Hugging Face's logo

Datasets:
Bc-AI
/
codva-it-data

Dataset card Files
xet
 Community
Bc-AI commited on
Commit
a5a8676
·
verified ·
1 Parent(s): eb281af

auto upload metadata.json

Browse files
Files changed (1) hide show
  1. metadata.json +272 -0
metadata.json ADDED
@@ -0,0 +1,272 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_name": "CodVa",
3
+ "pipeline_version": "codva_it_v4_merged_full",
4
+ "vocab_size": 50269,
5
+ "sequence_length": 2048,
6
+ "eos_token_id": 50256,
7
+ "im_start_id": 50257,
8
+ "im_end_id": 50258,
9
+ "train_samples": 2232347,
10
+ "val_samples": 42496,
11
+ "format": "one_example_per_row_with_loss_mask",
12
+ "label_ignore_index": -100,
13
+ "domain_token_budgets": {
14
+ "code": 600000000,
15
+ "math": 300000000,
16
+ "reasoning": 400000000,
17
+ "general": 300000000
18
+ },
19
+ "domain_token_actuals": {
20
+ "general": 19295136,
21
+ "code": 367961900,
22
+ "math": 211541727,
23
+ "reasoning": 403273403
24
+ },
25
+ "domain_doc_actuals": {
26
+ "general": 123592,
27
+ "code": 641821,
28
+ "math": 626712,
29
+ "reasoning": 732718
30
+ },
31
+ "hours_elapsed": 1.32,
32
+ "collection_stats": {
33
+ "ise-uiuc/Magicoder-OSS-Instruct-75K": {
34
+ "domain": "code",
35
+ "config": null,
36
+ "examples": 65409,
37
+ "skipped": 0,
38
+ "dupes": 9788
39
+ },
40
+ "ise-uiuc/Magicoder-Evol-Instruct-110K": {
41
+ "domain": "code",
42
+ "config": null,
43
+ "examples": 83077,
44
+ "skipped": 0,
45
+ "dupes": 28106
46
+ },
47
+ "m-a-p/CodeFeedback-Filtered-Instruction": {
48
+ "domain": "code",
49
+ "config": null,
50
+ "examples": 87571,
51
+ "skipped": 38,
52
+ "dupes": 68917
53
+ },
54
+ "Vezora/Tested-143k-Python-Alpaca": {
55
+ "domain": "code",
56
+ "config": null,
57
+ "examples": 59623,
58
+ "skipped": 0,
59
+ "dupes": 83704
60
+ },
61
+ "glaiveai/glaive-code-assistant-v3": {
62
+ "domain": "code",
63
+ "config": null,
64
+ "examples": 151040,
65
+ "skipped": 2,
66
+ "dupes": 3993
67
+ },
68
+ "TokenBender/code_instructions_122k_alpaca_style": {
69
+ "domain": "code",
70
+ "config": null,
71
+ "examples": 101376,
72
+ "skipped": 174,
73
+ "dupes": 7
74
+ },
75
+ "sahil2801/CodeAlpaca-20k": {
76
+ "domain": "code",
77
+ "config": null,
78
+ "examples": 15645,
79
+ "skipped": 252,
80
+ "dupes": 4125
81
+ },
82
+ "iamtarun/python_code_instructions_18k_alpaca": {
83
+ "domain": "code",
84
+ "config": null,
85
+ "examples": 3124,
86
+ "skipped": 13,
87
+ "dupes": 15475
88
+ },
89
+ "nickrosh/Evol-Instruct-Code-80k-v1": {
90
+ "domain": "code",
91
+ "config": null,
92
+ "examples": 11164,
93
+ "skipped": 252,
94
+ "dupes": 66848
95
+ },
96
+ "bigcode/self-oss-instruct-sc2-exec-filter-50k": {
97
+ "domain": "code",
98
+ "config": null,
99
+ "examples": 50661,
100
+ "skipped": 0,
101
+ "dupes": 0
102
+ },
103
+ "deepmind/code_contests": {
104
+ "domain": "code",
105
+ "config": null,
106
+ "examples": 13131,
107
+ "skipped": 194,
108
+ "dupes": 3
109
+ },
110
+ "TIGER-Lab/MathInstruct": {
111
+ "domain": "math",
112
+ "config": null,
113
+ "examples": 151040,
114
+ "skipped": 1,
115
+ "dupes": 8783
116
+ },
117
+ "gsm8k": {
118
+ "domain": "math",
119
+ "config": "main",
120
+ "examples": 7256,
121
+ "skipped": 0,
122
+ "dupes": 217
123
+ },
124
+ "microsoft/orca-math-word-problems-200k": {
125
+ "domain": "math",
126
+ "config": null,
127
+ "examples": 150016,
128
+ "skipped": 2,
129
+ "dupes": 5453
130
+ },
131
+ "math_qa": {
132
+ "domain": "math",
133
+ "config": null,
134
+ "examples": 0,
135
+ "skipped": 0,
136
+ "dupes": 0
137
+ },
138
+ "ChilleD/SVAMP": {
139
+ "domain": "math",
140
+ "config": null,
141
+ "examples": 700,
142
+ "skipped": 0,
143
+ "dupes": 0
144
+ },
145
+ "garage-bAInd/Open-Platypus": {
146
+ "domain": "math",
147
+ "config": null,
148
+ "examples": 16644,
149
+ "skipped": 5724,
150
+ "dupes": 2558
151
+ },
152
+ "AI-MO/NuminaMath-CoT": {
153
+ "domain": "math",
154
+ "config": null,
155
+ "examples": 100352,
156
+ "skipped": 0,
157
+ "dupes": 12026
158
+ },
159
+ "meta-math/MetaMathQA": {
160
+ "domain": "math",
161
+ "config": null,
162
+ "examples": 200704,
163
+ "skipped": 0,
164
+ "dupes": 9554
165
+ },
166
+ "Open-Orca/SlimOrca": {
167
+ "domain": "reasoning",
168
+ "config": null,
169
+ "examples": 151040,
170
+ "skipped": 0,
171
+ "dupes": 475
172
+ },
173
+ "teknium/OpenHermes-2.5": {
174
+ "domain": "reasoning",
175
+ "config": null,
176
+ "examples": 200704,
177
+ "skipped": 0,
178
+ "dupes": 16716
179
+ },
180
+ "Open-Orca/OpenOrca": {
181
+ "domain": "reasoning",
182
+ "config": null,
183
+ "examples": 151040,
184
+ "skipped": 0,
185
+ "dupes": 16972
186
+ },
187
+ "LDJnr/Capybara": {
188
+ "domain": "reasoning",
189
+ "config": null,
190
+ "examples": 14133,
191
+ "skipped": 0,
192
+ "dupes": 1873
193
+ },
194
+ "HuggingFaceH4/no_robots": {
195
+ "domain": "general",
196
+ "config": null,
197
+ "examples": 9500,
198
+ "skipped": 0,
199
+ "dupes": 0
200
+ },
201
+ "timdettmers/openassistant-guanaco": {
202
+ "domain": "reasoning",
203
+ "config": null,
204
+ "examples": 9825,
205
+ "skipped": 0,
206
+ "dupes": 1
207
+ },
208
+ "openbmb/UltraInteract_sft": {
209
+ "domain": "reasoning",
210
+ "config": null,
211
+ "examples": 84480,
212
+ "skipped": 0,
213
+ "dupes": 16416
214
+ },
215
+ "WizardLM/WizardLM_evol_instruct_V2_196k": {
216
+ "domain": "reasoning",
217
+ "config": null,
218
+ "examples": 73368,
219
+ "skipped": 1,
220
+ "dupes": 16400
221
+ },
222
+ "Anthropic/hh-rlhf": {
223
+ "domain": "general",
224
+ "config": "harmless-base",
225
+ "examples": 0,
226
+ "skipped": 0,
227
+ "dupes": 0
228
+ },
229
+ "HuggingFaceH4/ultrachat_200k": {
230
+ "domain": "reasoning",
231
+ "config": null,
232
+ "examples": 48128,
233
+ "skipped": 0,
234
+ "dupes": 0
235
+ },
236
+ "databricks/databricks-dolly-15k": {
237
+ "domain": "general",
238
+ "config": null,
239
+ "examples": 14862,
240
+ "skipped": 129,
241
+ "dupes": 20
242
+ },
243
+ "yahma/alpaca-cleaned": {
244
+ "domain": "general",
245
+ "config": null,
246
+ "examples": 51200,
247
+ "skipped": 130,
248
+ "dupes": 3
249
+ }
250
+ },
251
+ "special_tokens": {
252
+ "<|im_start|>": 50257,
253
+ "<|im_end|>": 50258,
254
+ "<|code_start|>": 50259,
255
+ "<|code_end|>": 50260,
256
+ "<|math_start|>": 50261,
257
+ "<|math_end|>": 50262,
258
+ "<|solution|>": 50263,
259
+ "<|doc_end|>": 50264,
260
+ "<|domain_code|>": 50265,
261
+ "<|domain_math|>": 50266,
262
+ "<|domain_general|>": 50267,
263
+ "<|domain_reasoning|>": 50268,
264
+ "<|endoftext|>": 50256
265
+ },
266
+ "anchor_repeat": 3,
267
+ "note": "v2 + v3 datasets fully merged. Overlap with pretrain is intentional \u2014 SFT teaches format/turn-taking on known content. Identity anchors repeated 3x to ensure CodVa name/persona locks in. Train with 2-3 epochs for best results on this dataset size. | UltraChat 200k train_sft appended: 150,000 examples (was 2,082,347 \u2192 now 2,232,347).",
268
+ "ultrachat_appended": true,
269
+ "ultrachat_examples": 150000,
270
+ "ultrachat_skipped": 0,
271
+ "ultrachat_dupes": 0
272
+ }