Update README.md
Browse files
README.md
CHANGED
@@ -5,10 +5,11 @@ license: mit
|
|
5 |
### Model Name: CodeOptimus - Supervised Instruction Fine-tuning [Mistral 7B Instruct](https://mistral.ai/news/announcing-mistral-7b/) using qLora.
|
6 |
|
7 |
## Prerequisites For Reproduction
|
8 |
-
1. **GPU**: Requires powerful GPUs
|
9 |
-
2. **
|
10 |
-
3. **
|
11 |
-
4. **
|
|
|
12 |
|
13 |
## Samples showing Adaptive Learning Capabilities
|
14 |
|
|
|
5 |
### Model Name: CodeOptimus - Supervised Instruction Fine-tuning [Mistral 7B Instruct](https://mistral.ai/news/announcing-mistral-7b/) using qLora.
|
6 |
|
7 |
## Prerequisites For Reproduction
|
8 |
+
1. **GPU**: Requires powerful GPUs - I used 7 Nvidia A100s.
|
9 |
+
2. **Train Time**: 1 week.
|
10 |
+
3. **Python Packages**: Install requirements.txt.
|
11 |
+
4. **Dataset**: Download [code_instructions_122k_alpaca_style](https://huggingface.co/datasets/TokenBender/code_instructions_122k_alpaca_style) plus some custom curated dataset
|
12 |
+
5. **Mistra-7B-Instruct-v0.1**: Download [mistralai/Mistral-7B-Instruct-v0.1 ](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) pytorch bin weights
|
13 |
|
14 |
## Samples showing Adaptive Learning Capabilities
|
15 |
|