https://github.com/shibing624/text2vec-encoder
**Text2vecEncoder** wraps the text2vec model with jina. It encodes text data into dense vectors.
https://github.com/shibing624/text2vec-encoder
Last synced: 7 months ago
JSON representation
**Text2vecEncoder** wraps the text2vec model with jina. It encodes text data into dense vectors.
- Host: GitHub
- URL: https://github.com/shibing624/text2vec-encoder
- Owner: shibing624
- License: apache-2.0
- Created: 2022-05-19T14:39:00.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2022-07-16T02:29:40.000Z (about 3 years ago)
- Last Synced: 2025-03-01T00:41:40.114Z (7 months ago)
- Language: Python
- Size: 28.3 KB
- Stars: 4
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Text2vecEncoder
**Text2vecEncoder** wraps the torch-version of transformers from huggingface. It encodes text data into dense vectors.
**Text2vecEncoder** receives [`Documents`](https://docs.jina.ai/fundamentals/document/) with `text` attributes.
The `text` attribute represents the text to be encoded. This Executor will encode each `text` into a dense vector and store them in the `embedding` attribute of the `Document`.
## Usage
From source deployment:
examples: [examples/client_demo.py](examples/client_demo.py)
```python
from docarray import Document, DocumentArray
da = DocumentArray([Document(text='如何更换花呗绑定银行卡'), Document(text='hello'), Document(text='你好'), ])
r = da.post('jinahub://Text2vecEncoder')
print(r.to_json())
```
output:
```shell
"embedding": [-0.0004445354570634663, -0.2973471283 ...]
```
Use the prebuilt images from Jina Hub in your Flow and encode an text into a dense vector.
```python
from jina import Flow, Document
f = Flow().add(uses='jinahub+docker://Text2vecEncoder')
doc = Document(content='如何更换花呗绑定银行卡')
with f:
f.post(on='/encode', inputs=doc, on_done=lambda resp: print(resp.docs[0].embedding))
```
### Set `volumes`
With the `volumes` attribute, you can map the cache directory to your local cache directory, in order to avoid downloading
the model each time you start the Flow.
```python
from jina import Flow
flow = Flow().add(
uses='jinahub+docker://Text2vecEncoder',
volumes='.cache/huggingface:/root/.cache/huggingface'
)
```
Alternatively, you can reference the docker image in the `yml` config and specify the `volumes` configuration.
`flow.yml`:
```yaml
jtype: Flow
executors:
- name: encoder
uses: 'jinahub+docker://Text2vecEncoder'
volumes: '.cache/huggingface:/root/.cache/huggingface'
```
And then use it like so:
```python
from jina import Flow
flow = Flow.load_config('flow.yml')
```
### Use other pre-trained models
You can specify the model to use with the parameter `pretrained_model_name_or_path`:
```python
from jina import Flow, Document
f = Flow().add(
uses='jinahub+docker://Text2vecEncoder',
uses_with={'pretrained_model_name_or_path': 'sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2'}
)
doc = Document(content='如何更换花呗绑定银行卡')
with f:
f.post(on='/encode', inputs=doc, on_done=lambda resp: print(resp.docs[0].embedding))
```
You can check the supported pre-trained models [here](https://huggingface.co/transformers/pretrained_models.html)
### Use GPUs
To enable GPU, you can set the `device` parameter to a cuda device.
Make sure your machine is cuda-compatible.
If you're using a docker container, make sure to add the `gpu` tag and enable
GPU access to Docker with `gpus='all'`.
Furthermore, make sure you satisfy the prerequisites mentioned in
[Executor on GPU tutorial](https://docs.jina.ai/tutorials/gpu_executor/#prerequisites).
```python
from jina import Flow, Document
f = Flow().add(
uses='jinahub+docker://Text2vecEncoder/gpu',
uses_with={'device': 'cuda'}, gpus='all'
)
doc = Document(content='如何更换花呗绑定银行卡')
with f:
f.post(on='/encode', inputs=doc, on_done=lambda resp: print(resp.docs[0].embedding))
```
## Reference
- [Huggingface Transformers](https://huggingface.co/transformers/pretrained_models.html)