Hugging Face's logo

rinna
/
japanese-gpt-neox-3.6b-instruction-sft

Text Generation
Transformers
PyTorch
Safetensors
Japanese
gpt_neox
lm
nlp
text-generation-inference
Model card Files
xet
Community
2

add AIBOM

#2
by fatima113 - opened
base: refs/heads/main
←
from: refs/pr/2
Discussion Files changed
+152
-0
Files changed (1) hide show
  1. rinna_japanese-gpt-neox-3.6b-instruction-sft.json +152 -0
rinna_japanese-gpt-neox-3.6b-instruction-sft.json ADDED
@@ -0,0 +1,152 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bomFormat": "CycloneDX",
3
+ "specVersion": "1.6",
4
+ "serialNumber": "urn:uuid:8fbc208b-8a4d-4877-a1c1-5d07a22f5cc0",
5
+ "version": 1,
6
+ "metadata": {
7
+ "timestamp": "2025-06-05T09:41:50.547933+00:00",
8
+ "component": {
9
+ "type": "machine-learning-model",
10
+ "bom-ref": "rinna/japanese-gpt-neox-3.6b-instruction-sft-e7a06d20-f7da-5712-b713-897284b52f74",
11
+ "name": "rinna/japanese-gpt-neox-3.6b-instruction-sft",
12
+ "externalReferences": [
13
+ {
14
+ "url": "https://huggingface.co/rinna/japanese-gpt-neox-3.6b-instruction-sft",
15
+ "type": "documentation"
16
+ }
17
+ ],
18
+ "modelCard": {
19
+ "modelParameters": {
20
+ "task": "text-generation",
21
+ "architectureFamily": "gpt_neox",
22
+ "modelArchitecture": "GPTNeoXForCausalLM",
23
+ "datasets": [
24
+ {
25
+ "ref": "Anthropic/hh-rlhf-5da2bc0b-1d42-54c8-8b53-3770b07b011a"
26
+ },
27
+ {
28
+ "ref": "stanfordnlp/SHP-f9be653c-ec3d-5323-8491-cc46c65f1623"
29
+ }
30
+ ]
31
+ },
32
+ "properties": [
33
+ {
34
+ "name": "library_name",
35
+ "value": "transformers"
36
+ },
37
+ {
38
+ "name": "base_model",
39
+ "value": "rinna/japanese-gpt-neox-3.6b"
40
+ }
41
+ ]
42
+ },
43
+ "authors": [
44
+ {
45
+ "name": "rinna"
46
+ }
47
+ ],
48
+ "licenses": [
49
+ {
50
+ "license": {
51
+ "id": "MIT",
52
+ "url": "https://spdx.org/licenses/MIT.html"
53
+ }
54
+ }
55
+ ],
56
+ "description": "This repository provides a Japanese GPT-NeoX model of 3.6 billion parameters. The model is based on [`rinna/japanese-gpt-neox-3.6b`](https://huggingface.co/rinna/japanese-gpt-neox-3.6b) and has been finetuned to serve as an instruction-following conversational agent.* **Model architecture**A 36-layer, 2816-hidden-size transformer-based language model.* **Finetuning**The finetuning data is the subset of the following datasets and has been translated into Japanese.* [Anthropic HH RLHF data](https://huggingface.co/datasets/Anthropic/hh-rlhf)* [FLAN Instruction Tuning data](https://github.com/google-research/FLAN)* [Stanford Human Preferences Dataset](https://huggingface.co/datasets/stanfordnlp/SHP)The data will **not** be released.* **Model Series**| Variant | Link || :-- | :--|| 3.6B PPO | https://huggingface.co/rinna/japanese-gpt-neox-3.6b-instruction-ppo || 3.6B SFT-v2 | https://huggingface.co/rinna/japanese-gpt-neox-3.6b-instruction-sft-v2 || 3.6B SFT | https://huggingface.co/rinna/japanese-gpt-neox-3.6b-instruction-sft || 3.6B pretrained | https://huggingface.co/rinna/japanese-gpt-neox-3.6b |* **Contributors**[Tianyu Zhao](https://huggingface.co/tianyuz) and [Kei Sawada](https://huggingface.co/keisawada)* **Release date**March 17, 2023",
57
+ "tags": [
58
+ "transformers",
59
+ "pytorch",
60
+ "safetensors",
61
+ "gpt_neox",
62
+ "text-generation",
63
+ "lm",
64
+ "nlp",
65
+ "ja",
66
+ "dataset:Anthropic/hh-rlhf",
67
+ "dataset:stanfordnlp/SHP",
68
+ "arxiv:2404.01657",
69
+ "base_model:rinna/japanese-gpt-neox-3.6b",
70
+ "base_model:finetune:rinna/japanese-gpt-neox-3.6b",
71
+ "license:mit",
72
+ "autotrain_compatible",
73
+ "text-generation-inference",
74
+ "region:us"
75
+ ]
76
+ }
77
+ },
78
+ "components": [
79
+ {
80
+ "type": "data",
81
+ "bom-ref": "Anthropic/hh-rlhf-5da2bc0b-1d42-54c8-8b53-3770b07b011a",
82
+ "name": "Anthropic/hh-rlhf",
83
+ "data": [
84
+ {
85
+ "type": "dataset",
86
+ "bom-ref": "Anthropic/hh-rlhf-5da2bc0b-1d42-54c8-8b53-3770b07b011a",
87
+ "name": "Anthropic/hh-rlhf",
88
+ "contents": {
89
+ "url": "https://huggingface.co/datasets/Anthropic/hh-rlhf",
90
+ "properties": [
91
+ {
92
+ "name": "license",
93
+ "value": "mit"
94
+ }
95
+ ]
96
+ },
97
+ "governance": {
98
+ "owners": [
99
+ {
100
+ "organization": {
101
+ "name": "Anthropic",
102
+ "url": "https://huggingface.co/Anthropic"
103
+ }
104
+ }
105
+ ]
106
+ },
107
+ "description": "\n\t\n\t\t\n\t\tDataset Card for HH-RLHF\n\t\n\n\n\t\n\t\t\n\t\tDataset Summary\n\t\n\nThis repository provides access to two different kinds of data:\n\nHuman preference data about helpfulness and harmlessness from Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback. These data are meant to train preference (or reward) models for subsequent RLHF training. These data are not meant for supervised training of dialogue agents. Training dialogue agents on these data is likely to lead\u2026 See the full description on the dataset page: https://huggingface.co/datasets/Anthropic/hh-rlhf."
108
+ }
109
+ ]
110
+ },
111
+ {
112
+ "type": "data",
113
+ "bom-ref": "stanfordnlp/SHP-f9be653c-ec3d-5323-8491-cc46c65f1623",
114
+ "name": "stanfordnlp/SHP",
115
+ "data": [
116
+ {
117
+ "type": "dataset",
118
+ "bom-ref": "stanfordnlp/SHP-f9be653c-ec3d-5323-8491-cc46c65f1623",
119
+ "name": "stanfordnlp/SHP",
120
+ "contents": {
121
+ "url": "https://huggingface.co/datasets/stanfordnlp/SHP",
122
+ "properties": [
123
+ {
124
+ "name": "task_categories",
125
+ "value": "text-generation, question-answering"
126
+ },
127
+ {
128
+ "name": "language",
129
+ "value": "en"
130
+ },
131
+ {
132
+ "name": "size_categories",
133
+ "value": "100K<n<1M"
134
+ }
135
+ ]
136
+ },
137
+ "governance": {
138
+ "owners": [
139
+ {
140
+ "organization": {
141
+ "name": "stanfordnlp",
142
+ "url": "https://huggingface.co/stanfordnlp"
143
+ }
144
+ }
145
+ ]
146
+ },
147
+ "description": "\n\t\n\t\t\n\t\t\ud83d\udea2 Stanford Human Preferences Dataset (SHP)\n\t\n\nIf you mention this dataset in a paper, please cite the paper: Understanding Dataset Difficulty with V-Usable Information (ICML 2022).\n\n\t\n\t\t\n\t\tSummary\n\t\n\nSHP is a dataset of 385K collective human preferences over responses to questions/instructions in 18 different subject areas, from cooking to legal advice.\nThe preferences are meant to reflect the helpfulness of one response over another, and are intended to be used for training RLHF\u2026 See the full description on the dataset page: https://huggingface.co/datasets/stanfordnlp/SHP."
148
+ }
149
+ ]
150
+ }
151
+ ]
152
+ }