Commit
•
0982a20
1
Parent(s):
c19e855
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,89 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
**Code-Llama-3-8B**
|
| 6 |
+
|
| 7 |
+
|
| 8 |
+
This Model is trained on refined version of my dataset [Code-290k-ShareGPT](https://huggingface.co/datasets/ajibawa-2023/Code-290k-ShareGPT).
|
| 9 |
+
|
| 10 |
+
Besides this it is trained on following datasets:
|
| 11 |
+
|
| 12 |
+
[Code-Feedback](https://huggingface.co/datasets/m-a-p/Code-Feedback)
|
| 13 |
+
|
| 14 |
+
[orca-math-word-problems-200k](https://huggingface.co/datasets/microsoft/orca-math-word-problems-200k)
|
| 15 |
+
|
| 16 |
+
[Openhermes](https://huggingface.co/datasets/teknium/openhermes)
|
| 17 |
+
|
| 18 |
+
The idea was to check how this Model will perform with both Code & Maths datasets. This model is very good with Coding.
|
| 19 |
+
Maths outputs are also very good. You can test out this model.
|
| 20 |
+
|
| 21 |
+
This Model is trained on massive datasets so the results are very good. You can check the Examples given below.
|
| 22 |
+
I have used ChatML prompt format.
|
| 23 |
+
|
| 24 |
+
This is Fully Finetuned Model. Quantized model will be updated very soon.
|
| 25 |
+
|
| 26 |
+
**GGUF & Exllama**
|
| 27 |
+
|
| 28 |
+
GGUF: TBA
|
| 29 |
+
|
| 30 |
+
Exllama v2: TBA
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
**Training:**
|
| 36 |
+
|
| 37 |
+
Entire dataset was trained on 4 x A100 80GB. For 2 epoch, training took more than 160 Hours. Axolotl codebase was used for training purpose.
|
| 38 |
+
Entire data is trained on Llama-3-8B by Meta.
|
| 39 |
+
|
| 40 |
+
**Example Prompt:**
|
| 41 |
+
This model uses **ChatML** prompt format.
|
| 42 |
+
|
| 43 |
+
```
|
| 44 |
+
<|im_start|>system
|
| 45 |
+
You are a helpful Coding assistant.<|im_end|>
|
| 46 |
+
<|im_start|>user
|
| 47 |
+
{prompt}<|im_end|>
|
| 48 |
+
<|im_start|>assistant
|
| 49 |
+
|
| 50 |
+
```
|
| 51 |
+
You can modify above Prompt as per your requirement.
|
| 52 |
+
One example will be:
|
| 53 |
+
```
|
| 54 |
+
This is a conversation with your helpful Coding assistant. Assistant can generate Code in various Programming Languages along with necessary explanation.
|
| 55 |
+
```
|
| 56 |
+
|
| 57 |
+
I want to say special Thanks to the Open Source community for helping & guiding me to better understand the AI/Model development.
|
| 58 |
+
|
| 59 |
+
Thank you for your love & support.
|
| 60 |
+
|
| 61 |
+
|
| 62 |
+
**Example Output**
|
| 63 |
+
|
| 64 |
+
Example 1
|
| 65 |
+
|
| 66 |
+
|
| 67 |
+
|
| 68 |
+
|
| 69 |
+
Example 2
|
| 70 |
+
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
|
| 74 |
+
Example 3
|
| 75 |
+
|
| 76 |
+
|
| 77 |
+
|
| 78 |
+
|
| 79 |
+
Example 4
|
| 80 |
+
|
| 81 |
+
|
| 82 |
+
|
| 83 |
+
|
| 84 |
+
Example 5
|
| 85 |
+
|
| 86 |
+
|
| 87 |
+
|
| 88 |
+
|
| 89 |
+
|