Update README.md
Browse files
README.md
CHANGED
|
@@ -25,6 +25,23 @@ This model has been finetuned from [GPT-J](https://huggingface.co/EleutherAI/gpt
|
|
| 25 |
- **License:** Apache-2
|
| 26 |
- **Finetuned from model [optional]:** [GPT-J](https://huggingface.co/EleutherAI/gpt-j-6B)
|
| 27 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
### Model Sources [optional]
|
| 29 |
|
| 30 |
<!-- Provide the basic links for the model. -->
|
|
|
|
| 25 |
- **License:** Apache-2
|
| 26 |
- **Finetuned from model [optional]:** [GPT-J](https://huggingface.co/EleutherAI/gpt-j-6B)
|
| 27 |
|
| 28 |
+
|
| 29 |
+
We have released several versions of our finetuned GPT-J model using [different dataset versions](https://huggingface.co/datasets/nomic-ai/gpt4all-j-prompt-generations)
|
| 30 |
+
|
| 31 |
+
- v1.0: The original model trained on the v1.0 dataset
|
| 32 |
+
- v1.1-breezy: Trained on afiltered dataset where we removed all instances of AI language model
|
| 33 |
+
- v1.2-jazzy: Trained on a filtered dataset where we also removed instances like I'm sorry, I can't answer... and AI language model
|
| 34 |
+
|
| 35 |
+
To download a model with a specific revision run
|
| 36 |
+
|
| 37 |
+
```python
|
| 38 |
+
from transformers import AutoModelForCausalLM
|
| 39 |
+
|
| 40 |
+
model = AutoModelForCausalLM.from_pretrained("nomic-ai/gpt4all-j", revision="v1.2-jazzy")
|
| 41 |
+
```
|
| 42 |
+
|
| 43 |
+
Downloading without specifying `revision` defaults to `main`/`v1.0`.
|
| 44 |
+
|
| 45 |
### Model Sources [optional]
|
| 46 |
|
| 47 |
<!-- Provide the basic links for the model. -->
|