File size: 1,379 Bytes
3ed6c38
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
pipeline_tag: text-generation
inference: true
widget:
- text: 'Hello!'
  example_title: Hello world
  group: Python
library_name: transformers
---

This model is randomly initialized, using the config from [databricks/dbrx-instruct](https://huggingface.co/databricks/dbrx-instruct) but with hidden dimension = 256. 
Note the model is in float16.

Codes:
```python
import transformers
import torch
import os
from huggingface_hub import create_repo, upload_folder

source_model_id = 'databricks/dbrx-instruct'
save_path = '/tmp/yujiepan/dbrx-tiny256-random'
repo_id = 'yujiepan/dbrx-tiny256-random'

config = transformers.AutoConfig.from_pretrained(
    source_model_id, trust_remote_code=True)
config.attn_config.kv_n_heads = 4
config.d_model = 256
config.ffn_config.ffn_hidden_size = 512
config.n_heads = 8
config.n_layers = 2

model = transformers.AutoModelForCausalLM.from_config(
    config, trust_remote_code=True)
model = model.half()
model.save_pretrained(save_path)

tokenizer = transformers.AutoTokenizer.from_pretrained(
    source_model_id, trust_remote_code=True)
tokenizer.save_pretrained(save_path)

result = transformers.pipelines.pipeline(
    'text-generation',
    model=model.float(), tokenizer=tokenizer)('Hello')
print(result)

os.system(f'ls -alh {save_path}')
create_repo(repo_id, exist_ok=True)
upload_folder(repo_id=repo_id, folder_path=save_path)
```