README.md

---
license: llama2
train: false
inference: false
pipeline_tag: text-generation
---

This is an experimental <a href="https://github.com/mobiusml/hqq/">HQQ</a> 2-bit quantized <a href="https://huggingface.co/meta-llama/Llama-2-7b-chat-hf"> Llama2-7B-chat model </a> using a low-rank adapter to improve the performance (referred to as <a href="https://mobiusml.github.io/1bit_blog/">HQQ+</a>).  

Quantizing small models at extreme low-bits is a challenging task. The purpose of this model is to show the community what to expect when fine-tuning such models. 
We notice that, when given more specialized data, the low-bit model can even outperform the full-precision model at some tasks.

This version offloads the meta-data to the CPU, so only the 2-bit weights and the low-rank adapters are stored in the GPU memory. 

## Datasets
The adapter was trained via SFT on random subsets of the following:

### Base Model
* <a href="https://huggingface.co/datasets/wikitext">wikitext-2-raw-v1</a> (full)

### Chat Model
* <a href="https://huggingface.co/datasets/timdettmers/openassistant-guana"> timdettmers/openassistant-guanaco </a> (full)
* <a href="https://huggingface.co/datasets/icrosoft/orca-math-word-problems-200k"> microsoft/orca-math-word-problems-200k </a> (10K)
* <a href="https://huggingface.co/datasets/meta-math/MetaMathQA"> meta-math/MetaMathQA </a> (10K)
* <a href="https://huggingface.co/datasets/HuggingFaceH4/ultrafeedback_binarized"> HuggingFaceH4/ultrafeedback_binarized </a> (10K - chosen answers only)

## Performance
| Models            | Llama2-7B (fp16)| Llama2-7B (HQQ 2-bit)| Llama2-7B (HQQ+ 2-bit)| Quip# (2-bit)|
|-------------------|------------------|------------------|------------------|------------------|
| Wiki Perpexlity   | 5.18             |             6.06 |      <b>5.14</b> |         8.54     |
| VRAM (GB)         |    13.5          |      <b>2.6</b>  |    2.69          |         2.72     |
| forward time (sec)|   <b>0.1<b>      |    0.221         |     0.27         |      0.353       |

| Models            | Llama2-7B-chat (fp16)| Llama2-7B-chat (HQQ 2-bit)| Llama2-7B-chat (HQQ+ 2-bit)|
|-------------------|------------------|------------------|------------------|
| ARC (25-shot)     |    53.67         |  45.56   |  47.01  |
| HellaSwag (10-shot)|   78.56         |  73.59   |  73.74  |
| MMLU (5-shot)     |    48.16         |  43.18   | 43.33   |
| TruthfulQA-MC2    |    45.32         |  43.1    |  42.66  |
| Winogrande (5-shot)|   72.53         |   67.32  |  71.51  |
| GSM8K (5-shot)    |    23.12         |   9.7    |   28.43 |
| Average           |    53.56         |  47.08   |  51.11  |

## Usage
First, install the latest version of <a href="https://github.com/mobiusml/hqq/">HQQ</a>:
```
pip install git+https://github.com/mobiusml/hqq.git
```
Then you can use the sample code below:
``` Python
from hqq.engine.hf import HQQModelForCausalLM, AutoTokenizer

#Load the model
model_id = 'mobiuslabsgmbh/Llama-2-7b-chat-hf_2bitgs8_hqq' 
model     = HQQModelForCausalLM.from_quantized(model_id, adapter='adapter_v0.1.lora')
tokenizer = AutoTokenizer.from_pretrained(model_id)

#Setup Inference Mode
tokenizer.add_bos_token = False
tokenizer.add_eos_token = False
if not tokenizer.pad_token: tokenizer.add_special_tokens({'pad_token': '[PAD]'})
model.config.use_cache  = True
model.eval();

# Optional: torch compile for faster inference
# model = torch.compile(model)

#Streaming Inference
import torch, transformers
from threading import Thread

def chat_processor(chat, max_new_tokens=100, do_sample=True, device='cuda'):
    tokenizer.use_default_system_prompt = False
    streamer = transformers.TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)

    generate_params = dict(
        tokenizer("<s> [INST] " + chat + " [/INST] ", return_tensors="pt").to(device),
        streamer=streamer,
        max_new_tokens=max_new_tokens,
        do_sample=do_sample,
        pad_token_id=tokenizer.pad_token_id,
        top_p=0.90 if do_sample else None,
        top_k=50 if do_sample else None,
        temperature= 0.6 if do_sample else None,
        num_beams=1,
        repetition_penalty=1.2,
    )

    t = Thread(target=model.generate, kwargs=generate_params)
    t.start()
    
    print("User: ", chat); 
    print("Assistant: ");
    outputs = ""
    for text in streamer:
        outputs += text
        print(text, end="", flush=True)

    torch.cuda.empty_cache()
  
    return outputs
```

### Example
``` Python
outputs = chat_processor("If you had 5 apples yesterday and you ate 2 today morning, how many apples do you have this evening?", max_new_tokens=1000, do_sample=False)
```
```
User:  If you had 5 apples yesterday and you ate 2 today morning, how many apples do you have this evening?
Assistant: 
You started with 5 apples.You ate 2 of them so now you have 5-2=3 apples left.So by the evening you will still have 3 apples.
```