|
--- |
|
license: wtfpl |
|
datasets: |
|
- HuggingFaceH4/no_robots |
|
pipeline_tag: text-generation |
|
--- |
|
|
|
# MAMBA (2.8B) 🐍 fine-tuned on OpenHerms |
|
|
|
Model Card is still WIP! |
|
|
|
|
|
## Base model info |
|
|
|
Mamba is a new state space model architecture showing promising performance on information-dense data such as language modeling, where previous subquadratic models fall short of Transformers. |
|
It is based on the line of progress on [structured state space models](https://github.com/state-spaces/s4), |
|
with an efficient hardware-aware design and implementation in the spirit of [FlashAttention](https://github.com/Dao-AILab/flash-attention). |
|
|
|
## Dataset info |
|
|
|
TBA |
|
|
|
|
|
## Usage |
|
|
|
```sh |
|
pip install transformers |
|
pip install causal-conv1d<=1.0.2 |
|
pip install mamba-ssm |
|
``` |
|
|
|
```py |
|
import torch |
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
from mamba_ssm.models.mixer_seq_simple import MambaLMHeadModel |
|
|
|
CHAT_TEMPLATE_ID = "HuggingFaceH4/zephyr-7b-beta" |
|
|
|
device = torch.device("cuda" if torch.cuda.is_available() else "cpu") |
|
model_name = "clibrain/mamba-2.8b-instruct-openhermes" |
|
|
|
eos_token = "<|endoftext|>" |
|
tokenizer = AutoTokenizer.from_pretrained(model_name) |
|
tokenizer.eos_token = eos_token |
|
tokenizer.pad_token = tokenizer.eos_token |
|
tokenizer.chat_template = AutoTokenizer.from_pretrained(CHAT_TEMPLATE_ID).chat_template |
|
|
|
model = MambaLMHeadModel.from_pretrained( |
|
model_name, device=device, dtype=torch.float16) |
|
|
|
history_dict: list[dict[str, str]] = [] |
|
prompt = "Tell me 5 sites to visit in Spain" |
|
history_dict.append(dict(role="user", content=prompt)) |
|
|
|
input_ids = tokenizer.apply_chat_template( |
|
history_dict, return_tensors="pt", add_generation_prompt=True |
|
).to(device) |
|
|
|
out = model.generate( |
|
input_ids=input_ids, |
|
max_length=2000, |
|
temperature=0.9, |
|
top_p=0.7, |
|
eos_token_id=tokenizer.eos_token_id, |
|
) |
|
|
|
decoded = tokenizer.batch_decode(out) |
|
assistant_message = ( |
|
decoded[0].split("<|assistant|>\n")[-1].replace(eos, "") |
|
) |
|
|
|
print(assistant_message) |
|
``` |
|
|
|
|
|
## Gradio Demo |
|
|
|
```sh |
|
git clone https://github.com/mrm8488/mamba-chat.git |
|
cd mamba-chat |
|
|
|
pip install -r requirements.txt |
|
pip install -q gradio==4.8.0 |
|
|
|
python app.py \ |
|
--model clibrain/mamba-2.8b-instruct-openhermes \ |
|
--share |
|
``` |
|
## Evaluations |
|
|
|
Coming soon! |
|
|
|
|
|
## Acknowledgments |
|
|
|
Thanks to [mamba-chat](https://github.com/havenhq/mamba-chat/tree/main) for heavily inspiring our work |
