Update README.md
Browse files
README.md
CHANGED
|
@@ -32,7 +32,7 @@ widget:
|
|
| 32 |
---
|
| 33 |
# pip-code-bandit
|
| 34 |
|
| 35 |
-
[
|
| 36 |
|
| 37 |
[colab_notebook](https://colab.research.google.com/drive/10av3SxFf0Psx_IkmZbcUhiVznStV5pVS?usp=sharing)
|
| 38 |
|
|
@@ -51,9 +51,9 @@ widget:
|
|
| 51 |
|
| 52 |
Given a goal and tools, can AI intelligently use the tools to reach the goal?\
|
| 53 |
What if it has a meagre 1.3b params/neurons akin to that of an owl? Can it follow instructions and plan to reach a goal?\
|
| 54 |
-
|
| 55 |
Releasing **pip-code-bandit** and **pipflow**\
|
| 56 |
-
A `model` and a `library` to manage and run goal
|
| 57 |
|
| 58 |
|
| 59 |
## Model attributes
|
|
@@ -77,16 +77,16 @@ A `model` and a `library` to manage and run goal oriented agentic system.
|
|
| 77 |
```
|
| 78 |
|
| 79 |
|
| 80 |
-
## How we
|
| 81 |
|
| 82 |
We used a simulator to simulate environments where the model could play games to achieve goals, given a set of actions available to it.
|
| 83 |
-
All the model could do was find the right action and config to incur positive reward.
|
| 84 |
-
The reward policy is around the concept of model going to a stable state of zero net sum reward for both good and bad behaviour.
|
| 85 |
-
In this
|
| 86 |
|
| 87 |
## License
|
| 88 |
```bash
|
| 89 |
-
complete open
|
| 90 |
```
|
| 91 |
|
| 92 |
## Usage
|
|
@@ -124,15 +124,15 @@ curl -X 'POST' \
|
|
| 124 |
-d 'model_name=PipableAI%2Fpip-code-bandit&prompt="YOUR PROMPT"&max_new_tokens=400'
|
| 125 |
```
|
| 126 |
|
| 127 |
-
Alternatively, you can directly access UI endpoint at https://playground.pipable.ai/docs#/default/infer_infer_post.
|
| 128 |
|
| 129 |
|
| 130 |
|
| 131 |
### Library Usage
|
| 132 |
|
| 133 |
-
|
| 134 |
|
| 135 |
-
For detailed usage refer to the [colab_notebook](https://colab.research.google.com/drive/10av3SxFf0Psx_IkmZbcUhiVznStV5pVS?usp=sharing)
|
| 136 |
|
| 137 |
|
| 138 |
|
|
|
|
| 32 |
---
|
| 33 |
# pip-code-bandit
|
| 34 |
|
| 35 |
+
[PipableAI](https://www.pipable.ai/)
|
| 36 |
|
| 37 |
[colab_notebook](https://colab.research.google.com/drive/10av3SxFf0Psx_IkmZbcUhiVznStV5pVS?usp=sharing)
|
| 38 |
|
|
|
|
| 51 |
|
| 52 |
Given a goal and tools, can AI intelligently use the tools to reach the goal?\
|
| 53 |
What if it has a meagre 1.3b params/neurons akin to that of an owl? Can it follow instructions and plan to reach a goal?\
|
| 54 |
+
It can!\
|
| 55 |
Releasing **pip-code-bandit** and **pipflow**\
|
| 56 |
+
A `model` and a `library` to manage and run goal-oriented agentic system.
|
| 57 |
|
| 58 |
|
| 59 |
## Model attributes
|
|
|
|
| 77 |
```
|
| 78 |
|
| 79 |
|
| 80 |
+
## How did we build it?
|
| 81 |
|
| 82 |
We used a simulator to simulate environments where the model could play games to achieve goals, given a set of actions available to it.
|
| 83 |
+
All the model could do was find the right action and config to incur a positive reward.
|
| 84 |
+
The reward policy is around the concept of a model going to a stable state of zero net sum reward for both good and bad behaviour.
|
| 85 |
+
In this setup, the model, which was pre-trained on code, function documentation, and similar OS datasets, was RL-tuned for reliability and instruction-following.
|
| 86 |
|
| 87 |
## License
|
| 88 |
```bash
|
| 89 |
+
complete open-sourced - apache 2.0. License
|
| 90 |
```
|
| 91 |
|
| 92 |
## Usage
|
|
|
|
| 124 |
-d 'model_name=PipableAI%2Fpip-code-bandit&prompt="YOUR PROMPT"&max_new_tokens=400'
|
| 125 |
```
|
| 126 |
|
| 127 |
+
Alternatively, you can directly access the UI endpoint at https://playground.pipable.ai/docs#/default/infer_infer_post.
|
| 128 |
|
| 129 |
|
| 130 |
|
| 131 |
### Library Usage
|
| 132 |
|
| 133 |
+
To directly use the model's capabilities without putting extra effort into schemas and prompts, try to use [pipflow](https://github.com/PipableAI/pipflow).
|
| 134 |
|
| 135 |
+
For detailed usage, refer to the [colab_notebook](https://colab.research.google.com/drive/10av3SxFf0Psx_IkmZbcUhiVznStV5pVS?usp=sharing)
|
| 136 |
|
| 137 |
|
| 138 |
|