File size: 3,776 Bytes
1888e1e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 |
## Model Details
This model is an int4 model with group_size 128 of [google/gemma-2b](https://huggingface.co/google/gemma-2b) generated by [intel/auto-round](https://github.com/intel/auto-round).
### Use the model
### INT4 Inference with AutoGPTQ's kernel
Install the latest [AutoGPTQ](https://github.com/AutoGPTQ/AutoGPTQ) from source first
```python
##pip install auto-gptq[triton]
##pip install triton==2.2.0
from transformers import AutoModelForCausalLM, AutoTokenizer
quantized_model_dir = "Intel/gemma-2b-int4-inc"
tokenizer = AutoTokenizer.from_pretrained(quantized_model_dir)
model = AutoModelForCausalLM.from_pretrained(quantized_model_dir,
device_map="auto",
trust_remote_code=False,
)
tokenizer = AutoTokenizer.from_pretrained(quantized_model_dir, use_fast=True)
text = "There is a girl who likes adventure,"
inputs = tokenizer(text, return_tensors="pt").to(model.device)
print(tokenizer.decode(model.generate(**inputs, max_new_tokens=50, do_sample=True)[0]))
```
### Evaluate the model
Install [lm-eval-harness](https://github.com/EleutherAI/lm-evaluation-harness.git) from source, and the git id we used is 96d185fa6232a5ab685ba7c43e45d1dbb3bb906d
pip install auto-gptq[triton]
pip install triton==2.2.0
Please note that there is a discrepancy between the baseline result and the official data, which is a known issue within the official model card community.
```bash
lm_eval --model hf --model_args pretrained="Intel/gemma-2b-int4-inc",autogptq=True,gptq_use_triton=True --device cuda:0 --tasks lambada_openai,hellaswag,piqa,winogrande,truthfulqa_mc1,openbookqa,boolq,rte,arc_easy,arc_challenge,mmlu --batch_size 16
```
| Metric | FP16 | int4 |
| -------------- | ------ | ------ |
| Avg. | 0.5383 | 0.5338 |
| mmlu | 0.3337 | 0.3276 |
| lambada_openai | 0.6398 | 0.6319 |
| hellaswag | 0.5271 | 0.5161 |
| winogrande | 0.6472 | 0.6472 |
| piqa | 0.7699 | 0.7622 |
| truthfulqa_mc1 | 0.2203 | 0.2191 |
| openbookqa | 0.3020 | 0.2980 |
| boolq | 0.6939 | 0.6939 |
| rte | 0.6426 | 0.6498 |
| arc_easy | 0.7424 | 0.7348 |
| arc_challenge | 0.4019 | 0.3908 |
### Reproduce the model
Here is the sample command to reproduce the model
```bash
git clone https://github.com/intel/auto-round
cd auto-round/examples/language-modeling
pip install -r requirements.txt
python3 main.py \
--model_name google/gemma-2b \
--device 0 \
--group_size 128 \
--bits 4 \
--iters 400 \
--use_quant_input \
--deployment_device 'gpu' \
--output_dir "./tmp_autoround"
```
## Ethical Considerations and Limitations
The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs.
Therefore, before deploying any applications of the model, developers should perform safety testing.
## Caveats and Recommendations
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.
Here are a couple of useful links to learn more about Intel's AI software:
* Intel Neural Compressor [link](https://github.com/intel/neural-compressor)
* Intel Extension for Transformers [link](https://github.com/intel/intel-extension-for-transformers)
## Disclaimer
The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.
|