weblab-10b / README.md
matsuo-lab's picture
Update README.md
d6fc432
---
license: cc-by-nc-4.0
---
# weblab-10b
# Overview
This repository provides a Japanese-centric multilingual GPT-NeoX model of 10 billion parameters.
* **Library**
The model was trained using code based on [EleutherAI/gpt-neox](https://github.com/EleutherAI/gpt-neox).
* **Model architecture**
A 36-layer, 4864-hidden-size transformer-based language model.
* **Pre-training**
The model was trained on around **600B** tokens from a mixture of the following corpora.
- [Japanese C4](https://huggingface.co/datasets/mc4)
- [The Pile](https://huggingface.co/datasets/EleutherAI/pile)
* **Model Series**
| Variant | Link |
| :-- | :--|
| weblab-10b-instruction-sft | https://huggingface.co/matsuo-lab/weblab-10b-instruction-sft |
| weblab-10b | https://huggingface.co/matsuo-lab/weblab-10b |
* **Authors**
Takeshi Kojima
---
# Benchmarking
* **Japanese benchmark : JGLUE 8-task (2023-08-27)**
- *We used [Stability-AI/lm-evaluation-harness](https://github.com/Stability-AI/lm-evaluation-harness/tree/2f1583c0735eacdfdfa5b7d656074b69577b6774) library for evaluation.*
- *The 8-task average accuracy is based on results of JCommonsenseQA-1.1, JNLI-1.1, MARC-ja-1.1, JSQuAD-1.1, jaqket_v2-0.2, xlsum_ja-1.0, xwinograd_ja, and mgsm-1.0.*
- *model loading is performed with float16, and evaluation is performed with template version 0.3 using the few-shot in-context learning.*
- *The number of few-shots is 3,3,3,2,1,1,0,5.*
- *special_tokens_map.json is modified to avoid errors during the evaluation of the second half benchmarks. As a result, the results of the first half benchmarks became slightly different.*
model | average | jcommonsenseqa | jnli | marc_ja | jsquad | jaqket_v2 | xlsum_ja | xwinograd_ja | mgsm
| :-- | :-- | :-- | :-- | :-- | :-- | :-- | :-- | :-- | :-- |
weblab-10b-instruction-sft | 59.11 | 74.62 | 66.56 | 95.49 | 78.34 | 63.32 | 20.57 | 71.95 | 2
weblab-10b | 50.74 | 66.58 | 53.74 | 82.07 | 62.94 | 56.19 | 10.03 | 71.95 | 2.4
* **Japanese benchmark : JGLUE 4-task (2023-08-18)**
- *We used [Stability-AI/lm-evaluation-harness](https://github.com/Stability-AI/lm-evaluation-harness/tree/2f1583c0735eacdfdfa5b7d656074b69577b6774) library for evaluation.*
- *The 4-task average accuracy is based on results of JCommonsenseQA-1.1, JNLI-1.1, MARC-ja-1.1, and JSQuAD-1.1.*
- *model loading is performed with float16, and evaluation is performed with template version 0.3 using the few-shot in-context learning.*
- *The number of few-shots is 3,3,3,2.*
| Model | Average | JCommonsenseQA | JNLI | MARC-ja | JSQuAD |
| :-- | :-- | :-- | :-- | :-- | :-- |
| weblab-10b-instruction-sft | 78.78 | 74.35 | 65.65 | 96.06 | 79.04 |
| weblab-10b | 66.38 | 65.86 | 54.19 | 84.49 | 60.98 |
---
# How to use the model
~~~~python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("matsuo-lab/weblab-10b")
model = AutoModelForCausalLM.from_pretrained("matsuo-lab/weblab-10b", torch_dtype=torch.float16)
if torch.cuda.is_available():
model = model.to("cuda")
text = "吾輩は猫である。"
token_ids = tokenizer.encode(text, add_special_tokens=False, return_tensors="pt")
with torch.no_grad():
output_ids = model.generate(
token_ids.to(model.device),
max_new_tokens=100,
do_sample=True,
temperature=0.7,
top_p=0.95
)
output = tokenizer.decode(output_ids.tolist()[0])
print(output)
~~~~
---
# Licenese
[cc-by-nc-4.0](https://creativecommons.org/licenses/by-nc/4.0/)