Dataset

Nemotron-Pretraining-Indic-Specialized-v1

Name: Nemotron-Pretraining-Indic-Specialized-v1
Creator: nvidia

by nvidia

Nemotron-Pretraining-Indic-Specialized-v1 Dataset Description: Nemotron-Pretraining-Indic-Specialized-v1 is the Specialized portion of the Nemotron-Pretraining-Indic corpus. It keeps only the records whose sourcerepo is nvidia/Nemotron-Pretraining-Specializ…

No licensetext-generation 6 0100M < n < 1B rows

published 26 Oct 2026

Open on Hugging Face

Preview

First 5 rows of or-Latn-IN / train, from the Hugging Face dataset viewer.

uuid	text	language_code	source_text	intermediate_text	evaluation_result	overall_score	intermediate_evaluation_result	intermediate_overall_score	metadata	_source_repo	_source_config	_subset_key	_tier	_source_file	license	token_count_source	tokenizer_model	source_language	language	source_uuid
5a623282-396c-4720-9de0-c5bffca55ccc	def max_gcd_of_substring_length_and_lcm(s: str) -> int: """Eka contiguous substring ra length ebong sehi substring re thiba samasta akshara ra alphabetic positions (a=1, b=2, …, z=26) ra LCM madhyare sarbadhika GCD khojantu. `s` ra pratyeka sambhabya substring pain, nimnalikhita ganana kara…	or-Latn-IN	def max_gcd_of_substring_length_and_lcm(s: str) -> int: """Find the maximum GCD between the length of a contiguous substring and the LCM of the alphabetic positions (a=1, b=2, …, z=26) of all letters in that substring. For each possible substring of `s`, compute length = number…	def max_gcd_of_substring_length_and_lcm(s: str) -> int: """ଏକ contiguous substring ର ଲମ୍ବ (length) ଏବଂ ସେହି substring ରେ ଥିବା ସମସ୍ତ ଅକ୍ଷରର alphabetic positions (a=1, b=2, …, z=26) ର LCM ମଧ୍ୟରେ ସର୍ବାଧିକ GCD ଖୋଜନ୍ତୁ। `s` ର ପ୍ରତ୍ୟେକ ସମ୍ଭାବ୍ୟ substring ପାଇଁ, ନିମ୍ନଲିଖିତ ଗଣନା କରନ୍ତୁ: leng…	{"analysis":"The translation is highly accurate. It correctly converts Odia script to Latin script (Romanization) while preserving the Python code structure, docstrings, and comments. No faithfulness, script compliance, or formatting errors were found. The Romanization is consistent and fluent.","e…	100	{"analysis":"The translation is of high professional quality. It accurately preserves the technical meaning of the Python docstring and comments. There are no major errors in faithfulness or format. A minor style point is the use of English terms like 'contiguous substring' and 'character' in paren…	100	{"category": "Nemotron-Pretraining-Code-Concepts", "models_used": "gpt-oss-20b,gpt-oss-120b", "function_name": "max_gcd_of_substring_length_and_lcm", "tags": "algorithms.max-subarray,functionality.algorithm.string-operations,functionality.math.number-theory.gcd-lcm,algorithms.strings.manipulation",…	nvidia/Nemotron-Pretraining-Specialized-v1.1	Nemotron-Pretraining-Code-Concepts	specialized_v11_code_concepts	top_priority	Nemotron-Pretraining-Code-Concepts/part_000013.parquet	cc-by-4.0	429	google/gemma-4-31B-it	English	Odia	ff3a4683-8c72-46fc-902c-bc220fca3f2d
70348d79-2431-4875-8a9d-d893db49982a	from typing import List def longest_common_palindromic_subrow_length(matrix: List[List[int]]) -> int: """ Eka matrix ra ati kamare duiti row re thiba eka contiguous subarray ra sarbadhika lamba pherai diantu, jaha nije eka palindrome ate. Jadi epari kounasi subarray nathae, tebe 0 pher…	or-Latn-IN	from typing import List def longest_common_palindromic_subrow_length(matrix: List[List[int]]) -> int: """ Return the maximum length of a contiguous subarray that appears in at least two rows of a matrix and is itself a palindrome. If no such subarray exists, return 0. Examples …	from typing import List def longest_common_palindromic_subrow_length(matrix: List[List[int]]) -> int: """ ଏକ ମାଟ୍ରିକ୍ସର ଅତି କମରେ ଦୁଇଟି row ରେ ଥିବା ଏକ contiguous subarray ର ସର୍ବାଧିକ ଲମ୍ବ ଫେରାଇ ଦିଅନ୍ତୁ, ଯାହା ନିଜେ ଏକ palindrome ଅଟେ। ଯଦି ଏପରି କୌଣସି subarray ନଥାଏ, ତେବେ 0 ଫେରାଇ ଦିଅନ୍ତୁ। …	{"analysis":"The translation is highly accurate. The Odia text within the docstring was correctly romanized into Latin script while maintaining the exact meaning and technical context. All Python code, docstring structure, and examples were perfectly preserved. No errors found.","error_analysis":[]…	100	{"analysis":"The translation is highly accurate. It correctly translates the docstring into Odia while preserving all Python code, indentation, and formatting. The use of English technical terms like 'row', 'contiguous subarray', and 'palindrome' is acceptable in a technical programming context to …	100	{"category": "Nemotron-Pretraining-Code-Concepts", "models_used": "gpt-oss-20b,gpt-oss-120b", "function_name": "longest_common_palindromic_subrow_length", "tags": "data-structures.arrays.matrix,algorithms.technique.two-pointer,algorithms.arrays.intersection,algorithms.strings.palindrome", "problem_…	nvidia/Nemotron-Pretraining-Specialized-v1.1	Nemotron-Pretraining-Code-Concepts	specialized_v11_code_concepts	top_priority	Nemotron-Pretraining-Code-Concepts/part_000013.parquet	cc-by-4.0	514	google/gemma-4-31B-it	English	Odia	9e11cd3b-a748-4a31-ab7d-5c2a09323f94
a5160184-dddb-42ab-a7ce-8fffb516c662	from typing import List def find_first_index_in_csv_numbers(csv_numbers: str, target: int) -> int: """Comma-separated sankhyagudikara eka talika ku parse karantu, pratyeka entry ku int re paribartana karantu, ta'pare target ra prathama abastha pain sarala bhabare khojantu. >>> find_fir…	or-Latn-IN	from typing import List def find_first_index_in_csv_numbers(csv_numbers: str, target: int) -> int: """Parse a comma‑separated list of numbers, convert each entry to an int in place, then linearly search for the first occurrence of the target. >>> find_first_index_in_csv_numbers("1, 2, …	from typing import List def find_first_index_in_csv_numbers(csv_numbers: str, target: int) -> int: """କମା-ପୃଥକିତ (comma-separated) ସଂଖ୍ୟାଗୁଡ଼ିକର ଏକ ତାଲିକାକୁ ପାର୍ସ (parse) କରନ୍ତୁ, ପ୍ରତ୍ୟେକ ଏଣ୍ଟ୍ରିକୁ int ରେ ପରିବର୍ତ୍ତନ କରନ୍ତୁ, ତା'ପରେ ଟାର୍ଗେଟ୍‌ର ପ୍ରଥମ ଅବସ୍ଥା ପାଇଁ ସରଳ ଭାବରେ ଖୋଜନ୍ତୁ। >>> fin…	{"analysis":"The translation is highly accurate. It successfully romanized all Odia prose while maintaining the code structure and logic. One minor fluency error was found in the transcription of 'ଆସିପାରୁଥିବା' as 'ansiparuthiba' (likely a typo for 'asiparuthiba').","error_analysis":[{"error_categor…	95	{"analysis":"The translation is high quality. It preserves the Python code structure and docstring formatting perfectly. There are a few instances of code-switching/transliteration (e.g., 'a linear search'), which are common and acceptable in technical Odia contexts to maintain clarity. No major fa…	95	{"category": "Nemotron-Pretraining-Code-Concepts", "models_used": "gpt-oss-20b,gpt-oss-120b", "function_name": "find_first_index_in_csv_numbers", "tags": "functionality.data-processing.conversion,algorithms.technique.in-place,algorithms.technique.linear-search,functionality.data-processing.text-par…	nvidia/Nemotron-Pretraining-Specialized-v1.1	Nemotron-Pretraining-Code-Concepts	specialized_v11_code_concepts	top_priority	Nemotron-Pretraining-Code-Concepts/part_000013.parquet	cc-by-4.0	376	google/gemma-4-31B-it	English	Odia	1bf2bbe2-853b-4f34-97a5-a8b9f037fe01
243972b0-4639-474e-9ab1-f2327546bd6d	from typing import List def convert_and_search(arr: List[str], target: int) -> int: """Numeric strings ra eka list ku in-place tarikare integers re paribarttan karantu ebang target ra index pherai diantu. Ehi paribarttan mula list ku paribarttita kare; ehara pare linear search karajae. …	or-Latn-IN	from typing import List def convert_and_search(arr: List[str], target: int) -> int: """Convert list of numeric strings in-place to integers and return the index of target. The conversion modifies the original list; afterwards perform linear search. If target is not found the function r…	from typing import List def convert_and_search(arr: List[str], target: int) -> int: """Numeric strings ର ଏକ ଲିଷ୍ଟକୁ in-place ତରିକାରେ integers ରେ ପରିବର୍ତ୍ତନ କରନ୍ତୁ ଏବଂ target ର index ଫେରାଇ ଦିଅନ୍ତୁ। ଏହି ପରିବର୍ତ୍ତନ ମୂଳ ଲିଷ୍ଟକୁ ପରିବର୍ତ୍ତିତ କରେ; ଏହା ପରେ linear search କରାଯାଏ। ଯଦି target ମିଳେ…	{"analysis":"The translation is highly accurate. All Odia script has been correctly Romanized into Latin script. Code structure, docstrings, and comments are perfectly preserved. There are no faithfulness or script compliance errors. Only a very minor fluency nuance exists in the Romanization of 'k…	100	{"analysis":"The translation is highly accurate. It preserves the Python code structure and docstrings perfectly. There is heavy use of English technical terms (in-place, integers, linear search), which is standard and acceptable in Odia technical contexts to maintain clarity. No major or critical …	98	{"category": "Nemotron-Pretraining-Code-Concepts", "models_used": "gpt-oss-20b,gpt-oss-120b", "function_name": "convert_and_search", "tags": "functionality.data-processing.conversion,algorithms.technique.in-place,algorithms.technique.linear-search,functionality.data-processing.text-parsing", "probl…	nvidia/Nemotron-Pretraining-Specialized-v1.1	Nemotron-Pretraining-Code-Concepts	specialized_v11_code_concepts	top_priority	Nemotron-Pretraining-Code-Concepts/part_000013.parquet	cc-by-4.0	224	google/gemma-4-31B-it	English	Odia	163b9521-0656-4fdf-b2da-140d1ac1cee9
f6c87f93-25db-419e-bb2b-5154ad94f505	from typing import Dict, List, Tuple def has_balanced_prefixed_path( graph: Dict[int, List[Tuple[int, str]]], start: int, end: int, prefix: str ) -> bool: """directed graph re source node ru target node parjyanta emiti eka path achhi ki nahi janch karantu, jenthare path aba…	or-Latn-IN	from typing import Dict, List, Tuple def has_balanced_prefixed_path( graph: Dict[int, List[Tuple[int, str]]], start: int, end: int, prefix: str ) -> bool: """Check if there is a path from source node to target node in the directed graph such that the concatenation of the ed…	from typing import Dict, List, Tuple def has_balanced_prefixed_path( graph: Dict[int, List[Tuple[int, str]]], start: int, end: int, prefix: str ) -> bool: """directed graph ରେ source node ରୁ target node ପର୍ଯ୍ୟନ୍ତ ଏମିତି ଏକ path ଅଛି କି ନାହିଁ ଯାଞ୍ଚ କରନ୍ତୁ, ଯେଉଁଠାରେ path ଅବଧି e…	{"analysis":"The translation is highly accurate and complies with all script and formatting constraints. There are two minor fluency/typo issues: 'dindhara' is an awkward romanization/typo for 'diya jaithiba' (given) and 'lanu' is a typo for 'laku' or 'laku-ku' (to). No major or critical errors wer…	88	{"analysis":"The translation is highly accurate. It preserves the Python code and docstring structure perfectly. The Odia translation of the comments and docstring is fluent and technically correct. There are no Major or Critical errors. A few Minor style choices involve mixing English technical te…	100	{"category": "Nemotron-Pretraining-Code-Concepts", "models_used": "gpt-oss-20b,gpt-oss-120b", "function_name": "has_balanced_prefixed_path", "tags": "functionality.algorithm.graph,functionality.parser.parentheses,algorithms.problem.general.solving,algorithms.strings.prefix", "problem_prompt": "Plea…	nvidia/Nemotron-Pretraining-Specialized-v1.1	Nemotron-Pretraining-Code-Concepts	specialized_v11_code_concepts	top_priority	Nemotron-Pretraining-Code-Concepts/part_000013.parquet	cc-by-4.0	694	google/gemma-4-31B-it	English	Odia	46158319-0da7-4df4-aae8-11bf1aa62523

Cite this

No licensetabulartextCite

Nemotron-Pretraining-Indic-Specialized-v1

Preview

Cite this

Tags