Update README.md
Browse files
README.md
CHANGED
|
@@ -2,11 +2,16 @@
|
|
| 2 |
license: apache-2.0
|
| 3 |
language:
|
| 4 |
- en
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
---
|
| 6 |
|
| 7 |
# RedPajama-INCITE-Instruct-7B-v0.1
|
| 8 |
|
| 9 |
RedPajama-INCITE-Instruct-7B-v0.1, is a large transformer-based language model developed by Together Computer and trained on the RedPajama-Data-1T dataset.
|
|
|
|
| 10 |
|
| 11 |
## Model Details
|
| 12 |
- **Developed by**: Together Computer.
|
|
@@ -21,7 +26,7 @@ Please note that the model requires `transformers` version >= 4.25.1.
|
|
| 21 |
|
| 22 |
## GPU Inference
|
| 23 |
|
| 24 |
-
This requires a GPU with
|
| 25 |
|
| 26 |
```python
|
| 27 |
import torch
|
|
@@ -54,7 +59,7 @@ Paris
|
|
| 54 |
|
| 55 |
## GPU Inference in Int8
|
| 56 |
|
| 57 |
-
This requires a GPU with
|
| 58 |
|
| 59 |
To run inference with int8, please ensure you have installed accelerate and bitandbytes. You can install them with the following command:
|
| 60 |
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
language:
|
| 4 |
- en
|
| 5 |
+
datasets:
|
| 6 |
+
- togethercomputer/RedPajama-Data-1T
|
| 7 |
+
- Muennighoff/P3
|
| 8 |
+
- Muennighoff/natural-instructions
|
| 9 |
---
|
| 10 |
|
| 11 |
# RedPajama-INCITE-Instruct-7B-v0.1
|
| 12 |
|
| 13 |
RedPajama-INCITE-Instruct-7B-v0.1, is a large transformer-based language model developed by Together Computer and trained on the RedPajama-Data-1T dataset.
|
| 14 |
+
The model was further fine-tuned on the data of [GPT-JT](https://huggingface.co/togethercomputer/GPT-JT-6B-v1), with exclusion of tasks that overlap with the HELM core scenarios.
|
| 15 |
|
| 16 |
## Model Details
|
| 17 |
- **Developed by**: Together Computer.
|
|
|
|
| 26 |
|
| 27 |
## GPU Inference
|
| 28 |
|
| 29 |
+
This requires a GPU with 16GB memory.
|
| 30 |
|
| 31 |
```python
|
| 32 |
import torch
|
|
|
|
| 59 |
|
| 60 |
## GPU Inference in Int8
|
| 61 |
|
| 62 |
+
This requires a GPU with 12GB memory.
|
| 63 |
|
| 64 |
To run inference with int8, please ensure you have installed accelerate and bitandbytes. You can install them with the following command:
|
| 65 |
|