Hugging Face's logo

bigcode
/
starcoder2-15b-instruct-v0.1

Text Generation
Transformers
Safetensors
starcoder2
code
conversational
Eval Results
text-generation-inference
Model card Files Community
6

add AIBOM

#6
by fatima113 - opened
base: refs/heads/main
←
from: refs/pr/6
Discussion Files changed
+164
-0
Files changed (1) hide show
  1. bigcode_starcoder2-15b-instruct-v0.1.json +164 -0
bigcode_starcoder2-15b-instruct-v0.1.json ADDED
@@ -0,0 +1,164 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bomFormat": "CycloneDX",
3
+ "specVersion": "1.6",
4
+ "serialNumber": "urn:uuid:19c1f27e-d53a-45dc-94ba-a0f9fb30f79a",
5
+ "version": 1,
6
+ "metadata": {
7
+ "timestamp": "2025-07-10T08:49:17.278628+00:00",
8
+ "component": {
9
+ "type": "machine-learning-model",
10
+ "bom-ref": "bigcode/starcoder2-15b-instruct-v0.1-e7aea605-679e-5203-beb5-6bb2732ebe54",
11
+ "name": "bigcode/starcoder2-15b-instruct-v0.1",
12
+ "externalReferences": [
13
+ {
14
+ "url": "https://huggingface.co/bigcode/starcoder2-15b-instruct-v0.1",
15
+ "type": "documentation"
16
+ }
17
+ ],
18
+ "modelCard": {
19
+ "modelParameters": {
20
+ "task": "text-generation",
21
+ "architectureFamily": "starcoder2",
22
+ "modelArchitecture": "Starcoder2ForCausalLM",
23
+ "datasets": [
24
+ {
25
+ "ref": "bigcode/self-oss-instruct-sc2-exec-filter-50k-87afcdf4-328e-5009-ba04-8150239e25f2"
26
+ }
27
+ ]
28
+ },
29
+ "properties": [
30
+ {
31
+ "name": "library_name",
32
+ "value": "transformers"
33
+ },
34
+ {
35
+ "name": "base_model",
36
+ "value": "bigcode/starcoder2-15b"
37
+ }
38
+ ],
39
+ "quantitativeAnalysis": {
40
+ "performanceMetrics": [
41
+ {
42
+ "slice": "dataset: livecodebench-codegeneration",
43
+ "type": "pass@1",
44
+ "value": 20.4
45
+ },
46
+ {
47
+ "slice": "dataset: livecodebench-selfrepair",
48
+ "type": "pass@1",
49
+ "value": 20.9
50
+ },
51
+ {
52
+ "slice": "dataset: livecodebench-testoutputprediction",
53
+ "type": "pass@1",
54
+ "value": 29.8
55
+ },
56
+ {
57
+ "slice": "dataset: livecodebench-codeexecution",
58
+ "type": "pass@1",
59
+ "value": 28.1
60
+ },
61
+ {
62
+ "slice": "dataset: humaneval",
63
+ "type": "pass@1",
64
+ "value": 72.6
65
+ },
66
+ {
67
+ "slice": "dataset: humanevalplus",
68
+ "type": "pass@1",
69
+ "value": 63.4
70
+ },
71
+ {
72
+ "slice": "dataset: mbpp",
73
+ "type": "pass@1",
74
+ "value": 75.2
75
+ },
76
+ {
77
+ "slice": "dataset: mbppplus",
78
+ "type": "pass@1",
79
+ "value": 61.2
80
+ },
81
+ {
82
+ "slice": "dataset: ds-1000",
83
+ "type": "pass@1",
84
+ "value": 40.6
85
+ }
86
+ ]
87
+ }
88
+ },
89
+ "authors": [
90
+ {
91
+ "name": "bigcode"
92
+ }
93
+ ],
94
+ "licenses": [
95
+ {
96
+ "license": {
97
+ "name": "bigcode-openrail-m"
98
+ }
99
+ }
100
+ ],
101
+ "description": "We introduce StarCoder2-15B-Instruct-v0.1, the very first entirely self-aligned code Large Language Model (LLM) trained with a fully permissive and transparent pipeline. Our open-source pipeline uses StarCoder2-15B to generate thousands of instruction-response pairs, which are then used to fine-tune StarCoder-15B itself without any human annotations or distilled data from huge and proprietary LLMs.- **Model:** [bigcode/starcoder2-15b-instruct-v0.1](https://huggingface.co/bigcode/starcoder2-instruct-15b-v0.1)- **Code:** [bigcode-project/starcoder2-self-align](https://github.com/bigcode-project/starcoder2-self-align)- **Dataset:** [bigcode/self-oss-instruct-sc2-exec-filter-50k](https://huggingface.co/datasets/bigcode/self-oss-instruct-sc2-exec-filter-50k/)- **Authors:**[Yuxiang Wei](https://yuxiang.cs.illinois.edu),[Federico Cassano](https://federico.codes/),[Jiawei Liu](https://jw-liu.xyz),[Yifeng Ding](https://yifeng-ding.com),[Naman Jain](https://naman-ntc.github.io),[Harm de Vries](https://www.harmdevries.com),[Leandro von Werra](https://twitter.com/lvwerra),[Arjun Guha](https://www.khoury.northeastern.edu/home/arjunguha/main/home/),[Lingming Zhang](https://lingming.cs.illinois.edu).![self-alignment pipeline](https://huggingface.co/datasets/bigcode/starcoder2-instruct-assets/resolve/main/method.png)",
102
+ "tags": [
103
+ "transformers",
104
+ "safetensors",
105
+ "starcoder2",
106
+ "text-generation",
107
+ "code",
108
+ "conversational",
109
+ "dataset:bigcode/self-oss-instruct-sc2-exec-filter-50k",
110
+ "arxiv:2410.24198",
111
+ "base_model:bigcode/starcoder2-15b",
112
+ "base_model:finetune:bigcode/starcoder2-15b",
113
+ "license:bigcode-openrail-m",
114
+ "model-index",
115
+ "autotrain_compatible",
116
+ "text-generation-inference",
117
+ "endpoints_compatible",
118
+ "region:us"
119
+ ]
120
+ }
121
+ },
122
+ "components": [
123
+ {
124
+ "type": "data",
125
+ "bom-ref": "bigcode/self-oss-instruct-sc2-exec-filter-50k-87afcdf4-328e-5009-ba04-8150239e25f2",
126
+ "name": "bigcode/self-oss-instruct-sc2-exec-filter-50k",
127
+ "data": [
128
+ {
129
+ "type": "dataset",
130
+ "bom-ref": "bigcode/self-oss-instruct-sc2-exec-filter-50k-87afcdf4-328e-5009-ba04-8150239e25f2",
131
+ "name": "bigcode/self-oss-instruct-sc2-exec-filter-50k",
132
+ "contents": {
133
+ "url": "https://huggingface.co/datasets/bigcode/self-oss-instruct-sc2-exec-filter-50k",
134
+ "properties": [
135
+ {
136
+ "name": "pretty_name",
137
+ "value": "StarCoder2-15b Self-Alignment Dataset (50K)"
138
+ },
139
+ {
140
+ "name": "configs",
141
+ "value": "Name of the dataset subset: default {\"split\": \"train\", \"path\": \"data/train-*\"}"
142
+ },
143
+ {
144
+ "name": "license",
145
+ "value": "odc-by"
146
+ }
147
+ ]
148
+ },
149
+ "governance": {
150
+ "owners": [
151
+ {
152
+ "organization": {
153
+ "name": "bigcode",
154
+ "url": "https://huggingface.co/bigcode"
155
+ }
156
+ }
157
+ ]
158
+ },
159
+ "description": "Final self-alignment training dataset for StarCoder2-Instruct. \n\nseed: Contains the seed Python function\nconcepts: Contains the concepts generated from the seed\ninstruction: Contains the instruction generated from the concepts\nresponse: Contains the execution-validated response to the instruction\n\nThis dataset utilizes seed Python functions derived from the MultiPL-T pipeline.\n"
160
+ }
161
+ ]
162
+ }
163
+ ]
164
+ }