-
Notifications
You must be signed in to change notification settings - Fork 328
Pull requests: huggingface/nanotron
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Bump the actions group with 3 updates
dependencies
Pull requests that update a dependency file
github_actions
Pull requests that update GitHub Actions code
#425
opened Sep 17, 2026 by
dependabot
Bot
Loading…
fix(config): name the offending keys when a union fails to match
#424
opened Sep 17, 2026 by
AloysJehwin
Loading…
fix(helpers): stop subtracting warmup twice from the LR decay horizon
#423
opened Sep 15, 2026 by
AloysJehwin
Loading…
fix(trainer): guard loss_avg computation when pipeline parallelism is active
#422
opened Sep 11, 2026 by
AloysJehwin
Loading…
fix(parametrization): pass full config to parametrizator_cls instead of config.model
#421
opened Sep 9, 2026 by
AloysJehwin
Loading…
fix(helpers): compute remaining stage steps from stage end so resume LR decays correctly
#420
opened Sep 8, 2026 by
AloysJehwin
Loading…
fix(trainer): clarify microbatch assertion to reference batch_accumulation_per_replica
#419
opened Sep 6, 2026 by
AloysJehwin
Loading…
fix(llama): restore attention_output indentation and fix init_model_randomly config arg
#418
opened Sep 5, 2026 by
AloysJehwin
Loading…
4 tasks
fix(config): deserialize all model_config dict types in ModelArgs.__post_init__
#416
opened Jul 29, 2026 by
AloysJehwin
Loading…
fix pp is enabled, the loss_avg log reports an error
#412
opened Jul 6, 2026 by
liguodongiot
Loading…
1 of 6 tasks
feat: opt-in torch.compile support via GeneralArgs.torch_compile
#411
opened Jul 6, 2026 by
py-ai-dev
Loading…
4 tasks
feat: fused RoPE kernel for LLaMA via torchembed (3–4x faster on CUDA)
#410
opened Jul 6, 2026 by
py-ai-dev
Loading…
4 tasks
add missing docstrings to assert_fail_except_rank_with, before_tbi_sanity_checks and after_tbi_sanity_checks
#409
opened Jul 4, 2026 by
RavSinghChandan
Loading…
add missing docstrings to set_random_seed, set_random_state and get_synced_random_state
#408
opened Jul 4, 2026 by
RavSinghChandan
Loading…
add missing docstrings to get_parameter_and_parent_module, get_untyped_storage, tensor_from_untyped_storage and find_free_port
#407
opened Jul 4, 2026 by
RavSinghChandan
Loading…
fix the error of passing in configuration object when llama calls par…
#406
opened Jul 3, 2026 by
liguodongiot
Loading…
1 of 6 tasks
Use per-group initial_lr for constant warmup (fix muP custom-LR groups)
#404
opened Jun 5, 2026 by
lollinng
Loading…
Previous Next
ProTip!
Updated in the last three days: updated:>2026-09-18.