|
--- |
|
library_name: peft |
|
base_model: DAMO-NLP-MT/polylm-1.7b |
|
--- |
|
|
|
# [Reproducing] Stanford Alpaca: An Instruction-following LLaMA Model |
|
This is the repo for reproducing [Stanford Alpaca : An Instruction-following LLaMA Model](https://github.com/tatsu-lab/stanford_alpaca/blob/main/README.md). We finetune some of LlaMa2-based large language model using medical QA dataset. The repo contains: |
|
|
|
- The [5K data](#dataset) conversations between patients and physicians used for fine-tuning the model. |
|
- The code for [Preparation data](#data-preparation). |
|
- The code for [Fine Tuning the Model](#fine-tuning). |
|
- The link for [Testing the Model](#testing-the-model). |
|
|
|
## Dataset |
|
We using the 5k generated dataset by [Chat Doctor](https://github.com/Kent0n-Li/ChatDoctor). The dataset is a generated conversations between patients and physicians from ChatGPT GenMedGPT-5k and disease database. Dataset also currated and modified to Indonesian Language Based. |
|
|
|
[`GenMedGPT-5k-id.json`](https://github.com/gilangcy/stanford-alpaca/blob/main/GenMedGPT-5k-id.json) contains 5K instruction-following data we used for fine-tuning the LlaMa model. This JSON file is a list of dictionaries, each dictionary contains the following fields: |
|
|
|
- `instruction`: `str`, describes the task the model should perform. Each of the 52K instructions is unique. |
|
- `input`: `str`, optional context or input for the task. For example, when the instruction is "Summarize the following article", the input is the article. Around 40% of the examples have an input. |
|
- `output`: `str`, the answer to the instruction as generated by `text-davinci-003`. |
|
|
|
If you're interested in fine-tuning with your own data, it's essential to adhere to the default prompt format that the model used during its pre-training phase. The prompt for LlaMa 2 is structured similarly to this: |
|
|
|
``` |
|
<s>[INST] <<SYS>> |
|
{{ instruction }} |
|
<</SYS>> |
|
|
|
{{ input }} [/INST] {{ output }} </s> |
|
``` |
|
|
|
Meanwhile, the prompt for PolyLM and InternLM (adapted to Indonesian) is structured similarly to this: |
|
|
|
``` |
|
Di bawah ini adalah instruksi yang menjelaskan tugas, dipasangkan dengan masukan yang memberikan konteks lebih lanjut. Tulis tanggapan yang melengkapi permintaan dengan tepat. |
|
|
|
Instruksi: |
|
{instruction} |
|
|
|
Masukan: |
|
{input} |
|
|
|
Tanggapan: |
|
{output} |
|
``` |
|
|
|
## Finetuning the Model |
|
We fine-tune our models based on the step from Stanford Alpaca. We choose to train some LLama-based model. The model that we finetune are PolyLM-1.7B, LlaMa-2-7B, InternLM-7B with the following hyperparameters: |
|
|
|
| Hyperparameter | PolyLM-1.7B | LLaMA-7B | InternLM-7B | |
|
|----------------|------------ |----------|-------------| |
|
| Batch size | 128 | 128 | 128 | |
|
| Learning rate | 3e-4 | 3e-4 | 3e-4 | |
|
| Epochs | 3 | 3 | 3 | |
|
| Max length | 256 | 256 | 256 | |
|
| Weight decay | 0 | 0 | 0 | |
|
|
|
To reproduce our fine-tuning runs for LLaMA, first install the requirements |
|
|
|
``` |
|
pip install -r requirements.txt |
|
``` |
|
|
|
The code for finetuning is available at [`fine-tuning.ipynb`](https://github.com/gilangcy/stanford-alpaca/blob/main/fine-tuning.ipynb) with four sections of pre-preocessing data, fine-tuning with LlaMa 2, fine-tuning with PolyLM, and fine-tuning with InternLM. |
|
|
|
## Training procedure |
|
|
|
|
|
The following `bitsandbytes` quantization config was used during training: |
|
- quant_method: bitsandbytes |
|
- load_in_8bit: True |
|
- load_in_4bit: False |
|
- llm_int8_threshold: 6.0 |
|
- llm_int8_skip_modules: None |
|
- llm_int8_enable_fp32_cpu_offload: False |
|
- llm_int8_has_fp16_weight: False |
|
- bnb_4bit_quant_type: fp4 |
|
- bnb_4bit_use_double_quant: False |
|
- bnb_4bit_compute_dtype: float32 |
|
|
|
### Framework versions |
|
|
|
|
|
- PEFT 0.6.0.dev0 |
|
|
|
|
|
## Testing the Model |
|
|
|
These are link for test the fine-tuned model : |
|
|
|
1. [PolyLM-1.7B](https://huggingface.co/spaces/dennyaw/polylm1.7b) |
|
2. [LlaMa-2-7B](https://huggingface.co/spaces/dennyaw/Llama-2-7b-finetuned) |
|
3. [InternLM-7B](https://huggingface.co/spaces/dennyaw/internlm-7b-finetuned) |
|
|
|
### Authors |
|
|
|
All interns below contributed equally and the order is determined by random draw. |
|
|
|
- [Denny Andriana Wahyu](https://www.linkedin.com/in/denny-aw/) |
|
- [Fadli Aulawi Al Ghiffari](https://www.linkedin.com/in/fadli-aulawi-al-ghiffari-9b4990148/) |
|
- [Gilang Catur Yudishtira](https://www.linkedin.com/in/gilangcy/) |
|
|
|
All advised by [Firqa Aqilla Noor Arasyi](https://www.linkedin.com/in/firqaana/) |
|
|
