Skip to content

Pull requests: huggingface/trl

New pull request New
33 Open 1,036 Closed
33 Open 1,036 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

setup_chat_format: throw error if there is already a template in base model
#2252 opened Oct 18, 2024 by ngxson Loading…
1 of 5 tasks
1
Adjust padding in batch generation
#2251 opened Oct 18, 2024 by gaetanlop Loading…
3 tasks done
Refactor log_reports.py for Improved Logging, File Processing, and Slack Payload Handling
#2249 opened Oct 18, 2024 by Mefisto04 Loading…
1
Conversational dataset support for KTOTrainer
#2248 opened Oct 18, 2024 by qgallouedec Loading…
5 tasks
1
1
Bump the minimum transformers version to v4.46
#2245 opened Oct 17, 2024 by qgallouedec • Draft
5 tasks
1
[Judges] use the pair-judges in online-preference trainers
#2243 opened Oct 16, 2024 by kashif Loading…
12
Data mixer Integration
#2240 opened Oct 16, 2024 by August-murr • Draft
3 of 5 tasks
1
[SFT VLM] Added support for Molmo models via standalone script sft_vlm_molmo
#2236 opened Oct 15, 2024 by sergiopaniego Loading…
2 of 5 tasks
1
[online-DPO] evaluaiton step error 🐛 bug Something isn't working
#2231 opened Oct 15, 2024 by kashif Loading…
1
1
Remove ds_config scheuduler params to prevent deepseed from creating scheduler for ref_model
#2224 opened Oct 11, 2024 by Ben-Schneider-code Loading…
2 of 5 tasks
1
Refactor DPO data processing
#2209 opened Oct 9, 2024 by qgallouedec Loading…
5 tasks
26
fixed: OverflowError: out of range integral type conversion attempted
#2206 opened Oct 9, 2024 by himanshushukla12 Loading…
1 of 5 tasks
1
13
Add VAS to TRL ✨ enhancement New feature or request
#2195 opened Oct 7, 2024 by idanshen Loading…
[CGPO] CGPO Trainer (single task single objective) ✨ enhancement New feature or request
#2190 opened Oct 6, 2024 by gaetanlop • Draft
9 of 11 tasks
1
Change KTO tokenization to use DPO's 🏋 KTO Related to KTO
#2187 opened Oct 6, 2024 by kawine Loading…
15
[CGPO] Mixture of judges 👨‍⚖️ judge Related to judges
#2159 opened Oct 3, 2024 by gaetanlop Loading…
4 tasks done
18
Remove graph breaks for torch.compile() in padding free branch in DataCollatorForCompletionOnlyLM 🐛 bug Something isn't working 🏋 SFT Related to SFT
#2158 opened Oct 3, 2024 by Abhishek-TAMU Loading…
1 of 5 tasks
4
populate SUPPORTED_COMMANDS cli
#2157 opened Oct 2, 2024 by grumpyp Loading…
4 of 5 tasks
1
7
[Open discusion] Multistep dataset
#2148 opened Oct 1, 2024 by qgallouedec • Draft
4 tasks
4
DPO trainer supports num_logits_to_keep to save memory 🏋 DPO Related to DPO
#2129 opened Sep 26, 2024 by xyangk Loading…
3 of 5 tasks
@kashif
3
Process-supervised RM Trainer
#2127 opened Sep 26, 2024 by gaetanlop • Draft
5 tasks done
59
[SCoRE] initial score stage 1
#2115 opened Sep 24, 2024 by kashif • Draft
1
Remove deprecated args in trainers
#2036 opened Sep 8, 2024 by qgallouedec • Draft
5 tasks
4
feat: add support for packing tokenized datasets
#2011 opened Sep 3, 2024 by kmehant Loading…
3 of 5 tasks
1
8
allow masking on consecutive messages with same roles
#2000 opened Aug 31, 2024 by lsy641 Loading…
4 of 5 tasks
ProTip! Type g p on any issue or pull request to go back to the pull request listing page.