Text Generation
Transformers
PyTorch
bloom
text-generation-inference
File size: 7,491 Bytes
22c0f1b
8b55706
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7f469d5
 
 
8b55706
22c0f1b
8b55706
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
---
language:
- ak
- ar
- as
- bm
- bn
- ca
- en
- es
- eu
- fon
- fr
- gu
- hi
- id
- ig
- ki
- kn
- lg
- ln
- ml
- mr
- ne
- nso
- ny
- or
- pa
- pt
- rn
- rw
- sn
- st
- sw
- ta
- te
- tn
- ts
- tum
- tw
- ur
- vi
- wo
- xh
- yo
- zh
- zu
- multilingual
license: bigscience-bloom-rail-1.0
inference: false
pipeline_tag: text-generation
---
### Quantized bigscience/bloom 1B3 with 8-bit weights

Heavily inspired by [Hivemind's GPT-J-6B with 8-bit weights](https://huggingface.co/hivemind/gpt-j-6B-8bit), this is a version of [bigscience/bloom](https://huggingface.co/bigscience/bloom-1b3) a ~1 billion parameters language model that you run and fine-tune with less memory.

Here, we also apply [LoRA (Low Rank Adaptation)](https://arxiv.org/abs/2106.09685) to reduce model size.

### How to fine-tune
TBA

### How to use

This model can be used by adapting Bloom original implementation. This is an adaptation from [Hivemind's GPT-J 8-bit](https://nbviewer.org/urls/huggingface.co/hivemind/gpt-j-6B-8bit/raw/main/convert-gpt-j.ipynb):

```python
import transformers
import torch
import torch.nn as nn
import torch.nn.functional as F

from bitsandbytes.functional import quantize_blockwise, dequantize_blockwise
from typing import Tuple
from torch.cuda.amp import custom_fwd, custom_bwd

class FrozenBNBLinear(nn.Module):
    def __init__(self, weight, absmax, code, bias=None):
        assert isinstance(bias, nn.Parameter) or bias is None
        super().__init__()
        self.out_features, self.in_features = weight.shape
        self.register_buffer("weight", weight.requires_grad_(False))
        self.register_buffer("absmax", absmax.requires_grad_(False))
        self.register_buffer("code", code.requires_grad_(False))
        self.adapter = None
        self.bias = bias
 
    def forward(self, input):
        output = DequantizeAndLinear.apply(input, self.weight, self.absmax, self.code, self.bias)
        if self.adapter:
            output += self.adapter(input)
        return output
 
    @classmethod
    def from_linear(cls, linear: nn.Linear) -> "FrozenBNBLinear":
        weights_int8, state = quantize_blockise_lowmemory(linear.weight)
        return cls(weights_int8, *state, linear.bias)
 
    def __repr__(self):
        return f"{self.__class__.__name__}({self.in_features}, {self.out_features})"
 
 
class DequantizeAndLinear(torch.autograd.Function): 
    @staticmethod
    @custom_fwd
    def forward(ctx, input: torch.Tensor, weights_quantized: torch.ByteTensor,
                absmax: torch.FloatTensor, code: torch.FloatTensor, bias: torch.FloatTensor):
        weights_deq = dequantize_blockwise(weights_quantized, absmax=absmax, code=code)
        ctx.save_for_backward(input, weights_quantized, absmax, code)
        ctx._has_bias = bias is not None
        return F.linear(input, weights_deq, bias)
 
    @staticmethod
    @custom_bwd
    def backward(ctx, grad_output: torch.Tensor):
        assert not ctx.needs_input_grad[1] and not ctx.needs_input_grad[2] and not ctx.needs_input_grad[3]
        input, weights_quantized, absmax, code = ctx.saved_tensors
        # grad_output: [*batch, out_features]
        weights_deq = dequantize_blockwise(weights_quantized, absmax=absmax, code=code)
        grad_input = grad_output @ weights_deq
        grad_bias = grad_output.flatten(0, -2).sum(dim=0) if ctx._has_bias else None
        return grad_input, None, None, None, grad_bias
 
 
class FrozenBNBEmbedding(nn.Module):
    def __init__(self, weight, absmax, code):
        super().__init__()
        self.num_embeddings, self.embedding_dim = weight.shape
        self.register_buffer("weight", weight.requires_grad_(False))
        self.register_buffer("absmax", absmax.requires_grad_(False))
        self.register_buffer("code", code.requires_grad_(False))
        self.adapter = None
 
    def forward(self, input, **kwargs):
        with torch.no_grad():
            # note: both quantuized weights and input indices are *not* differentiable
            weight_deq = dequantize_blockwise(self.weight, absmax=self.absmax, code=self.code)
            output = F.embedding(input, weight_deq, **kwargs)
        if self.adapter:
            output += self.adapter(input)
        return output 
 
    @classmethod
    def from_embedding(cls, embedding: nn.Embedding) -> "FrozenBNBEmbedding":
        weights_int8, state = quantize_blockise_lowmemory(embedding.weight)
        return cls(weights_int8, *state)
 
    def __repr__(self):
        return f"{self.__class__.__name__}({self.num_embeddings}, {self.embedding_dim})"
 
 
def quantize_blockise_lowmemory(matrix: torch.Tensor, chunk_size: int = 2 ** 20):
    assert chunk_size % 4096 == 0
    code = None
    chunks = []
    absmaxes = []
    flat_tensor = matrix.view(-1)
    for i in range((matrix.numel() - 1) // chunk_size + 1):
        input_chunk = flat_tensor[i * chunk_size: (i + 1) * chunk_size].clone()
        quantized_chunk, (absmax_chunk, code) = quantize_blockwise(input_chunk, code=code)
        chunks.append(quantized_chunk)
        absmaxes.append(absmax_chunk)
 
    matrix_i8 = torch.cat(chunks).reshape_as(matrix)
    absmax = torch.cat(absmaxes)
    return matrix_i8, (absmax, code)
 
 
def convert_to_int8(model):
    """Convert linear and embedding modules to 8-bit with optional adapters"""
    for module in list(model.modules()):
        for name, child in module.named_children():
            if isinstance(child, nn.Linear):
                print(name, child)
                setattr( 
                    module,
                    name,
                    FrozenBNBLinear(
                        weight=torch.zeros(child.out_features, child.in_features, dtype=torch.uint8),
                        absmax=torch.zeros((child.weight.numel() - 1) // 4096 + 1),
                        code=torch.zeros(256),
                        bias=child.bias,
                    ),
                )
            elif isinstance(child, nn.Embedding):
                setattr(
                    module,
                    name,
                    FrozenBNBEmbedding(
                        weight=torch.zeros(child.num_embeddings, child.embedding_dim, dtype=torch.uint8),
                        absmax=torch.zeros((child.weight.numel() - 1) // 4096 + 1),
                        code=torch.zeros(256),
                    )
                )

class BloomBlock(transformers.models.bloom.modeling_bloom.BloomBlock):
    def __init__(self, config, layer_number=None):
        super().__init__(config, layer_number)

        convert_to_int8(self.self_attention)
        convert_to_int8(self.mlp)


class BloomModel(transformers.models.bloom.modeling_bloom.BloomModel):
    def __init__(self, config):
        super().__init__(config)
        convert_to_int8(self)
        

class BloomForCausalLM(transformers.models.bloom.modeling_bloom.BloomForCausalLM):
    def __init__(self, config):
        super().__init__(config)
        convert_to_int8(self)
        
transformers.models.bloom.modeling_bloom.BloomBlock = BloomBlock

model_name = 'mrm8488/bloom-1b3-8bit'
model = BloomForCausalLM.from_pretrained(model_name, low_cpu_mem_usage=True)
tokenizer = BloomTokenizerFast.from_pretrained(model_name)

prompt = tokenizer("Given a table named salaries and columns id, created_at, salary, age. Creates a SQL to answer What is the average salary for 22 years old:", return_tensors='pt')
out = model.generate(**prompt, min_length=10, do_sample=True)
tokenizer.decode(out[0])
```