional[List[torch.FloatTensor]] = None,
925
+ inputs_embeds: Optional[torch.FloatTensor] = None,
926
+ labels: Optional[torch.LongTensor] = None,
927
+ use_cache: Optional[bool] = None,
928
+ output_attentions: Optional[bool] = None,
929
+ output_hidden_states: Optional[bool] = None,
930
+ return_dict: Optional[bool] = None,
931
+ ) -> Union[Tuple, SequenceClassifierOutputWithPast]:
932
+ r"""
933
+ labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*):
934
+ Labels for computing the sequence classification/regression loss. Indices should be in `[0, ...,
935
+ config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If
936
+ `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
937
+ """
938
+ return_dict = return_dict if return_dict is not None else self.config.use_return_dict
939
+
940
+ transformer_outputs = self.model(
941
+ input_ids,
942
+ past_key_values=past_key_values,
943
+ attention_mask=attention_mask,
944
+ inputs_embeds=inputs_embeds,
945
+ use_cache=use_cache,
946
+ output_attentions=output_attentions,
947
+ output_hidden_states=output_hidden_states,
948
+ return_dict=return_dict,
949
+ )
950
+ hidden_states = transformer_outputs[0]
951
+ logits = self.score(hidden_states)
952
+
953
+ if input_ids is not None:
954
+ batch_size = input_ids.shape[0]
955
+ else:
956
+ batch_size = inputs_embeds.shape[0]
957
+
958
+ if self.config.pad_token_id is None and batch_size != 1:
959
+ raise ValueError("Cannot handle batch sizes > 1 if no padding token is defined.")
960
+ if self.config.pad_token_id is None:
961
+ sequence_lengths = -1
962
+ else:
963
+ if input_ids is not None:
964
+ sequence_lengths = (torch.ne(input_ids, self.config.pad_token_id).sum(-1) - 1).to(logits.device)
965
+ else:
966
+ sequence_lengths = -1
967
+
968
+ pooled_logits = logits[torch.arange(batch_size, device=logits.device), sequence_lengths]
969
+
970
+ loss = None
971
+ if labels is not None:
972
+ if self.config.problem_type is None:
973
+ if self.num_labels == 1:
974
+ self.config.problem_type = "regression"
975
+ elif self.num_labels > 1 and (labels.dtype == torch.long or labels.dtype == torch.int):
976
+ self.config.problem_type = "single_label_classification"
977
+ else:
978
+ self.config.problem_type = "multi_label_classification"
979
+
980
+ if self.config.problem_type == "regression":
981
+ loss_fct = MSELoss()
982
+ if self.num_labels == 1:
983
+ loss = loss_fct(pooled_logits.squeeze(), labels.squeeze())
984
+ else:
985
+ loss = loss_fct(pooled_logits, labels)
986
+ elif self.config.problem_type == "single_label_classification":
987
+ loss_fct = CrossEntropyLoss()
988
+ loss = loss_fct(pooled_logits.view(-1, self.num_labels), labels.view(-1))
989
+ elif self.config.problem_type == "multi_label_classification":
990
+ loss_fct = BCEWithLogitsLoss()
991
+ loss = loss_fct(pooled_logits, labels)
992
+ if not return_dict:
993
+ output = (pooled_logits,) + transformer_outputs[1:]
994
+ return ((loss,) + output) if loss is not None else output
995
+
996
+ return SequenceClassifierOutputWithPast(
997
+ loss=loss,
998
+ logits=pooled_logits,
999
+ past_key_values=transformer_outputs.past_key_values,
1000
+ hidden_states=transformer_outputs.hidden_states,
1001
+ attentions=transformer_outputs.attentions,
1002
+ )
pytorch_model-00001-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c624f557b943e6d63d2c3c8094cbc6e64ae9e611fd225b477965246b34c8af3b
3
+ size 1944118389
pytorch_model-00002-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:83303dd7d88f278706701a6f1eab56893fcbf5299cd04cdcde8f4eb4ce250757
3
+ size 1933671335
pytorch_model-00003-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2a1d6d7994764dfec3b374bac33dbd37a48ad10e45980b21b3a53d0b7471405
3
+ size 1933671399
pytorch_model-00004-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cd9e42a3d0969d2b59cc407a1accf08087f83fb9a86fafb986421df94771acb7
3
+ size 1990294503
pytorch_model-00005-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c0fc7ecd4b1f636c1181286fd7af1f8972f914c02643e3a0163e346553bc105a
3
+ size 1990294503
pytorch_model-00006-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca852b08d79820a16b786519a2b953d9b6dcd82c4e23d8d44095af43a664fa96
3
+ size 1990294503
pytorch_model-00007-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f9753ebd6a2aa93e5c34a72444b0bc528efa82b2b3dc27b39a0abcb814c36241
3
+ size 1990303033
pytorch_model-00008-of-00008.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51aa754fa0f9e71dfe85534b7b2979b06a4fe511c6859ef2c2f3ffc737140fc4
3
+ size 819987370
pytorch_model.bin.index.json ADDED
@@ -0,0 +1,300 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 14592532480
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "pytorch_model-00008-of-00008.bin",
7
+ "model.embed_layer_norm.bias": "pytorch_model-00001-of-00008.bin",
8
+ "model.embed_layer_norm.weight": "pytorch_model-00001-of-00008.bin",
9
+ "model.embed_tokens.weight": "pytorch_model-00001-of-00008.bin",
10
+ "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00008.bin",
11
+ "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00008.bin",
12
+ "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00008.bin",
13
+ "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00008.bin",
14
+ "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00008.bin",
15
+ "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00008.bin",
16
+ "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00008.bin",
17
+ "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00008.bin",
18
+ "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00008.bin",
19
+ "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00008.bin",
20
+ "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00008.bin",
21
+ "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00008.bin",
22
+ "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00008.bin",
23
+ "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00008.bin",
24
+ "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00008.bin",
25
+ "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00008.bin",
26
+ "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00008.bin",
27
+ "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00008.bin",
28
+ "model.layers.10.input_layernorm.weight": "pytorch_model-00003-of-00008.bin",
29
+ "model.layers.10.mlp.down_proj.weight": "pytorch_model-00003-of-00008.bin",
30
+ "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00003-of-00008.bin",
31
+ "model.layers.10.mlp.up_proj.weight": "pytorch_model-00003-of-00008.bin",
32
+ "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00003-of-00008.bin",
33
+ "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00003-of-00008.bin",
34
+ "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00003-of-00008.bin",
35
+ "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00003-of-00008.bin",
36
+ "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00003-of-00008.bin",
37
+ "model.layers.11.input_layernorm.weight": "pytorch_model-00003-of-00008.bin",
38
+ "model.layers.11.mlp.down_proj.weight": "pytorch_model-00003-of-00008.bin",
39
+ "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00003-of-00008.bin",
40
+ "model.layers.11.mlp.up_proj.weight": "pytorch_model-00003-of-00008.bin",
41
+ "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00003-of-00008.bin",
42
+ "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00003-of-00008.bin",
43
+ "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00003-of-00008.bin",
44
+ "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00003-of-00008.bin",
45
+ "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00003-of-00008.bin",
46
+ "model.layers.12.input_layernorm.weight": "pytorch_model-00004-of-00008.bin",
47
+ "model.layers.12.mlp.down_proj.weight": "pytorch_model-00004-of-00008.bin",
48
+ "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00004-of-00008.bin",
49
+ "model.layers.12.mlp.up_proj.weight": "pytorch_model-00004-of-00008.bin",
50
+ "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00004-of-00008.bin",
51
+ "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00003-of-00008.bin",
52
+ "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00003-of-00008.bin",
53
+ "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00003-of-00008.bin",
54
+ "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00003-of-00008.bin",
55
+ "model.layers.13.input_layernorm.weight": "pytorch_model-00004-of-00008.bin",
56
+ "model.layers.13.mlp.down_proj.weight": "pytorch_model-00004-of-00008.bin",
57
+ "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00004-of-00008.bin",
58
+ "model.layers.13.mlp.up_proj.weight": "pytorch_model-00004-of-00008.bin",
59
+ "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00004-of-00008.bin",
60
+ "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00004-of-00008.bin",
61
+ "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00004-of-00008.bin",
62
+ "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00004-of-00008.bin",
63
+ "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00004-of-00008.bin",
64
+ "model.layers.14.input_layernorm.weight": "pytorch_model-00004-of-00008.bin",
65
+ "model.layers.14.mlp.down_proj.weight": "pytorch_model-00004-of-00008.bin",
66
+ "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00004-of-00008.bin",
67
+ "model.layers.14.mlp.up_proj.weight": "pytorch_model-00004-of-00008.bin",
68
+ "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00004-of-00008.bin",
69
+ "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00004-of-00008.bin",
70
+ "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00004-of-00008.bin",
71
+ "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00004-of-00008.bin",
72
+ "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00004-of-00008.bin",
73
+ "model.layers.15.input_layernorm.weight": "pytorch_model-00004-of-00008.bin",
74
+ "model.layers.15.mlp.down_proj.weight": "pytorch_model-00004-of-00008.bin",
75
+ "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00004-of-00008.bin",
76
+ "model.layers.15.mlp.up_proj.weight": "pytorch_model-00004-of-00008.bin",
77
+ "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00004-of-00008.bin",
78
+ "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00004-of-00008.bin",
79
+ "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00004-of-00008.bin",
80
+ "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00004-of-00008.bin",
81
+ "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00004-of-00008.bin",
82
+ "model.layers.16.input_layernorm.weight": "pytorch_model-00004-of-00008.bin",
83
+ "model.layers.16.mlp.down_proj.weight": "pytorch_model-00004-of-00008.bin",
84
+ "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00004-of-00008.bin",
85
+ "model.layers.16.mlp.up_proj.weight": "pytorch_model-00004-of-00008.bin",
86
+ "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00004-of-00008.bin",
87
+ "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00004-of-00008.bin",
88
+ "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00004-of-00008.bin",
89
+ "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00004-of-00008.bin",
90
+ "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00004-of-00008.bin",
91
+ "model.layers.17.input_layernorm.weight": "pytorch_model-00005-of-00008.bin",
92
+ "model.layers.17.mlp.down_proj.weight": "pytorch_model-00005-of-00008.bin",
93
+ "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00005-of-00008.bin",
94
+ "model.layers.17.mlp.up_proj.weight": "pytorch_model-00005-of-00008.bin",
95
+ "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00005-of-00008.bin",
96
+ "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00004-of-00008.bin",
97
+ "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00005-of-00008.bin",
98
+ "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00004-of-00008.bin",
99
+ "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00004-of-00008.bin",
100
+ "model.layers.18.input_layernorm.weight": "pytorch_model-00005-of-00008.bin",
101
+ "model.layers.18.mlp.down_proj.weight": "pytorch_model-00005-of-00008.bin",
102
+ "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00005-of-00008.bin",
103
+ "model.layers.18.mlp.up_proj.weight": "pytorch_model-00005-of-00008.bin",
104
+ "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00005-of-00008.bin",
105
+ "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00005-of-00008.bin",
106
+ "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00005-of-00008.bin",
107
+ "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00005-of-00008.bin",
108
+ "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00005-of-00008.bin",
109
+ "model.layers.19.input_layernorm.weight": "pytorch_model-00005-of-00008.bin",
110
+ "model.layers.19.mlp.down_proj.weight": "pytorch_model-00005-of-00008.bin",
111
+ "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00005-of-00008.bin",
112
+ "model.layers.19.mlp.up_proj.weight": "pytorch_model-00005-of-00008.bin",
113
+ "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00005-of-00008.bin",
114
+ "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00005-of-00008.bin",
115
+ "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00005-of-00008.bin",
116
+ "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00005-of-00008.bin",
117
+ "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00005-of-00008.bin",
118
+ "model.layers.2.input_layernorm.weight": "pytorch_model-00002-of-00008.bin",
119
+ "model.layers.2.mlp.down_proj.weight": "pytorch_model-00001-of-00008.bin",
120
+ "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00008.bin",
121
+ "model.layers.2.mlp.up_proj.weight": "pytorch_model-00002-of-00008.bin",
122
+ "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00002-of-00008.bin",
123
+ "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00008.bin",
124
+ "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00008.bin",
125
+ "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00008.bin",
126
+ "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00008.bin",
127
+ "model.layers.20.input_layernorm.weight": "pytorch_model-00005-of-00008.bin",
128
+ "model.layers.20.mlp.down_proj.weight": "pytorch_model-00005-of-00008.bin",
129
+ "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00005-of-00008.bin",
130
+ "model.layers.20.mlp.up_proj.weight": "pytorch_model-00005-of-00008.bin",
131
+ "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00005-of-00008.bin",
132
+ "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00005-of-00008.bin",
133
+ "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00005-of-00008.bin",
134
+ "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00005-of-00008.bin",
135
+ "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00005-of-00008.bin",
136
+ "model.layers.21.input_layernorm.weight": "pytorch_model-00005-of-00008.bin",
137
+ "model.layers.21.mlp.down_proj.weight": "pytorch_model-00005-of-00008.bin",
138
+ "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00005-of-00008.bin",
139
+ "model.layers.21.mlp.up_proj.weight": "pytorch_model-00005-of-00008.bin",
140
+ "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00005-of-00008.bin",
141
+ "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00005-of-00008.bin",
142
+ "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00005-of-00008.bin",
143
+ "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00005-of-00008.bin",
144
+ "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00005-of-00008.bin",
145
+ "model.layers.22.input_layernorm.weight": "pytorch_model-00006-of-00008.bin",
146
+ "model.layers.22.mlp.down_proj.weight": "pytorch_model-00006-of-00008.bin",
147
+ "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00006-of-00008.bin",
148
+ "model.layers.22.mlp.up_proj.weight": "pytorch_model-00006-of-00008.bin",
149
+ "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00006-of-00008.bin",
150
+ "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00005-of-00008.bin",
151
+ "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00006-of-00008.bin",
152
+ "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00005-of-00008.bin",
153
+ "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00006-of-00008.bin",
154
+ "model.layers.23.input_layernorm.weight": "pytorch_model-00006-of-00008.bin",
155
+ "model.layers.23.mlp.down_proj.weight": "pytorch_model-00006-of-00008.bin",
156
+ "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00006-of-00008.bin",
157
+ "model.layers.23.mlp.up_proj.weight": "pytorch_model-00006-of-00008.bin",
158
+ "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00006-of-00008.bin",
159
+ "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00006-of-00008.bin",
160
+ "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00006-of-00008.bin",
161
+ "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00006-of-00008.bin",
162
+ "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00006-of-00008.bin",
163
+ "model.layers.24.input_layernorm.weight": "pytorch_model-00006-of-00008.bin",
164
+ "model.layers.24.mlp.down_proj.weight": "pytorch_model-00006-of-00008.bin",
165
+ "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00006-of-00008.bin",
166
+ "model.layers.24.mlp.up_proj.weight": "pytorch_model-00006-of-00008.bin",
167
+ "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00006-of-00008.bin",
168
+ "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00006-of-00008.bin",
169
+ "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00006-of-00008.bin",
170
+ "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00006-of-00008.bin",
171
+ "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00006-of-00008.bin",
172
+ "model.layers.25.input_layernorm.weight": "pytorch_model-00006-of-00008.bin",
173
+ "model.layers.25.mlp.down_proj.weight": "pytorch_model-00006-of-00008.bin",
174
+ "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00006-of-00008.bin",
175
+ "model.layers.25.mlp.up_proj.weight": "pytorch_model-00006-of-00008.bin",
176
+ "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00006-of-00008.bin",
177
+ "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00006-of-00008.bin",
178
+ "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00006-of-00008.bin",
179
+ "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00006-of-00008.bin",
180
+ "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00006-of-00008.bin",
181
+ "model.layers.26.input_layernorm.weight": "pytorch_model-00006-of-00008.bin",
182
+ "model.layers.26.mlp.down_proj.weight": "pytorch_model-00006-of-00008.bin",
183
+ "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00006-of-00008.bin",
184
+ "model.layers.26.mlp.up_proj.weight": "pytorch_model-00006-of-00008.bin",
185
+ "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00006-of-00008.bin",
186
+ "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00006-of-00008.bin",
187
+ "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00006-of-00008.bin",
188
+ "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00006-of-00008.bin",
189
+ "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00006-of-00008.bin",
190
+ "model.layers.27.input_layernorm.weight": "pytorch_model-00007-of-00008.bin",
191
+ "model.layers.27.mlp.down_proj.weight": "pytorch_model-00007-of-00008.bin",
192
+ "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00007-of-00008.bin",
193
+ "model.layers.27.mlp.up_proj.weight": "pytorch_model-00007-of-00008.bin",
194
+ "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00007-of-00008.bin",
195
+ "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00007-of-00008.bin",
196
+ "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00007-of-00008.bin",
197
+ "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00006-of-00008.bin",
198
+ "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00007-of-00008.bin",
199
+ "model.layers.28.input_layernorm.weight": "pytorch_model-00007-of-00008.bin",
200
+ "model.layers.28.mlp.down_proj.weight": "pytorch_model-00007-of-00008.bin",
201
+ "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00007-of-00008.bin",
202
+ "model.layers.28.mlp.up_proj.weight": "pytorch_model-00007-of-00008.bin",
203
+ "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00007-of-00008.bin",
204
+ "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00007-of-00008.bin",
205
+ "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00007-of-00008.bin",
206
+ "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00007-of-00008.bin",
207
+ "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00007-of-00008.bin",
208
+ "model.layers.29.input_layernorm.weight": "pytorch_model-00007-of-00008.bin",
209
+ "model.layers.29.mlp.down_proj.weight": "pytorch_model-00007-of-00008.bin",
210
+ "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00007-of-00008.bin",
211
+ "model.layers.29.mlp.up_proj.weight": "pytorch_model-00007-of-00008.bin",
212
+ "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00007-of-00008.bin",
213
+ "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00007-of-00008.bin",
214
+ "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00007-of-00008.bin",
215
+ "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00007-of-00008.bin",
216
+ "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00007-of-00008.bin",
217
+ "model.layers.3.input_layernorm.weight": "pytorch_model-00002-of-00008.bin",
218
+ "model.layers.3.mlp.down_proj.weight": "pytorch_model-00002-of-00008.bin",
219
+ "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00002-of-00008.bin",
220
+ "model.layers.3.mlp.up_proj.weight": "pytorch_model-00002-of-00008.bin",
221
+ "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00002-of-00008.bin",
222
+ "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00002-of-00008.bin",
223
+ "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00002-of-00008.bin",
224
+ "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00002-of-00008.bin",
225
+ "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00002-of-00008.bin",
226
+ "model.layers.30.input_layernorm.weight": "pytorch_model-00007-of-00008.bin",
227
+ "model.layers.30.mlp.down_proj.weight": "pytorch_model-00007-of-00008.bin",
228
+ "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00007-of-00008.bin",
229
+ "model.layers.30.mlp.up_proj.weight": "pytorch_model-00007-of-00008.bin",
230
+ "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00007-of-00008.bin",
231
+ "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00007-of-00008.bin",
232
+ "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00007-of-00008.bin",
233
+ "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00007-of-00008.bin",
234
+ "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00007-of-00008.bin",
235
+ "model.layers.31.input_layernorm.weight": "pytorch_model-00007-of-00008.bin",
236
+ "model.layers.31.mlp.down_proj.weight": "pytorch_model-00007-of-00008.bin",
237
+ "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00007-of-00008.bin",
238
+ "model.layers.31.mlp.up_proj.weight": "pytorch_model-00007-of-00008.bin",
239
+ "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00007-of-00008.bin",
240
+ "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00007-of-00008.bin",
241
+ "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00007-of-00008.bin",
242
+ "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00007-of-00008.bin",
243
+ "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00007-of-00008.bin",
244
+ "model.layers.4.input_layernorm.weight": "pytorch_model-00002-of-00008.bin",
245
+ "model.layers.4.mlp.down_proj.weight": "pytorch_model-00002-of-00008.bin",
246
+ "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00002-of-00008.bin",
247
+ "model.layers.4.mlp.up_proj.weight": "pytorch_model-00002-of-00008.bin",
248
+ "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00002-of-00008.bin",
249
+ "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00002-of-00008.bin",
250
+ "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00002-of-00008.bin",
251
+ "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00002-of-00008.bin",
252
+ "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00002-of-00008.bin",
253
+ "model.layers.5.input_layernorm.weight": "pytorch_model-00002-of-00008.bin",
254
+ "model.layers.5.mlp.down_proj.weight": "pytorch_model-00002-of-00008.bin",
255
+ "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00002-of-00008.bin",
256
+ "model.layers.5.mlp.up_proj.weight": "pytorch_model-00002-of-00008.bin",
257
+ "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00002-of-00008.bin",
258
+ "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00002-of-00008.bin",
259
+ "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00002-of-00008.bin",
260
+ "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00002-of-00008.bin",
261
+ "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00002-of-00008.bin",
262
+ "model.layers.6.input_layernorm.weight": "pytorch_model-00002-of-00008.bin",
263
+ "model.layers.6.mlp.down_proj.weight": "pytorch_model-00002-of-00008.bin",
264
+ "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00002-of-00008.bin",
265
+ "model.layers.6.mlp.up_proj.weight": "pytorch_model-00002-of-00008.bin",
266
+ "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00002-of-00008.bin",
267
+ "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00002-of-00008.bin",
268
+ "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00002-of-00008.bin",
269
+ "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00002-of-00008.bin",
270
+ "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00002-of-00008.bin",
271
+ "model.layers.7.input_layernorm.weight": "pytorch_model-00003-of-00008.bin",
272
+ "model.layers.7.mlp.down_proj.weight": "pytorch_model-00003-of-00008.bin",
273
+ "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00002-of-00008.bin",
274
+ "model.layers.7.mlp.up_proj.weight": "pytorch_model-00003-of-00008.bin",
275
+ "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00003-of-00008.bin",
276
+ "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00002-of-00008.bin",
277
+ "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00002-of-00008.bin",
278
+ "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00002-of-00008.bin",
279
+ "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00002-of-00008.bin",
280
+ "model.layers.8.input_layernorm.weight": "pytorch_model-00003-of-00008.bin",
281
+ "model.layers.8.mlp.down_proj.weight": "pytorch_model-00003-of-00008.bin",
282
+ "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00003-of-00008.bin",
283
+ "model.layers.8.mlp.up_proj.weight": "pytorch_model-00003-of-00008.bin",
284
+ "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00003-of-00008.bin",
285
+ "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00003-of-00008.bin",
286
+ "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00003-of-00008.bin",
287
+ "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00003-of-00008.bin",
288
+ "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00003-of-00008.bin",
289
+ "model.layers.9.input_layernorm.weight": "pytorch_model-00003-of-00008.bin",
290
+ "model.layers.9.mlp.down_proj.weight": "pytorch_model-00003-of-00008.bin",
291
+ "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00003-of-00008.bin",
292
+ "model.layers.9.mlp.up_proj.weight": "pytorch_model-00003-of-00008.bin",
293
+ "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00003-of-00008.bin",
294
+ "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00003-of-00008.bin",
295
+ "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00003-of-00008.bin",
296
+ "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00003-of-00008.bin",
297
+ "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00003-of-00008.bin",
298
+ "model.norm.weight": "pytorch_model-00007-of-00008.bin"
299
+ }
300
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [],
3
+ "bos_token": {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": true,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ "eos_token": {
11
+ "content": "</s>",
12
+ "lstrip": false,
13
+ "normalized": true,
14
+ "rstrip": false,
15
+ "single_word": false
16
+ },
17
+ "unk_token": {
18
+ "content": "<unk>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<pad>",
26
+ "lstrip": false,
27
+ "normalized": true,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
tokenization_bluelm.py ADDED
@@ -0,0 +1,251 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+
21
+ """Tokenization classes for BlueLM."""
22
+ import os
23
+ from shutil import copyfile
24
+ from typing import Any, Dict, List, Optional, Tuple
25
+
26
+ import sentencepiece as spm
27
+
28
+ from transformers.tokenization_utils import AddedToken, PreTrainedTokenizer
29
+ from transformers.utils import logging
30
+
31
+
32
+ logger = logging.get_logger(__name__)
33
+
34
+ VOCAB_FILES_NAMES = {"vocab_file": "tokenizer.model"}
35
+
36
+ PRETRAINED_VOCAB_FILES_MAP = {
37
+ "vocab_file": {},
38
+ "tokenizer_file": {},
39
+ }
40
+ PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES = {}
41
+
42
+
43
+ class BlueLMTokenizer(PreTrainedTokenizer):
44
+ """
45
+ Construct a BlueLM tokenizer. Based on byte-level Byte-Pair-Encoding.
46
+
47
+ Args:
48
+ vocab_file (`str`):
49
+ Path to the vocabulary file.
50
+ """
51
+
52
+ vocab_files_names = VOCAB_FILES_NAMES
53
+ pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP
54
+ max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES
55
+ model_input_names = ["input_ids", "attention_mask"]
56
+
57
+ def __init__(
58
+ self,
59
+ vocab_file,
60
+ unk_token="<unk>",
61
+ bos_token="<s>",
62
+ eos_token="</s>",
63
+ pad_token=None,
64
+ sp_model_kwargs: Optional[Dict[str, Any]] = None,
65
+ add_bos_token=True,
66
+ add_eos_token=False,
67
+ clean_up_tokenization_spaces=False,
68
+ **kwargs,
69
+ ):
70
+ self.sp_model_kwargs = {} if sp_model_kwargs is None else sp_model_kwargs
71
+ bos_token = AddedToken(bos_token, lstrip=False, rstrip=False) if isinstance(bos_token, str) else bos_token
72
+ eos_token = AddedToken(eos_token, lstrip=False, rstrip=False) if isinstance(eos_token, str) else eos_token
73
+ unk_token = AddedToken(unk_token, lstrip=False, rstrip=False) if isinstance(unk_token, str) else unk_token
74
+ pad_token = AddedToken(pad_token, lstrip=False, rstrip=False) if isinstance(pad_token, str) else pad_token
75
+ super().__init__(
76
+ bos_token=bos_token,
77
+ eos_token=eos_token,
78
+ unk_token=unk_token,
79
+ pad_token=pad_token,
80
+ add_bos_token=add_bos_token,
81
+ add_eos_token=add_eos_token,
82
+ sp_model_kwargs=self.sp_model_kwargs,
83
+ clean_up_tokenization_spaces=clean_up_tokenization_spaces,
84
+ **kwargs,
85
+ )
86
+ self.vocab_file = vocab_file
87
+ self.add_bos_token = add_bos_token
88
+ self.add_eos_token = add_eos_token
89
+ self.sp_model = spm.SentencePieceProcessor(**self.sp_model_kwargs)
90
+ self.sp_model.Load(vocab_file)
91
+
92
+ def __getstate__(self):
93
+ state = self.__dict__.copy()
94
+ state["sp_model"] = None
95
+ return state
96
+
97
+ def __setstate__(self, d):
98
+ self.__dict__ = d
99
+ self.sp_model = spm.SentencePieceProcessor(**self.sp_model_kwargs)
100
+ self.sp_model.Load(self.vocab_file)
101
+
102
+ @property
103
+ def vocab_size(self):
104
+ """Returns vocab size"""
105
+ return self.sp_model.get_piece_size()
106
+
107
+ def get_vocab(self):
108
+ """Returns vocab as a dict"""
109
+ vocab = {self.convert_ids_to_tokens(i): i for i in range(self.vocab_size)}
110
+ vocab.update(self.added_tokens_encoder)
111
+ return vocab
112
+
113
+ def _tokenize(self, text):
114
+ """Returns a tokenized string."""
115
+ return self.sp_model.encode(text, out_type=str)
116
+
117
+ def _convert_token_to_id(self, token):
118
+ """Converts a token (str) in an id using the vocab."""
119
+ return self.sp_model.piece_to_id(token)
120
+
121
+ def _convert_id_to_token(self, index):
122
+ """Converts an index (integer) in a token (str) using the vocab."""
123
+ token = self.sp_model.IdToPiece(index)
124
+ return token
125
+
126
+ def convert_tokens_to_string(self, tokens):
127
+ """Converts a sequence of tokens (string) in a single string."""
128
+ current_sub_tokens = []
129
+ out_string = ""
130
+ prev_is_special = False
131
+ for i, token in enumerate(tokens):
132
+ # make sure that special tokens are not decoded using sentencepiece model
133
+ if token in self.all_special_tokens:
134
+ if not prev_is_special and i != 0:
135
+ out_string += " "
136
+ out_string += self.sp_model.decode(current_sub_tokens) + token
137
+ prev_is_special = True
138
+ current_sub_tokens = []
139
+ else:
140
+ current_sub_tokens.append(token)
141
+ prev_is_special = False
142
+ out_string += self.sp_model.decode(current_sub_tokens)
143
+ return out_string
144
+
145
+ def save_vocabulary(self, save_directory, filename_prefix: Optional[str] = None) -> Tuple[str]:
146
+ """
147
+ Save the vocabulary and special tokens file to a directory.
148
+
149
+ Args:
150
+ save_directory (`str`):
151
+ The directory in which to save the vocabulary.
152
+
153
+ Returns:
154
+ `Tuple(str)`: Paths to the files saved.
155
+ """
156
+ if not os.path.isdir(save_directory):
157
+ logger.error(f"Vocabulary path ({save_directory}) should be a directory")
158
+ return
159
+ out_vocab_file = os.path.join(
160
+ save_directory, (filename_prefix + "-" if filename_prefix else "") + VOCAB_FILES_NAMES["vocab_file"]
161
+ )
162
+
163
+ if os.path.abspath(self.vocab_file) != os.path.abspath(out_vocab_file) and os.path.isfile(self.vocab_file):
164
+ copyfile(self.vocab_file, out_vocab_file)
165
+ elif not os.path.isfile(self.vocab_file):
166
+ with open(out_vocab_file, "wb") as fi:
167
+ content_spiece_model = self.sp_model.serialized_model_proto()
168
+ fi.write(content_spiece_model)
169
+
170
+ return (out_vocab_file,)
171
+
172
+ def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
173
+ bos_token_id = [self.bos_token_id] if self.add_bos_token else []
174
+ eos_token_id = [self.eos_token_id] if self.add_eos_token else []
175
+
176
+ output = bos_token_id + token_ids_0 + eos_token_id
177
+
178
+ if token_ids_1 is not None:
179
+ output = output + bos_token_id + token_ids_1 + eos_token_id
180
+
181
+ return output
182
+
183
+ def get_special_tokens_mask(
184
+ self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None, already_has_special_tokens: bool = False
185
+ ) -> List[int]:
186
+ """
187
+ Retrieve sequence ids from a token list that has no special tokens added. This method is called when adding
188
+ special tokens using the tokenizer `prepare_for_model` method.
189
+
190
+ Args:
191
+ token_ids_0 (`List[int]`):
192
+ List of IDs.
193
+ token_ids_1 (`List[int]`, *optional*):
194
+ Optional second list of IDs for sequence pairs.
195
+ already_has_special_tokens (`bool`, *optional*, defaults to `False`):
196
+ Whether or not the token list is already formatted with special tokens for the model.
197
+
198
+ Returns:
199
+ `List[int]`: A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.
200
+ """
201
+ if already_has_special_tokens:
202
+ return super().get_special_tokens_mask(
203
+ token_ids_0=token_ids_0, token_ids_1=token_ids_1, already_has_special_tokens=True
204
+ )
205
+
206
+ bos_token_id = [1] if self.add_bos_token else []
207
+ eos_token_id = [1] if self.add_eos_token else []
208
+
209
+ if token_ids_1 is None:
210
+ return bos_token_id + ([0] * len(token_ids_0)) + eos_token_id
211
+ return (
212
+ bos_token_id
213
+ + ([0] * len(token_ids_0))
214
+ + eos_token_id
215
+ + bos_token_id
216
+ + ([0] * len(token_ids_1))
217
+ + eos_token_id
218
+ )
219
+
220
+ def create_token_type_ids_from_sequences(
221
+ self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
222
+ ) -> List[int]:
223
+ """
224
+ Creates a mask from the two sequences passed to be used in a sequence-pair classification task. An ALBERT
225
+ sequence pair mask has the following format:
226
+
227
+ ```
228
+ 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1
229
+ | first sequence | second sequence |
230
+ ```
231
+
232
+ if token_ids_1 is None, only returns the first portion of the mask (0s).
233
+
234
+ Args:
235
+ token_ids_0 (`List[int]`):
236
+ List of ids.
237
+ token_ids_1 (`List[int]`, *optional*):
238
+ Optional second list of IDs for sequence pairs.
239
+
240
+ Returns:
241
+ `List[int]`: List of [token type IDs](../glossary#token-type-ids) according to the given sequence(s).
242
+ """
243
+ bos_token_id = [self.bos_token_id] if self.add_bos_token else []
244
+ eos_token_id = [self.eos_token_id] if self.add_eos_token else []
245
+
246
+ output = [0] * len(bos_token_id + token_ids_0 + eos_token_id)
247
+
248
+ if token_ids_1 is not None:
249
+ output += [1] * len(bos_token_id + token_ids_1 + eos_token_id)
250
+
251
+ return output
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5ed07a4a6a74d6a69f56478892da8a06fbaa29dc27ff4d957fda6237643150b
3
+ size 1609668
tokenizer_config.json ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "auto_map": {
3
+ "AutoTokenizer": ["tokenization_bluelm.BlueLMTokenizer", null]
4
+ },
5
+ "add_bos_token": true,
6
+ "add_eos_token": false,
7
+ "bos_token": {
8
+ "__type": "AddedToken",
9
+ "content": "<s>",
10
+ "lstrip": false,
11
+ "normalized": true,
12
+ "rstrip": false,
13
+ "single_word": false
14
+ },
15
+ "clean_up_tokenization_spaces": false,
16
+ "eos_token": {
17
+ "__type": "AddedToken",
18
+ "content": "</s>",
19
+ "lstrip": false,
20
+ "normalized": true,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "model_max_length": 1000000000000000019884624838656,
25
+ "pad_token": {
26
+ "__type": "AddedToken",
27
+ "content": "<pad>",
28
+ "lstrip": false,
29
+ "normalized": true,
30
+ "rstrip": false,
31
+ "single_word": false
32
+ },
33
+ "sp_model_kwargs": {},
34
+ "tokenizer_class": "BlueLMTokenizer",
35
+ "unk_token": {
36
+ "__type": "AddedToken",
37
+ "content": "<unk>",
38
+ "lstrip": false,
39
+ "normalized": true,
40
+ "rstrip": false,
41
+ "single_word": false
42
+ }
43
+ }