77 lines
3.2 KiB
Markdown
77 lines
3.2 KiB
Markdown
|
---
|
||
|
|
datasets:
|
||
|
|
- cognitivecomputations/dolphin
|
||
|
|
- jondurbin/airoboros-2.2.1
|
||
|
|
- cognitivecomputations/dolphin-coder
|
||
|
|
- teknium/openhermes
|
||
|
|
- ise-uiuc/Magicoder-OSS-Instruct-75K
|
||
|
|
- ise-uiuc/Magicoder-Evol-Instruct-110K
|
||
|
|
- m-a-p/Code-Feedback
|
||
|
|
- m-a-p/CodeFeedback-Filtered-Instruction
|
||
|
|
language:
|
||
|
|
- en
|
||
|
|
license: bigcode-openrail-m
|
||
|
|
---
|
||
|
|
|
||
|
|
DolphinCoder StarCoder2 15b 🐬
|
||
|
|
|
||
|
|
sponsored by [latitude.sh](https://www.latitude.sh/).
|
||
|
|
|
||
|
|
[](https://discord.gg/cognitivecomputations)
|
||
|
|
Discord: https://discord.gg/cognitivecomputations
|
||
|
|
|
||
|
|
<img src="https://cdn-uploads.huggingface.co/production/uploads/63111b2d88942700629f5771/ldkN1J0WIDQwU4vutGYiD.png" width="600" />
|
||
|
|
|
||
|
|
This model is based on StarCoder2-15b and is subject to bigcode-openrail-m license.
|
||
|
|
|
||
|
|
This Dolphin is *really good* at coding, I trained with a lot of coding data.
|
||
|
|
|
||
|
|
This model is uncensored. I have filtered the dataset to remove alignment and bias. This makes the model more compliant. You are advised to implement your own alignment layer before exposing the model as a service. It will be highly compliant to any requests, even unethical ones. Please read my blog post about uncensored models. https://erichartford.com/uncensored-models
|
||
|
|
You are responsible for any content you create using this model. Enjoy responsibly.
|
||
|
|
|
||
|
|
## Training
|
||
|
|
It took 3 days to train 3 epochs on 8x H100s using qLoRA and Axolotl
|
||
|
|
|
||
|
|
Prompt format:
|
||
|
|
This model uses ChatML prompt format.
|
||
|
|
```
|
||
|
|
<|im_start|>system
|
||
|
|
You are DolphinCoder, a helpful AI programming assistant.<|im_end|>
|
||
|
|
<|im_start|>user
|
||
|
|
{prompt}<|im_end|>
|
||
|
|
<|im_start|>assistant
|
||
|
|
|
||
|
|
```
|
||
|
|
|
||
|
|
Example:
|
||
|
|
```
|
||
|
|
<|im_start|>system
|
||
|
|
You are DolphinCoder, a master at software engineering and coding in any programming language.
|
||
|
|
<|im_start|>user
|
||
|
|
Please write me a program in golang that parses all the lines in a file, and reverses them character-wise, and saves it to a new file.
|
||
|
|
<|im_start|>assistant
|
||
|
|
```
|
||
|
|
|
||
|
|
## Quantized models
|
||
|
|
|
||
|
|
- [gguf](https://huggingface.co/dagbs/dolphincoder-starcoder2-15b-GGUF)
|
||
|
|
|
||
|
|
- [ExLlamaV2](https://huggingface.co/bartowski/dolphincoder-starcoder2-15b-exl2)
|
||
|
|
|
||
|
|
## Gratitude
|
||
|
|
- This model was made possible by the generous sponsorship of [latitude.sh](https://www.latitude.sh/).
|
||
|
|
- Huge thank you to [BigCode](https://www.bigcode-project.org/) for training and publishing the weights of StarCoder2
|
||
|
|
- HUGE Thank you to the dataset authors: @ise-uiuc, @teknium, @m-a-p
|
||
|
|
- And HUGE thanks to @winglian and the Axolotl contributors for making the best training framework!
|
||
|
|
- [<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
|
||
|
|
- Thank you to all the other people in the Open Source AI community who have taught me and helped me along the way.
|
||
|
|
|
||
|
|
## Example Output
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
[If you would like to financially support my efforts](https://ko-fi.com/erichartford)
|
||
|
|
|
||
|
|
[swag](https://fa7113.myshopify.com/)
|