forked from alxndrTL/mamba.py
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmamba_lm.py
171 lines (121 loc) · 6.11 KB
/
mamba_lm.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
from dataclasses import dataclass, fields, asdict
import json
import torch
import torch.nn as nn
import torch.nn.functional as F
from mamba import Mamba, MambaConfig, RMSNorm
"""
Encapsulates a Mamba model as language model. It has an embedding layer, and a LM head which maps the model output to logits.
"""
# TODO generate function : batch size != 1 ? (for now B=1)
# TODO generate function : top-p sampling
@dataclass
class MambaLMConfig(MambaConfig):
vocab_size: int = 32000
pad_vocab_size_multiple: int = 8
def __post_init__(self):
super().__post_init__()
if self.vocab_size % self.pad_vocab_size_multiple != 0:
self.vocab_size += (self.pad_vocab_size_multiple - self.vocab_size % self.pad_vocab_size_multiple)
def to_mamba_config(self) -> MambaConfig:
mamba_config_fields = {field.name for field in fields(MambaConfig)}
filtered_dict = {k: v for k, v in asdict(self).items() if k in mamba_config_fields}
return MambaConfig(**filtered_dict)
# adapted from https://github.com/johnma2006/mamba-minimal
def from_pretrained(name: str):
"""
Returns a model loaded with pretrained weights pulled from HuggingFace.
Args:
name: As of now, supports
* 'state-spaces/mamba-2.8b-slimpj'
* 'state-spaces/mamba-2.8b'
* 'state-spaces/mamba-1.4b'
* 'state-spaces/mamba-790m'
* 'state-spaces/mamba-370m'
* 'state-spaces/mamba-130m'
Returns:
model: a Mamba model configured with the proper parameters and initialized with the proper weights
"""
from transformers.utils import WEIGHTS_NAME, CONFIG_NAME
from transformers.utils.hub import cached_file
def load_config_hf(model_name):
resolved_archive_file = cached_file(model_name, CONFIG_NAME, _raise_exceptions_for_missing_entries=False)
return json.load(open(resolved_archive_file))
def load_state_dict_hf(model_name):
resolved_archive_file = cached_file(model_name, WEIGHTS_NAME, _raise_exceptions_for_missing_entries=False)
return torch.load(resolved_archive_file, weights_only=True, map_location='cpu', mmap=True)
# copy config data
config_data = load_config_hf(name)
config = MambaLMConfig(d_model=config_data['d_model'], n_layers=config_data['n_layer'], vocab_size=config_data['vocab_size'])
model = MambaLM(config)
# copy weights
state_dict = load_state_dict_hf(name)
new_state_dict = {}
for key in state_dict:
if key == 'backbone.embedding.weight' or key == 'backbone.norm_f.weight':
new_key = key.replace('backbone.', '')
else:
new_key = key.replace('backbone', 'mamba')
new_state_dict[new_key] = state_dict[key]
model.load_state_dict(new_state_dict)
return model
class MambaLM(nn.Module):
def __init__(self, lm_config: MambaLMConfig):
super().__init__()
self.lm_config = lm_config
self.config = lm_config.to_mamba_config()
self.embedding = nn.Embedding(self.lm_config.vocab_size, self.config.d_model)
self.mamba = Mamba(self.config)
self.norm_f = RMSNorm(self.config.d_model)
self.lm_head = nn.Linear(self.config.d_model, self.lm_config.vocab_size, bias=False)
self.lm_head.weight = self.embedding.weight
def forward(self, tokens):
# tokens : (B, L)
# logits : (B, L, vocab_size)
x = self.embedding(tokens)
x = self.mamba(x)
x = self.norm_f(x)
logits = self.lm_head(x)
return logits
def step(self, token, caches):
# token : (B)
# caches : [cache(layer) for all layers], cache : (h, inputs)
# logits : (B, vocab_size)
# caches : [cache(layer) for all layers], cache : (h, inputs)
x = self.embedding(token)
x, caches = self.mamba.step(x, caches)
x = self.norm_f(x)
logits = self.lm_head(x)
return logits, caches
# TODO process prompt in parallel, and pass in sequential mode when prompt is finished ?
def generate(self, tokenizer, prompt: str, num_tokens: int = 50, batch_size: int = 1, sample: bool = True, top_k: int = 40, temperature: float = 1.0):
self.eval()
input_ids = tokenizer(prompt, return_tensors='pt').input_ids.to(next(self.parameters()).device) # (1, num_tokens)
input_ids = input_ids.repeat(batch_size, 1)
# caches is a list of cache, one per layer
# cache is composed of : the hidden state, and the last d_conv-1 inputs
# the hidden state because the update is like an RNN
# the last d_conv-1 inputs because they are used in a 1d convolution (usually d_conv=4 so this is not large)
caches = [(None, torch.zeros(batch_size, self.config.d_inner, self.config.d_conv-1, device=input_ids.device)) for _ in range(self.config.n_layers)]
for i in range(input_ids.size(1) + num_tokens - 1):
with torch.no_grad():
# forward the new output, get new cache
next_token_logits, caches = self.step(input_ids[:, i], caches) # (batch_size, vocab_size), caches
# sample (no sampling when the prompt is being processed)
if i+1 >= input_ids.size(1):
probs = F.softmax(next_token_logits / temperature, dim=-1) # (batch_size, vocab_size)
if top_k is not None:
values, _ = torch.topk(probs, k=top_k) # (batch_size, k) ordered from lowest to biggest
probs[probs < values[:, -1, None]] = 0
probs = probs / probs.sum(axis=1, keepdims=True)
if sample:
next_token = torch.multinomial(probs, num_samples=1).squeeze(1) # (batch_size)
else:
next_token = torch.argmax(probs, dim=-1) # (batch_size)
input_ids = torch.cat([input_ids, next_token.unsqueeze(1)], dim=1)
outputs = [tokenizer.decode(output.tolist()) for output in input_ids]
self.train()
if batch_size==1:
return outputs[0]
else:
return outputs