Dataset Viewer
Auto-converted to Parquet Duplicate
project
string
organization
string
version
string
model
string
rows
int64
sequence_length
int64
total_tokens
int64
sources
dict
tokenizer
dict
packing
dict
binary
dict
Acauã
Harpia AI Research
3.1
Nanbeige/Nanbeige4.2-3B
10,000
1,024
10,240,000
{ "code": { "dataset": "nvidia/OpenCodeInstruct", "sequences": 9700, "ratio": 0.97, "official_shards": 50, "fields": [ "id", "input", "output", "domain", "generation_algorithm", "llm_judgement", "unit_tests", "tests_execution_status", "average_...
{ "name": "Nanbeige/Nanbeige4.2-3B", "use_fast": false, "trust_remote_code": true, "chat_template": true, "enable_thinking": false }
{ "strategy": "continuous_stream", "sequence_length": 1024, "discard_incomplete": true, "shuffle": true, "shuffle_unit": "sequence", "seed": 42 }
{ "dtype": "uint32", "endianness": "little", "shape": [ 10000, 1024 ], "bytes": 40960000 }

No dataset card yet

Downloads last month
-