Skip to content

Navigation Menu

Appearance settings

- AI CODE CREATION
  GitHub CopilotWrite better code with AI
  GitHub Copilot appDirect agents from issue to merge
  MCP Registry^NewIntegrate external tools
- DEVELOPER WORKFLOWS
  ActionsAutomate any workflow
  CodespacesInstant dev environments
  IssuesPlan and track work
  Code ReviewManage code changes
- APPLICATION SECURITY
  GitHub Advanced SecurityFind and fix vulnerabilities
  Code securitySecure your code as you build
  Secret protectionStop leaks before they start
- EXPLORE
  Why GitHub
  Documentation
  Blog
  Changelog
  Marketplace
View all features
- BY COMPANY SIZE
  Enterprises
  Small and medium teams
  Startups
  Nonprofits
- BY USE CASE
  App Modernization
  DevSecOps
  DevOps
  CI/CD
  View all use cases
- BY INDUSTRY
  Healthcare
  Financial services
  Manufacturing
  Government
  View all industries
View all solutions
- EXPLORE BY TOPIC
  AI
  Software Development
  DevOps
  Security
  View all topics
- EXPLORE BY TYPE
  Customer stories
  Events & webinars
  Ebooks & reports
  Business insights
  GitHub Skills
- SUPPORT & SERVICES
  Documentation
  Customer support
  Community forum
  Trust center
  Partners
View all resources
- COMMUNITY
  GitHub SponsorsFund open source developers
- PROGRAMS
  Security Lab
  Maintainer Community
  Accelerator
  GitHub Stars
  Archive Program
- REPOSITORIES
  Topics
  Trending
  Collections
- ENTERPRISE SOLUTIONS
  Enterprise platformAI-powered developer platform
- AVAILABLE ADD-ONS
  GitHub Advanced SecurityEnterprise-grade security features
  Copilot for BusinessEnterprise-grade AI features
  Premium SupportEnterprise-grade 24/7 support
Pricing

Search code, repositories, users, issues, pull requests...

Search

Clear

Search syntax tips

Provide feedback

We read every piece of feedback, and take your input very seriously.

Include my email address so I can be contacted

Saved searches

Use saved searches to filter your results more quickly

Name

Query

To see all available qualifiers, see our documentation.

Appearance settings

You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session.

Dismiss alert

Uh oh!

There was an error while loading. Please reload this page.

LocalResearchGroup / llm-foundry Public

Notifications You must be signed in to change notification settings
Fork 2
Star 5

Code
Issues 9
Pull requests 6
Actions
Projects
Wiki
Security and quality
Insights

Additional navigation options

Code
Issues
Pull requests
Actions
Projects
Wiki
Security and quality
Insights

#42 add some splits for HuggingFaceTB/finemath dataset#43

Closed

tyoc213 wants to merge 50 commits into

mainLocalResearchGroup/llm-foundry:mainfrom

dataset-finemathLocalResearchGroup/llm-foundry:dataset-finemathCopy head branch name to clipboard

Conversation Commits50 (50)Checks Files changed

Closed

#42 add some splits for HuggingFaceTB/finemath dataset#43
tyoc213 wants to merge 50 commits into
mainfrom
dataset-finemath

Commits

Commits on Aug 30, 2025

#42 add some splits for HuggingFaceTB/finemath dataset
tyoc213
committed
#42 generate train and validation from single train split
tyoc213
committed
donf't shuffle
tyoc213
committed
Create test and validation for finemath-4plus and generate sets of 1M, 100k, 10k, 1k
tyoc213
committed
revert changes to convert_dataset_hf.py to generate finemath splits
tyoc213
committed
Add some logs and file checks
tyoc213
committed
fix missing import and path check
tyoc213
committed
Generate train/val for tulu-3-sft-olmo-2-mixture without aya source and ablations of 100k, 10k, 1k
tyoc213
committed
starting to work on split_hf_datasets.py

matdmiller
authored and
tyoc213
committed
Refactor to single file and multiple calls for datasets: finemath, tulu and numina
tyoc213
committed
Adding configurations to allow script tokenize new datasets
tyoc213
committed
OK Splitting into train/test and starting exploring dataset tokenization
tyoc213
committed
little cleanup
tyoc213
committed
numina can now be tokenized, tulu is still failing
tyoc213
committed
split q/a messages column into prompt/response and remove unnecessary columns
tyoc213
committed
`data_prep/split_hf_datasets.py -h` for help and `data_prep/split_hf_datasets.py` to run all from start to finish (split+tokenize+upload works for finemath)
tyoc213
committed
Making use of convert_finetuning_dataset_from_args to tokenize tulu and nomina, token folder has same shape so all can be uploaded
tyoc213
committed
feat `download_repo.py` and `modal_script#pull_hf_to_folder`
tyoc213
committed
fix modal DATASETS_VOLUME_MOUNT_PATH and args
tyoc213
committed
fix autoscaling parameter https://modal.com/docs/guide/modal-1-0-migration
tyoc213
committed
add pull_hf_to_folder which downloads: tulu, numina and finemath hf splits
tyoc213
committed
add glaiveai/glaive-code-assistant-v3
tyoc213
committed
adding avelina python edu
tyoc213
committed
extra commas and path changes
tyoc213
committed
fix glaive -1M option added some extra logs
tyoc213
committed
revert back to `concurrency_limit` for modal < 0.73.76
tyoc213
committed
added missings glaive, avelinapythonedu configs
tyoc213
committed
missing cmdn line options in previous commit
tyoc213
committed
adding concat tokens for pretran datasets that indicates use the tokenizer; refactored dataset_constants_splits_config
tyoc213
committed
quickfix `"` by `'`
tyoc213
committed
fix concat_tokens value
tyoc213
committed
addking kind to make clear which datasets are instruct and which pretrain. Remove extra method
tyoc213
committed
Override ablations internal configuration with `--one-k`` to do only 1k split for any 3 steps (or all):
tyoc213
committed
missed 8192 tokens
tyoc213
committed
add constants directly in convert_dataset_hf so that it can be called stand alone from scripts folder
tyoc213
committed
tokenize each row and padd it up to max_length
tyoc213
committed
add glaive preprocessing after_pull on original dataset
tyoc213
committed
use tokenizer if tokenizer is specified without concat tokens
tyoc213
committed
pretrain set to use tokenizer with concat tokens = None
tyoc213
committed
adding system prompt chat template
tyoc213
committed
numina and tulu preproc after_pull
tyoc213
committed
`built_tokenizer` when tokenizer exist for pretraining dataset
tyoc213
committed
set correct max_seq_len for instruct and pretrain datasets
tyoc213
committed
preprocs methods need to return the colums that have the template applied
tyoc213
committed
`concat_tokens=max_seq_len` for pretrain datasets
tyoc213
committed
remove dead code
tyoc213
committed
extend timeout to process datasets, fix instruct tokenizer, fix padding that sets a row per sample in concat tokens path
tyoc213
committed
`no_wrap=True` to trim samples for pretrain data
tyoc213
committed
force enter loop if any token present and break; make batchsize=1 so that no errors are show because different sizes in batch (as not padded up to a length)
tyoc213
committed
add little more info on how to use
tyoc213
committed

Footer

© 2026 GitHub, Inc.

Footer navigation

Terms
Privacy
Security
Status
Community
Docs
Contact

You can’t perform that action at this time.