---
license: other
tags:
- axolotl
- finetune
- qlora
base_model: openchat/openchat-3.5-0106
datasets:
- hendrycks/competition_math
- allenai/ai2_arc
- camel-ai/physics
- camel-ai/chemistry
- camel-ai/biology
- camel-ai/math
- STEM-AI-mtl/Electrical-engineering
- openbookqa
- piqa
- metaeval/reclor
- mandyyyyii/scibench
- derek-thomas/ScienceQA
- sciq
- TIGER-Lab/ScienceEval
---
# ๐ฌ๐ฉโ๐ฌ Newton-7B
This model is a fine-tuned version of [openchat/openchat-3.5-0106](https://huggingface.co/openchat/openchat-3.5-0106) on datasets related to science.
This model is fine-tuned using [QLoRa](https://arxiv.org/abs/2305.14314) and [axolotl](https://github.com/OpenAccess-AI-Collective/axolotl).
This model's training was sponsored by [sablo.ai](https://sablo.ai).
See axolotl config
axolotl version: `0.3.0`
```yaml
base_model: openchat/openchat-3.5-0106
model_type: MistralForCausalLM
tokenizer_type: LlamaTokenizer
is_mistral_derived_model: true
load_in_8bit: false
load_in_4bit: true
strict: false
datasets:
- path: merged_all.json
type:
field_instruction: instruction
field_output: output
format: "GPT4 Correct User: {instruction}<|end_of_turn|>GPT4 Correct Assistant:"
no_input_format: "GPT4 Correct User: {instruction}<|end_of_turn|>GPT4 Correct Assistant:"
dataset_prepared_path: last_run_prepared
val_set_size: 0.01 # not sure
output_dir: ./newton
adapter: qlora
lora_model_dir:
sequence_len: 8192
sample_packing: true
pad_to_sequence_len: true
lora_r: 128
lora_alpha: 64
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:
lora_target_modules:
- gate_proj
- down_proj
- up_proj
- q_proj
- v_proj
- k_proj
- o_proj
lora_modules_to_save:
- embed_tokens
- lm_head
wandb_project: huggingface
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:
hub_model_id: Weyaxi/newton-lora
save_safetensors: true
# change #
gradient_accumulation_steps: 12
micro_batch_size: 6
num_epochs: 2
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002
# change #
train_on_inputs: false
group_by_length: false
bf16: true
fp16: false
tf32: false
gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
warmup_steps: 10 # not sure
saves_per_epoch: 2
evals_per_epoch: 4
eval_table_size:
eval_table_max_new_tokens: 128
debug:
deepspeed:
weight_decay: 0.1 # not sure
fsdp:
fsdp_config:
special_tokens:
bos_token: ""
eos_token: ""
unk_token: ""
tokens:
- "<|end_of_turn|>"
- "<|pad_0|>"
```
# ๐ Datasets
You can find the dataset I used and the work I am doing with this datasets here:
https://huggingface.co/datasets/Weyaxi/sci-datasets
Following datasets were used in this model:
- ๐ [MATH](https://huggingface.co/datasets/hendrycks/competition_math)
- ๐ง [ARC](https://huggingface.co/datasets/allenai/ai2_arc) (Note: Only **train** part)
- ๐งฒ [camel-ai/physics](https://huggingface.co/datasets/camel-ai/physics)
- โ๏ธ [camel-ai/chemistry](https://huggingface.co/datasets/camel-ai/chemistry)
- ๐ฆ [camel-ai/biology](https://huggingface.co/datasets/camel-ai/biology)
- ๐ [camel-ai/math](https://huggingface.co/datasets/camel-ai/math)
- โก [STEM-AI-mtl/Electrical-engineering](https://huggingface.co/datasets/STEM-AI-mtl/Electrical-engineering)
- ๐ [openbookqa](https://huggingface.co/datasets/openbookqa)
- ๐ง [piqa](https://huggingface.co/datasets/piqa)
- ๐จ [reclor](https://huggingface.co/datasets/metaeval/reclor)
- ๐ฌ [scibench](https://github.com/mandyyyyii/scibench)
- ๐งช [ScienceQA](https://huggingface.co/datasets/derek-thomas/ScienceQA)
- ๐งฌ [sciq](https://huggingface.co/datasets/sciq)
- ๐ [ScienceEval](https://huggingface.co/datasets/TIGER-Lab/ScienceEval)
## ๐ ๏ธ Multiple Choice Question & Answer Datasets Conversion Progress
I used [mistralai/Mixtral-8x7B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1) to generate a reasonable and logical answer by providing it with the question and the answer key.
I used the [Together AI](https://www.together.ai) API for this task.
The following datasets are converted using this method:
- ๐ง [ARC](https://huggingface.co/datasets/allenai/ai2_arc) (Note: Only **train** part)
- ๐ [openbookqa](https://huggingface.co/datasets/openbookqa)
- ๐จ [reclor](https://huggingface.co/datasets/metaeval/reclor)
- ๐งฌ [sciq](https://huggingface.co/datasets/sciq)
# ๐ฌ Prompt Template
You can use this prompt template while using the model:
### GPT4 Correct [(Openchat)](https://huggingface.co/openchat/openchat-3.5-0106#conversation-templates)
```
GPT4 Correct User: {user}<|end_of_turn|>GPT4 Correct Assistant: {asistant}<|end_of_turn|>GPT4 Correct User: {user}<|end_of_turn|>GPT4 Correct Assistant:
```
You can also utilize the chat template method from the tokenizer config like here:
```python
messages = [
{"role": "user", "content": "Hello"},
{"role": "assistant", "content": "Hi"},
{"role": "user", "content": "How are you today?"}
]
tokens = tokenizer.apply_chat_template(messages, add_generation_prompt=True)
```
# ๐ค Acknowledgments
Thanks to [openchat](https://huggingface.co/openchat) team for fine-tuning an excellent model that I used as a base model.
Thanks to [@jondurbin](https://huggingface.co/jondurbin) for reformatting codes for some datasets: [bagel/data_sources](https://github.com/jondurbin/bagel/tree/main/bagel/data_sources)
Thanks to [Together AI](https://www.together.ai) for providing everyone with free credits, which I used to generate a dataset in multiple choice to explanations format.
Thanks to [Tim Dettmers](https://huggingface.co/timdettmers) for his excellent [QLoRA](https://arxiv.org/abs/2305.14314) work.
Thanks to all the dataset authors mentioned in the datasets section.
Thanks to [axolotl](https://github.com/OpenAccess-AI-Collective/axolotl) for making the repository I used to make this model.
Overall, thanks to all of the open soure AI community! ๐
[
](https://github.com/OpenAccess-AI-Collective/axolotl)
If you would like to support me:
[โ Buy Me a Coffee](https://www.buymeacoffee.com/weyaxi)