EleutherAI/polyglot-ko-3.8b
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: EleutherAI/polyglot-ko-3.8b
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-05-21 09:12:12 +08:00
commit a89ebbf245
18 changed files with 823 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

49
.gitattributes vendored Normal file
View File

@@ -0,0 +1,49 @@
*.7z filter=lfs diff=lfs merge=lfs -text
*.arrow filter=lfs diff=lfs merge=lfs -text
*.bin filter=lfs diff=lfs merge=lfs -text
*.bin.* filter=lfs diff=lfs merge=lfs -text
*.bz2 filter=lfs diff=lfs merge=lfs -text
*.ftz filter=lfs diff=lfs merge=lfs -text
*.gz filter=lfs diff=lfs merge=lfs -text
*.h5 filter=lfs diff=lfs merge=lfs -text
*.joblib filter=lfs diff=lfs merge=lfs -text
*.lfs.* filter=lfs diff=lfs merge=lfs -text
*.model filter=lfs diff=lfs merge=lfs -text
*.msgpack filter=lfs diff=lfs merge=lfs -text
*.onnx filter=lfs diff=lfs merge=lfs -text
*.ot filter=lfs diff=lfs merge=lfs -text
*.parquet filter=lfs diff=lfs merge=lfs -text
*.pb filter=lfs diff=lfs merge=lfs -text
*.pt filter=lfs diff=lfs merge=lfs -text
*.pth filter=lfs diff=lfs merge=lfs -text
*.rar filter=lfs diff=lfs merge=lfs -text
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
*.tar.* filter=lfs diff=lfs merge=lfs -text
*.tflite filter=lfs diff=lfs merge=lfs -text
*.tgz filter=lfs diff=lfs merge=lfs -text
*.xz filter=lfs diff=lfs merge=lfs -text
*.zip filter=lfs diff=lfs merge=lfs -text
*.zstandard filter=lfs diff=lfs merge=lfs -text
*.tfevents* filter=lfs diff=lfs merge=lfs -text
*.db* filter=lfs diff=lfs merge=lfs -text
*.ark* filter=lfs diff=lfs merge=lfs -text
**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text
**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text
**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text
*.safetensors filter=lfs diff=lfs merge=lfs -text
*.ckpt filter=lfs diff=lfs merge=lfs -text
*.gguf* filter=lfs diff=lfs merge=lfs -text
*.ggml filter=lfs diff=lfs merge=lfs -text
*.llamafile* filter=lfs diff=lfs merge=lfs -text
*.pt2 filter=lfs diff=lfs merge=lfs -text
*.mlmodel filter=lfs diff=lfs merge=lfs -text
*.npy filter=lfs diff=lfs merge=lfs -text
*.npz filter=lfs diff=lfs merge=lfs -text
*.pickle filter=lfs diff=lfs merge=lfs -text
*.pkl filter=lfs diff=lfs merge=lfs -text
*.tar filter=lfs diff=lfs merge=lfs -text
*.wasm filter=lfs diff=lfs merge=lfs -text
*.zst filter=lfs diff=lfs merge=lfs -text
*tfevents* filter=lfs diff=lfs merge=lfs -text
tokenizer.json filter=lfs diff=lfs merge=lfs -text

203
README.md Normal file
View File

@@ -0,0 +1,203 @@
---
language:
- ko
tags:
- pytorch
- causal-lm
license: apache-2.0
---
# Polyglot-Ko-3.8B
## Model Description
Polyglot-Ko is a series of large-scale Korean autoregressive language models made by the EleutherAI polyglot team.
| Hyperparameter | Value |
|----------------------|----------------------------------------------------------------------------------------------------------------------------------------|
| \\(n_{parameters}\\) | 3,809,974,272 |
| \\(n_{layers}\\) | 32 |
| \\(d_{model}\\) | 3,072 |
| \\(d_{ff}\\) | 12,288 |
| \\(n_{heads}\\) | 24 |
| \\(d_{head}\\) | 128 |
| \\(n_{ctx}\\) | 2,048 |
| \\(n_{vocab}\\) | 30,003 / 30,080 |
| Positional Encoding | [Rotary Position Embedding (RoPE)](https://arxiv.org/abs/2104.09864) |
| RoPE Dimensions | [64](https://github.com/kingoflolz/mesh-transformer-jax/blob/f2aa66e0925de6593dcbb70e72399b97b4130482/mesh_transformer/layers.py#L223) |
The model consists of 32 transformer layers with a model dimension of 3072, and a feedforward dimension of 12288. The model
dimension is split into 24 heads, each with a dimension of 128. Rotary Position Embedding (RoPE) is applied to 64
dimensions of each head. The model is trained with a tokenization vocabulary of 30003.
## Training data
Polyglot-Ko-3.8B was trained on 863 GB of Korean language data (1.2TB before processing), a large-scale dataset curated by [TUNiB](https://tunib.ai/). The data collection process has abided by South Korean laws. This dataset was collected for the purpose of training Polyglot-Ko models, so it will not be released for public use.
| Source |Size (GB) | Link |
|-------------------------------------|---------|------------------------------------------|
| Korean blog posts | 682.3 | - |
| Korean news dataset | 87.0 | - |
| Modu corpus | 26.4 |corpus.korean.go.kr |
| Korean patent dataset | 19.0 | - |
| Korean Q & A dataset | 18.1 | - |
| KcBert dataset | 12.7 | github.com/Beomi/KcBERT |
| Korean fiction dataset | 6.1 | - |
| Korean online comments | 4.2 | - |
| Korean wikipedia | 1.4 | ko.wikipedia.org |
| Clova call | < 1.0 | github.com/clovaai/ClovaCall |
| Naver sentiment movie corpus | < 1.0 | github.com/e9t/nsmc |
| Korean hate speech dataset | < 1.0 | - |
| Open subtitles | < 1.0 | opus.nlpl.eu/OpenSubtitles.php |
| AIHub various tasks datasets | < 1.0 |aihub.or.kr |
| Standard Korean language dictionary | < 1.0 | stdict.korean.go.kr/main/main.do |
Furthermore, in order to avoid the model memorizing and generating personally identifiable information (PII) in the training data, we masked out the following sensitive information in the pre-processing stage:
* `<|acc|>` : bank account number
* `<|rrn|>` : resident registration number
* `<|tell|>` : phone number
## Training procedure
Polyglot-Ko-3.8B was trained for 219 billion tokens over 105,000 steps on 256 A100 GPUs with the [GPT-NeoX framework](https://github.com/EleutherAI/gpt-neox). It was trained as an autoregressive language model, using cross-entropy loss to maximize the likelihood of predicting the next token.
## How to use
This model can be easily loaded using the `AutoModelForCausalLM` class:
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("EleutherAI/polyglot-ko-3.8b")
model = AutoModelForCausalLM.from_pretrained("EleutherAI/polyglot-ko-3.8b")
```
## Evaluation results
We evaluate Polyglot-Ko-3.8B on [KOBEST dataset](https://arxiv.org/abs/2204.04541), a benchmark with 5 downstream tasks, against comparable models such as skt/ko-gpt-trinity-1.2B-v0.5, kakaobrain/kogpt and facebook/xglm-7.5B, using the prompts provided in the paper.
The following tables show the results when the number of few-shot examples differ. You can reproduce these results using the [polyglot branch of lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness/tree/polyglot) and the following scripts. For a fair comparison, all models were run under the same conditions and using the same prompts. In the tables, `n` refers to the number of few-shot examples.
In case of WiC dataset, all models show random performance.
```console
python main.py \
--model gpt2 \
--model_args pretrained='EleutherAI/polyglot-ko-3.8b' \
--tasks kobest_copa,kobest_hellaswag \
--num_fewshot $YOUR_NUM_FEWSHOT \
--batch_size $YOUR_BATCH_SIZE \
--device $YOUR_DEVICE \
--output_path $/path/to/output/
```
### COPA (F1)
| Model | params | 0-shot | 5-shot | 10-shot | 50-shot |
|----------------------------------------------------------------------------------------------|--------|--------|--------|---------|---------|
| [skt/ko-gpt-trinity-1.2B-v0.5](https://huggingface.co/skt/ko-gpt-trinity-1.2B-v0.5) | 1.2B | 0.6696 | 0.6477 | 0.6419 | 0.6514 |
| [kakaobrain/kogpt](https://huggingface.co/kakaobrain/kogpt) | 6.0B | 0.7345 | 0.7287 | 0.7277 | 0.7479 |
| [facebook/xglm-7.5B](https://huggingface.co/facebook/xglm-7.5B) | 7.5B | 0.6723 | 0.6731 | 0.6769 | 0.7119 |
| [EleutherAI/polyglot-ko-1.3b](https://huggingface.co/EleutherAI/polyglot-ko-1.3b) | 1.3B | 0.7196 | 0.7193 | 0.7204 | 0.7206 |
| **[EleutherAI/polyglot-ko-3.8b](https://huggingface.co/EleutherAI/polyglot-ko-3.8b) (this)** | **3.8B** | **0.7595** | **0.7608** | **0.7638** | **0.7788** |
| [EleutherAI/polyglot-ko-5.8b](https://huggingface.co/EleutherAI/polyglot-ko-5.8b) | 5.8B | 0.7745 | 0.7676 | 0.7775 | 0.7887 |
| [EleutherAI/polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 12.8B | 0.7937 | 0.8108 | 0.8037 | 0.8369 |
<img src="https://github.com/EleutherAI/polyglot/assets/19511788/d5b49364-aed5-4467-bae2-5a322c8e2ceb" width="800px">
### HellaSwag (F1)
| Model | params | 0-shot | 5-shot | 10-shot | 50-shot |
|----------------------------------------------------------------------------------------------|--------|--------|--------|---------|---------|
| [skt/ko-gpt-trinity-1.2B-v0.5](https://huggingface.co/skt/ko-gpt-trinity-1.2B-v0.5) | 1.2B | 0.5243 | 0.5272 | 0.5166 | 0.5352 |
| [kakaobrain/kogpt](https://huggingface.co/kakaobrain/kogpt) | 6.0B | 0.5590 | 0.5833 | 0.5828 | 0.5907 |
| [facebook/xglm-7.5B](https://huggingface.co/facebook/xglm-7.5B) | 7.5B | 0.5665 | 0.5689 | 0.5565 | 0.5622 |
| [EleutherAI/polyglot-ko-1.3b](https://huggingface.co/EleutherAI/polyglot-ko-1.3b) | 1.3B | 0.5247 | 0.5260 | 0.5278 | 0.5427 |
| **[EleutherAI/polyglot-ko-3.8b](https://huggingface.co/EleutherAI/polyglot-ko-3.8b) (this)** | **3.8B** | **0.5707** | **0.5830** | **0.5670** | **0.5787** |
| [EleutherAI/polyglot-ko-5.8b](https://huggingface.co/EleutherAI/polyglot-ko-5.8b) | 5.8B | 0.5976 | 0.5998 | 0.5979 | 0.6208 |
| [EleutherAI/polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 12.8B | 0.5954 | 0.6306 | 0.6098 | 0.6118 |
<img src="https://github.com/EleutherAI/polyglot/assets/19511788/5acb60ac-161a-4ab3-a296-db4442e08b7f" width="800px">
### BoolQ (F1)
| Model | params | 0-shot | 5-shot | 10-shot | 50-shot |
|----------------------------------------------------------------------------------------------|--------|--------|--------|---------|---------|
| [skt/ko-gpt-trinity-1.2B-v0.5](https://huggingface.co/skt/ko-gpt-trinity-1.2B-v0.5) | 1.2B | 0.3356 | 0.4014 | 0.3640 | 0.3560 |
| [kakaobrain/kogpt](https://huggingface.co/kakaobrain/kogpt) | 6.0B | 0.4514 | 0.5981 | 0.5499 | 0.5202 |
| [facebook/xglm-7.5B](https://huggingface.co/facebook/xglm-7.5B) | 7.5B | 0.4464 | 0.3324 | 0.3324 | 0.3324 |
| [EleutherAI/polyglot-ko-1.3b](https://huggingface.co/EleutherAI/polyglot-ko-1.3b) | 1.3B | 0.3552 | 0.4751 | 0.4109 | 0.4038 |
| **[EleutherAI/polyglot-ko-3.8b](https://huggingface.co/EleutherAI/polyglot-ko-3.8b) (this)** | **3.8B** | **0.4320** | **0.5263** | **0.4930** | **0.4038** |
| [EleutherAI/polyglot-ko-5.8b](https://huggingface.co/EleutherAI/polyglot-ko-5.8b) | 5.8B | 0.4356 | 0.5698 | 0.5187 | 0.5236 |
| [EleutherAI/polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 12.8B | 0.4818 | 0.6041 | 0.6289 | 0.6448 |
<img src="https://github.com/EleutherAI/polyglot/assets/19511788/b74c23c0-01f3-4b68-9e10-a48e9aa052ab" width="800px">
### SentiNeg (F1)
| Model | params | 0-shot | 5-shot | 10-shot | 50-shot |
|----------------------------------------------------------------------------------------------|--------|--------|--------|---------|---------|
| [skt/ko-gpt-trinity-1.2B-v0.5](https://huggingface.co/skt/ko-gpt-trinity-1.2B-v0.5) | 1.2B | 0.6065 | 0.6878 | 0.7280 | 0.8413 |
| [kakaobrain/kogpt](https://huggingface.co/kakaobrain/kogpt) | 6.0B | 0.3747 | 0.8942 | 0.9294 | 0.9698 |
| [facebook/xglm-7.5B](https://huggingface.co/facebook/xglm-7.5B) | 7.5B | 0.3578 | 0.4471 | 0.3964 | 0.5271 |
| [EleutherAI/polyglot-ko-1.3b](https://huggingface.co/EleutherAI/polyglot-ko-1.3b) | 1.3B | 0.6790 | 0.6257 | 0.5514 | 0.7851 |
| **[EleutherAI/polyglot-ko-3.8b](https://huggingface.co/EleutherAI/polyglot-ko-3.8b) (this)** | **3.8B** | **0.4858** | **0.7950** | **0.7320** | **0.7851** |
| [EleutherAI/polyglot-ko-5.8b](https://huggingface.co/EleutherAI/polyglot-ko-5.8b) | 5.8B | 0.3394 | 0.8841 | 0.8808 | 0.9521 |
| [EleutherAI/polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 12.8B | 0.9117 | 0.9015 | 0.9345 | 0.9723 |
<img src="https://github.com/EleutherAI/polyglot/assets/19511788/95b56b19-d349-4b70-9ff9-94a5560f89ee" width="800px">
### WiC (F1)
| Model | params | 0-shot | 5-shot | 10-shot | 50-shot |
|----------------------------------------------------------------------------------------------|--------|--------|--------|---------|---------|
| [skt/ko-gpt-trinity-1.2B-v0.5](https://huggingface.co/skt/ko-gpt-trinity-1.2B-v0.5) | 1.2B | 0.3290 | 0.4313 | 0.4001 | 0.3621 |
| [kakaobrain/kogpt](https://huggingface.co/kakaobrain/kogpt) | 6.0B | 0.3526 | 0.4775 | 0.4358 | 0.4061 |
| [facebook/xglm-7.5B](https://huggingface.co/facebook/xglm-7.5B) | 7.5B | 0.3280 | 0.4903 | 0.4945 | 0.3656 |
| [EleutherAI/polyglot-ko-1.3b](https://huggingface.co/EleutherAI/polyglot-ko-1.3b) | 1.3B | 0.3297 | 0.4850 | 0.4650 | 0.3290 |
| **[EleutherAI/polyglot-ko-3.8b](https://huggingface.co/EleutherAI/polyglot-ko-3.8b) (this)** | **3.8B** | **0.3390** | **0.4944** | **0.4203** | **0.3835** |
| [EleutherAI/polyglot-ko-5.8b](https://huggingface.co/EleutherAI/polyglot-ko-5.8b) | 5.8B | 0.3913 | 0.4688 | 0.4189 | 0.3910 |
| [EleutherAI/polyglot-ko-12.8b](https://huggingface.co/EleutherAI/polyglot-ko-12.8b) | 12.8B | 0.3985 | 0.3683 | 0.3307 | 0.3273 |
<img src="https://github.com/EleutherAI/polyglot/assets/19511788/4de4a4c3-d7ac-4e04-8b0c-0d533fe88294" width="800px">
## Limitations and Biases
Polyglot-Ko has been trained to optimize next token prediction. Language models such as this are often used for a wide variety of tasks and it is important to be aware of possible unexpected outcomes. For instance, Polyglot-Ko will not always return the most factual or accurate response but the most statistically likely one. In addition, Polyglot may produce socially unacceptable or offensive content. We recommend having a human curator or other filtering mechanism to censor sensitive content.
## Citation and Related Information
### BibTeX entry
If you find our work useful, please consider citing:
```bibtex
@misc{ko2023technical,
title={A Technical Report for Polyglot-Ko: Open-Source Large-Scale Korean Language Models},
author={Hyunwoong Ko and Kichang Yang and Minho Ryu and Taekyoon Choi and Seungmu Yang and jiwung Hyun and Sungho Park},
year={2023},
eprint={2306.02254},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
### Licensing
All our models are licensed under the terms of the Apache License 2.0.
```
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
```
### Acknowledgement
This project was made possible thanks to the computing resources from [Stability.ai](https://stability.ai), and thanks to [TUNiB](https://tunib.ai) for providing a large-scale Korean dataset for this work.

26
config.json Normal file
View File

@@ -0,0 +1,26 @@
{
"_name_or_path": "./polyglot-ko-3.8b/",
"architectures": [
"GPTNeoXForCausalLM"
],
"bos_token_id": 0,
"classifier_dropout": 0.1,
"eos_token_id": 2,
"hidden_act": "gelu",
"hidden_size": 3072,
"initializer_range": 0.02,
"intermediate_size": 12288,
"layer_norm_eps": 1e-05,
"max_position_embeddings": 2048,
"model_type": "gpt_neox",
"num_attention_heads": 24,
"num_hidden_layers": 32,
"rotary_emb_base": 10000,
"rotary_pct": 0.5,
"tie_word_embeddings": false,
"torch_dtype": "float16",
"transformers_version": "4.29.2",
"use_cache": true,
"use_parallel_residual": true,
"vocab_size": 30080
}

1
configuration.json Normal file
View File

@@ -0,0 +1 @@
{"framework": "pytorch", "task": "text-generation", "allow_remote": true}

6
generation_config.json Normal file
View File

@@ -0,0 +1,6 @@
{
"_from_model_config": true,
"bos_token_id": 0,
"eos_token_id": 2,
"transformers_version": "4.29.2"
}

3
model-00001-of-00008.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:eb686e62d8f44c95de8a6cfe1a0c70fd21a80cf8853ad1557a7655e792548691
size 956859264

3
model-00002-of-00008.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:d906f3c3cfd9bffcb05cc95bbb9daf7ceb3cbeed08e9de4ce46c01230d73e38f
size 998596040

3
model-00003-of-00008.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:0e4f72d41324f36aea931311bceaf685fbde30ceb5de9eb03ad43c6d0cbf20ad
size 1002797426

3
model-00004-of-00008.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:dcaf868ec73c80c57644946344e1be703e9cde961f10eb7f6db89b5ae19ccbde
size 998596336

3
model-00005-of-00008.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:e140c307650709d0e070c6afbc5a6eca734a114eb436ecd4381f072efcf271c3
size 998596096

3
model-00006-of-00008.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:7ba948db6174d75df12ae550692cbedbdb990939247b6ea3a3cabd4e1661fc5a
size 1002797458

3
model-00007-of-00008.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:89054f5b0f9af5abd3aff9857cdc1365d5dc3f77b89f965c956325e663a1d389
size 998596336

3
model-00008-of-00008.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:efcafff4d10a4dec766a6e20d5830ac907b42dccab7cf9c60c3ad7aeb30de849
size 797387164

491
model.safetensors.index.json Normal file
View File

@@ -0,0 +1,491 @@
{
"metadata": {
"total_size": 7636729920.0
},
"weight_map": {
"embed_out.weight": "model-00008-of-00008.safetensors",
"gpt_neox.embed_in.weight": "model-00001-of-00008.safetensors",
"gpt_neox.final_layer_norm.bias": "model-00008-of-00008.safetensors",
"gpt_neox.final_layer_norm.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.0.attention.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.attention.dense.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.attention.dense.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.attention.masked_bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.attention.query_key_value.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.attention.query_key_value.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.attention.rotary_emb.inv_freq": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.input_layernorm.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.input_layernorm.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.mlp.dense_4h_to_h.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.mlp.dense_4h_to_h.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.mlp.dense_h_to_4h.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.mlp.dense_h_to_4h.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.post_attention_layernorm.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.0.post_attention_layernorm.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.attention.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.attention.dense.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.attention.dense.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.attention.masked_bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.attention.query_key_value.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.attention.query_key_value.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.attention.rotary_emb.inv_freq": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.input_layernorm.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.input_layernorm.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.mlp.dense_4h_to_h.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.mlp.dense_4h_to_h.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.mlp.dense_h_to_4h.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.mlp.dense_h_to_4h.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.post_attention_layernorm.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.1.post_attention_layernorm.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.10.attention.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.attention.dense.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.attention.dense.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.attention.masked_bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.attention.query_key_value.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.attention.query_key_value.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.attention.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.input_layernorm.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.input_layernorm.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.mlp.dense_4h_to_h.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.mlp.dense_4h_to_h.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.mlp.dense_h_to_4h.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.mlp.dense_h_to_4h.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.post_attention_layernorm.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.10.post_attention_layernorm.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.attention.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.attention.dense.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.attention.dense.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.attention.masked_bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.attention.query_key_value.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.attention.query_key_value.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.attention.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.input_layernorm.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.input_layernorm.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.mlp.dense_4h_to_h.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.mlp.dense_4h_to_h.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.mlp.dense_h_to_4h.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.mlp.dense_h_to_4h.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.post_attention_layernorm.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.11.post_attention_layernorm.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.12.attention.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.12.attention.dense.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.12.attention.dense.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.12.attention.masked_bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.12.attention.query_key_value.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.12.attention.query_key_value.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.12.attention.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
"gpt_neox.layers.12.input_layernorm.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.12.input_layernorm.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.12.mlp.dense_4h_to_h.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.12.mlp.dense_4h_to_h.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.12.mlp.dense_h_to_4h.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.12.mlp.dense_h_to_4h.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.12.post_attention_layernorm.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.12.post_attention_layernorm.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.13.attention.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.attention.dense.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.attention.dense.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.attention.masked_bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.attention.query_key_value.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.attention.query_key_value.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.attention.rotary_emb.inv_freq": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.input_layernorm.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.input_layernorm.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.mlp.dense_4h_to_h.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.mlp.dense_4h_to_h.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.mlp.dense_h_to_4h.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.mlp.dense_h_to_4h.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.post_attention_layernorm.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.13.post_attention_layernorm.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.attention.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.attention.dense.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.attention.dense.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.attention.masked_bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.attention.query_key_value.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.attention.query_key_value.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.attention.rotary_emb.inv_freq": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.input_layernorm.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.input_layernorm.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.mlp.dense_4h_to_h.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.mlp.dense_4h_to_h.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.mlp.dense_h_to_4h.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.mlp.dense_h_to_4h.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.post_attention_layernorm.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.14.post_attention_layernorm.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.attention.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.attention.dense.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.attention.dense.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.attention.masked_bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.attention.query_key_value.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.attention.query_key_value.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.attention.rotary_emb.inv_freq": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.input_layernorm.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.input_layernorm.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.mlp.dense_4h_to_h.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.mlp.dense_4h_to_h.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.mlp.dense_h_to_4h.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.mlp.dense_h_to_4h.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.post_attention_layernorm.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.15.post_attention_layernorm.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.attention.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.attention.dense.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.attention.dense.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.attention.masked_bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.attention.query_key_value.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.attention.query_key_value.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.attention.rotary_emb.inv_freq": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.input_layernorm.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.input_layernorm.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.mlp.dense_4h_to_h.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.16.mlp.dense_4h_to_h.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.16.mlp.dense_h_to_4h.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.16.mlp.dense_h_to_4h.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.16.post_attention_layernorm.bias": "model-00004-of-00008.safetensors",
"gpt_neox.layers.16.post_attention_layernorm.weight": "model-00004-of-00008.safetensors",
"gpt_neox.layers.17.attention.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.attention.dense.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.attention.dense.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.attention.masked_bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.attention.query_key_value.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.attention.query_key_value.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.attention.rotary_emb.inv_freq": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.input_layernorm.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.input_layernorm.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.mlp.dense_4h_to_h.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.mlp.dense_4h_to_h.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.mlp.dense_h_to_4h.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.mlp.dense_h_to_4h.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.post_attention_layernorm.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.17.post_attention_layernorm.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.attention.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.attention.dense.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.attention.dense.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.attention.masked_bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.attention.query_key_value.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.attention.query_key_value.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.attention.rotary_emb.inv_freq": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.input_layernorm.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.input_layernorm.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.mlp.dense_4h_to_h.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.mlp.dense_4h_to_h.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.mlp.dense_h_to_4h.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.mlp.dense_h_to_4h.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.post_attention_layernorm.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.18.post_attention_layernorm.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.attention.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.attention.dense.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.attention.dense.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.attention.masked_bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.attention.query_key_value.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.attention.query_key_value.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.attention.rotary_emb.inv_freq": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.input_layernorm.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.input_layernorm.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.mlp.dense_4h_to_h.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.mlp.dense_4h_to_h.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.mlp.dense_h_to_4h.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.mlp.dense_h_to_4h.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.post_attention_layernorm.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.19.post_attention_layernorm.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.2.attention.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.attention.dense.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.attention.dense.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.attention.masked_bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.attention.query_key_value.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.attention.query_key_value.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.attention.rotary_emb.inv_freq": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.input_layernorm.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.input_layernorm.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.mlp.dense_4h_to_h.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.mlp.dense_4h_to_h.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.mlp.dense_h_to_4h.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.mlp.dense_h_to_4h.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.post_attention_layernorm.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.2.post_attention_layernorm.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.20.attention.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.attention.dense.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.attention.dense.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.attention.masked_bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.attention.query_key_value.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.attention.query_key_value.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.attention.rotary_emb.inv_freq": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.input_layernorm.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.input_layernorm.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.mlp.dense_4h_to_h.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.20.mlp.dense_4h_to_h.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.20.mlp.dense_h_to_4h.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.mlp.dense_h_to_4h.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.post_attention_layernorm.bias": "model-00005-of-00008.safetensors",
"gpt_neox.layers.20.post_attention_layernorm.weight": "model-00005-of-00008.safetensors",
"gpt_neox.layers.21.attention.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.attention.dense.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.attention.dense.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.attention.masked_bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.attention.query_key_value.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.attention.query_key_value.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.attention.rotary_emb.inv_freq": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.input_layernorm.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.input_layernorm.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.mlp.dense_4h_to_h.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.mlp.dense_4h_to_h.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.mlp.dense_h_to_4h.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.mlp.dense_h_to_4h.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.post_attention_layernorm.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.21.post_attention_layernorm.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.attention.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.attention.dense.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.attention.dense.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.attention.masked_bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.attention.query_key_value.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.attention.query_key_value.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.attention.rotary_emb.inv_freq": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.input_layernorm.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.input_layernorm.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.mlp.dense_4h_to_h.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.mlp.dense_4h_to_h.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.mlp.dense_h_to_4h.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.mlp.dense_h_to_4h.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.post_attention_layernorm.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.22.post_attention_layernorm.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.attention.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.attention.dense.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.attention.dense.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.attention.masked_bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.attention.query_key_value.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.attention.query_key_value.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.attention.rotary_emb.inv_freq": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.input_layernorm.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.input_layernorm.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.mlp.dense_4h_to_h.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.mlp.dense_4h_to_h.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.mlp.dense_h_to_4h.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.mlp.dense_h_to_4h.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.post_attention_layernorm.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.23.post_attention_layernorm.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.attention.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.attention.dense.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.attention.dense.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.attention.masked_bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.attention.query_key_value.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.attention.query_key_value.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.attention.rotary_emb.inv_freq": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.input_layernorm.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.input_layernorm.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.mlp.dense_4h_to_h.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.mlp.dense_4h_to_h.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.mlp.dense_h_to_4h.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.mlp.dense_h_to_4h.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.post_attention_layernorm.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.24.post_attention_layernorm.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.25.attention.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.25.attention.dense.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.25.attention.dense.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.25.attention.masked_bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.25.attention.query_key_value.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.25.attention.query_key_value.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.25.attention.rotary_emb.inv_freq": "model-00006-of-00008.safetensors",
"gpt_neox.layers.25.input_layernorm.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.25.input_layernorm.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.25.mlp.dense_4h_to_h.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.25.mlp.dense_4h_to_h.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.25.mlp.dense_h_to_4h.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.25.mlp.dense_h_to_4h.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.25.post_attention_layernorm.bias": "model-00006-of-00008.safetensors",
"gpt_neox.layers.25.post_attention_layernorm.weight": "model-00006-of-00008.safetensors",
"gpt_neox.layers.26.attention.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.attention.dense.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.attention.dense.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.attention.masked_bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.attention.query_key_value.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.attention.query_key_value.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.attention.rotary_emb.inv_freq": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.input_layernorm.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.input_layernorm.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.mlp.dense_4h_to_h.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.mlp.dense_4h_to_h.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.mlp.dense_h_to_4h.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.mlp.dense_h_to_4h.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.post_attention_layernorm.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.26.post_attention_layernorm.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.attention.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.attention.dense.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.attention.dense.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.attention.masked_bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.attention.query_key_value.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.attention.query_key_value.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.attention.rotary_emb.inv_freq": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.input_layernorm.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.input_layernorm.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.mlp.dense_4h_to_h.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.mlp.dense_4h_to_h.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.mlp.dense_h_to_4h.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.mlp.dense_h_to_4h.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.post_attention_layernorm.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.27.post_attention_layernorm.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.attention.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.attention.dense.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.attention.dense.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.attention.masked_bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.attention.query_key_value.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.attention.query_key_value.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.attention.rotary_emb.inv_freq": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.input_layernorm.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.input_layernorm.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.mlp.dense_4h_to_h.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.mlp.dense_4h_to_h.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.mlp.dense_h_to_4h.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.mlp.dense_h_to_4h.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.post_attention_layernorm.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.28.post_attention_layernorm.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.attention.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.attention.dense.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.attention.dense.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.attention.masked_bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.attention.query_key_value.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.attention.query_key_value.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.attention.rotary_emb.inv_freq": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.input_layernorm.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.input_layernorm.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.mlp.dense_4h_to_h.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.29.mlp.dense_4h_to_h.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.29.mlp.dense_h_to_4h.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.29.mlp.dense_h_to_4h.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.29.post_attention_layernorm.bias": "model-00007-of-00008.safetensors",
"gpt_neox.layers.29.post_attention_layernorm.weight": "model-00007-of-00008.safetensors",
"gpt_neox.layers.3.attention.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.3.attention.dense.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.3.attention.dense.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.3.attention.masked_bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.3.attention.query_key_value.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.3.attention.query_key_value.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.3.attention.rotary_emb.inv_freq": "model-00001-of-00008.safetensors",
"gpt_neox.layers.3.input_layernorm.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.3.input_layernorm.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.3.mlp.dense_4h_to_h.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.3.mlp.dense_4h_to_h.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.3.mlp.dense_h_to_4h.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.3.mlp.dense_h_to_4h.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.3.post_attention_layernorm.bias": "model-00001-of-00008.safetensors",
"gpt_neox.layers.3.post_attention_layernorm.weight": "model-00001-of-00008.safetensors",
"gpt_neox.layers.30.attention.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.attention.dense.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.attention.dense.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.attention.masked_bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.attention.query_key_value.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.attention.query_key_value.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.attention.rotary_emb.inv_freq": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.input_layernorm.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.input_layernorm.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.mlp.dense_4h_to_h.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.mlp.dense_4h_to_h.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.mlp.dense_h_to_4h.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.mlp.dense_h_to_4h.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.post_attention_layernorm.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.30.post_attention_layernorm.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.attention.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.attention.dense.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.attention.dense.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.attention.masked_bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.attention.query_key_value.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.attention.query_key_value.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.attention.rotary_emb.inv_freq": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.input_layernorm.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.input_layernorm.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.mlp.dense_4h_to_h.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.mlp.dense_4h_to_h.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.mlp.dense_h_to_4h.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.mlp.dense_h_to_4h.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.post_attention_layernorm.bias": "model-00008-of-00008.safetensors",
"gpt_neox.layers.31.post_attention_layernorm.weight": "model-00008-of-00008.safetensors",
"gpt_neox.layers.4.attention.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.attention.dense.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.attention.dense.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.attention.masked_bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.attention.query_key_value.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.attention.query_key_value.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.attention.rotary_emb.inv_freq": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.input_layernorm.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.input_layernorm.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.mlp.dense_4h_to_h.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.mlp.dense_4h_to_h.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.mlp.dense_h_to_4h.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.mlp.dense_h_to_4h.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.post_attention_layernorm.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.4.post_attention_layernorm.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.attention.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.attention.dense.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.attention.dense.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.attention.masked_bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.attention.query_key_value.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.attention.query_key_value.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.attention.rotary_emb.inv_freq": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.input_layernorm.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.input_layernorm.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.mlp.dense_4h_to_h.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.mlp.dense_4h_to_h.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.mlp.dense_h_to_4h.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.mlp.dense_h_to_4h.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.post_attention_layernorm.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.5.post_attention_layernorm.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.attention.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.attention.dense.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.attention.dense.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.attention.masked_bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.attention.query_key_value.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.attention.query_key_value.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.attention.rotary_emb.inv_freq": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.input_layernorm.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.input_layernorm.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.mlp.dense_4h_to_h.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.mlp.dense_4h_to_h.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.mlp.dense_h_to_4h.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.mlp.dense_h_to_4h.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.post_attention_layernorm.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.6.post_attention_layernorm.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.attention.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.attention.dense.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.attention.dense.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.attention.masked_bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.attention.query_key_value.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.attention.query_key_value.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.attention.rotary_emb.inv_freq": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.input_layernorm.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.input_layernorm.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.mlp.dense_4h_to_h.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.7.mlp.dense_4h_to_h.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.7.mlp.dense_h_to_4h.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.mlp.dense_h_to_4h.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.post_attention_layernorm.bias": "model-00002-of-00008.safetensors",
"gpt_neox.layers.7.post_attention_layernorm.weight": "model-00002-of-00008.safetensors",
"gpt_neox.layers.8.attention.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.attention.dense.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.attention.dense.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.attention.masked_bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.attention.query_key_value.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.attention.query_key_value.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.attention.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.input_layernorm.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.input_layernorm.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.mlp.dense_4h_to_h.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.mlp.dense_4h_to_h.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.mlp.dense_h_to_4h.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.mlp.dense_h_to_4h.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.post_attention_layernorm.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.8.post_attention_layernorm.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.attention.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.attention.dense.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.attention.dense.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.attention.masked_bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.attention.query_key_value.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.attention.query_key_value.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.attention.rotary_emb.inv_freq": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.input_layernorm.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.input_layernorm.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.mlp.dense_4h_to_h.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.mlp.dense_4h_to_h.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.mlp.dense_h_to_4h.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.mlp.dense_h_to_4h.weight": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.post_attention_layernorm.bias": "model-00003-of-00008.safetensors",
"gpt_neox.layers.9.post_attention_layernorm.weight": "model-00003-of-00008.safetensors"
}
}

3
pytorch_model.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:c690d2e06597fb5f5f7c0e7444d417fe9f391787741a26b1c0fd953b0bdfbd6f
size 7754324845

11
special_tokens_map.json Normal file
View File

@@ -0,0 +1,11 @@
{
"additional_special_tokens": [
"<|endoftext|>",
"<|sep|>",
"<|acc|>",
"<|tel|>",
"<|rrn|>"
],
"eos_token": "<|endoftext|>",
"pad_token": "<|endoftext|>"
}

3
tokenizer.json Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:60e8bd123994badd563a603bce1a319dd7f5dc4798a476c2d88c671766db2f24
size 1652158

6
tokenizer_config.json Normal file
View File

@@ -0,0 +1,6 @@
{
"name_or_path": "EleutherAI/polyglot-ko-3.8b",
"eos_token": "<|endoftext|>",
"pad_token": "<|endoftext|>",
"tokenizer_class": "PreTrainedTokenizerFast"
}