2023-02-24 14:58:58 +00:00
# Banana
This page covers how to use the Banana ecosystem within LangChain.
It is broken into two parts: installation and setup, and then references to specific Banana wrappers.
## Installation and Setup
2023-02-27 01:53:57 +00:00
2023-03-09 04:16:09 +00:00
- Install with `pip install banana-dev`
2023-02-27 01:53:57 +00:00
- Get an Banana api key and set it as an environment variable (`BANANA_API_KEY`)
2023-02-24 14:58:58 +00:00
## Define your Banana Template
2023-02-27 01:53:57 +00:00
If you want to use an available language model template you can find one [here ](https://app.banana.dev/templates/conceptofmind/serverless-template-palmyra-base ).
This template uses the Palmyra-Base model by [Writer ](https://writer.com/product/api/ ).
2023-02-24 14:58:58 +00:00
You can check out an example Banana repository [here ](https://github.com/conceptofmind/serverless-template-palmyra-base ).
## Build the Banana app
2023-02-27 01:53:57 +00:00
Banana Apps must include the "output" key in the return json.
There is a rigid response structure.
2023-02-24 14:58:58 +00:00
```python
# Return the results as a dictionary
result = {'output': result}
```
An example inference function would be:
2023-02-27 01:53:57 +00:00
2023-02-24 14:58:58 +00:00
```python
def inference(model_inputs:dict) -> dict:
global model
global tokenizer
# Parse out your arguments
prompt = model_inputs.get('prompt', None)
if prompt == None:
return {'message': "No prompt provided"}
2023-02-27 01:53:57 +00:00
2023-02-24 14:58:58 +00:00
# Run the model
input_ids = tokenizer.encode(prompt, return_tensors='pt').cuda()
output = model.generate(
2023-02-27 01:53:57 +00:00
input_ids,
max_length=100,
do_sample=True,
top_k=50,
top_p=0.95,
num_return_sequences=1,
temperature=0.9,
early_stopping=True,
no_repeat_ngram_size=3,
num_beams=5,
length_penalty=1.5,
repetition_penalty=1.5,
2023-02-24 14:58:58 +00:00
bad_words_ids=[[tokenizer.encode(' ', add_prefix_space=True)[0]]]
)
result = tokenizer.decode(output[0], skip_special_tokens=True)
# Return the results as a dictionary
result = {'output': result}
return result
```
You can find a full example of a Banana app [here ](https://github.com/conceptofmind/serverless-template-palmyra-base/blob/main/app.py ).
## Wrappers
### LLM
2023-02-27 01:53:57 +00:00
There exists an Banana LLM wrapper, which you can access with
2023-02-24 14:58:58 +00:00
```python
from langchain.llms import Banana
```
You need to provide a model key located in the dashboard:
2023-02-27 01:53:57 +00:00
2023-02-24 14:58:58 +00:00
```python
llm = Banana(model_key="YOUR_MODEL_KEY")
2023-02-27 01:53:57 +00:00
```