ultranationalism
  • Joined on Oct 16, 2023
  • Organization
Public Activity Repositories Starred Repositories
1
 Following
0
 Followers
7
 Achievement Badge
0
Loading Heatmap…

ultranationalism synced commits to model-test-refactor at ultranationalism/diffusers from mirror

23 hours ago

ultranationalism synced commits to pull-request/2738 at ultranationalism/Megatron-LM from mirror

1 day ago

ultranationalism synced commits to pull-request/2716 at ultranationalism/Megatron-LM from mirror

  • 37e383d43c revert mtp tie output change that might have lead to error

1 day ago

ultranationalism synced commits to pull-request/2482 at ultranationalism/Megatron-LM from mirror

  • bdbe48792f Merge branch 'main' into robinz/capture_time_main
  • 3cf7a63fe9 Update oncall for next few weeks (#2748)
  • ede9ae4e8f chore: rotate oncall schedule
  • 876a04689b feat: manual registration mode for nccl-ub option when using megatron-fsdp (#2661) Signed-off-by: Youngeun Kwon <youngeunk@nvidia.com>
  • 0a771229e4 Update sequence packing case when dummy PackedSeqParams are used (#2743)
  • Compare 28 commits »

1 day ago

ultranationalism synced commits to main at ultranationalism/Megatron-LM from mirror

  • dd7c9f4f69 Prep work for migrating to types from ModuleSpec (#2668)
  • 3cf7a63fe9 Update oncall for next few weeks (#2748)
  • ede9ae4e8f chore: rotate oncall schedule
  • 876a04689b feat: manual registration mode for nccl-ub option when using megatron-fsdp (#2661) Signed-off-by: Youngeun Kwon <youngeunk@nvidia.com>
  • 0a771229e4 Update sequence packing case when dummy PackedSeqParams are used (#2743)
  • Compare 9 commits »

1 day ago

ultranationalism synced commits to dev at ultranationalism/Megatron-LM from mirror

  • 1068d775d6 Revert "[Dev] Remove calculation of padding token in moe routing loss (#2121)" (#2747) Signed-off-by: Charlie Truong <chtruong@nvidia.com>
  • 0b6714ec87 [Dev] Remove calculation of padding token in moe routing loss (#2121) Co-authored-by: Li Tao <lit@nvidia.com>
  • Compare 2 commits »

1 day ago

ultranationalism synced commits to main at ultranationalism/diffusers from mirror

  • f6b6a7181e Add z-image-omni-base implementation (#12857) * Add z-image-omni-base implementation * Merged into one transformer for Z-Image. * Fix bugs for controlnet after merging the main branch new feature. * Fix for auto_pipeline, Add Styling. * Refactor noise handling and modulation - Add select_per_token function for per-token value selection - Separate adaptive modulation logic - Cleanify t_noisy/clean variable naming - Move image_noise_mask handler from forward to pipeline * Styling & Formatting. * Rewrite code with more non-forward func & clean forward. 1.Change to one forward with shorter code with omni code (None). 2.Split out non-forward funcs: _build_unified_sequence, _prepare_sequence, patchify, pad. * Styling & Formatting. * Manual check fix-copies in controlnet, Add select_per_token, _patchify_image, _pad_with_ids; Styling. * Add Import in pipeline __init__.py. --------- Co-authored-by: Jerry Qilong Wu <xinglong.wql@alibaba-inc.com> Co-authored-by: YiYi Xu <yixu310@gmail.com>

2 days ago

ultranationalism synced commits to pull-request/2685 at ultranationalism/Megatron-LM from mirror

  • 592644541e Merge branch 'main' into stop_words
  • 5b1ef07031 Add oncall to all new PRs (#2734)
  • 8ea3b8d04a ci: Fix copy-pr-bot update (#2736) Signed-off-by: oliver könig <okoenig@nvidia.com>
  • 1c67e7e093 Update copy-pr-bot.yaml [skip ci]
  • cfd980b4e7 Add options to improve data loader initialization time, especially at scale (#2445) Signed-off-by: Antoni-Joan Solergibert <asolergibert@nvidia.com> Co-authored-by: Deepak Narayanan <2724038+deepakn94@users.noreply.github.com>
  • Compare 40 commits »

3 days ago

ultranationalism synced commits to pull-request/2121 at ultranationalism/Megatron-LM from mirror

3 days ago

ultranationalism synced commits to main at ultranationalism/Megatron-LM from mirror

3 days ago

ultranationalism synced commits to dev at ultranationalism/Megatron-LM from mirror

  • 46b550591a [Dev] Fix ep overlap missing final layernorm (#2691)

3 days ago

ultranationalism synced commits to main at ultranationalism/diffusers from mirror

  • 52766e6a69 Use `T5Tokenizer` instead of `MT5Tokenizer` (removed in Transformers v5.0+) (#12877) Use `T5Tokenizer` instead of `MT5Tokenizer` Given that the `MT5Tokenizer` in `transformers` is just a "re-export" of `T5Tokenizer` as per https://github.com/huggingface/transformers/blob/v4.57.3/src/transformers/models/mt5/tokenization_mt5.py )on latest available stable Transformers i.e., v4.57.3), this commit updates the imports to point to `T5Tokenizer` instead, so that those still work with Transformers v5.0.0rc0 onwards.

3 days ago

ultranationalism synced commits to device-map-direct at ultranationalism/diffusers from mirror

  • c61e455ce7 Merge branch 'main' into device-map-direct
  • 973a077c6a Cosmos Predict2.5 14b Conversion (#12863) 14b conversion
  • 0c4f6c9cff Add `OvisImagePipeline` in `AUTO_TEXT2IMAGE_PIPELINES_MAPPING` (#12876)
  • 262ce19bff Feature: Add Mambo-G Guidance as Guider (#12862) * Feature: Add Mambo-G Guidance to Qwen-Image Pipeline * change to guider implementation * fix copied code residual * Update src/diffusers/guiders/magnitude_aware_guidance.py * Apply style fixes --------- Co-authored-by: Pscgylotti <pscgylotti@github.com> Co-authored-by: YiYi Xu <yixu310@gmail.com> Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
  • f7753b1bc8 more update in modular (#12560) * move node registry to mellon * up * fix * modula rpipeline update: filter out none for input_names, fix default blocks for pipe.init() and allow user pass additional kwargs_type in a dict * qwen modular refactor, unpack before decode * update mellon node config, adding* to required_inputs and required_model_inputs * modularpipeline.from_pretrained: error out if no config found * add a component_names property to modular blocks to be consistent! * flux image_encoder -> vae_encoder * controlnet_bundle * refator MellonNodeConfig MellonPipelineConfig * refactor & simplify mellon utils * vae_image_encoder -> vae_encoder * mellon config save keep key order * style + copies * add kwargs input for zimage
  • Compare 18 commits »

3 days ago

ultranationalism synced commits to master at ultranationalism/apex from mirror

  • 3c98f93d1a [pre-commit.ci] pre-commit autoupdate (#1973)

3 days ago

ultranationalism synced commits to pull-request/2724 at ultranationalism/Megatron-LM from mirror

  • d1b016de44 Merge branch 'main' of github.com:wdykas/Megatron-LM into inference-memory-test
  • 4193f3aa9b remove fp16 assert in moe_grouped_gemm & EP (#2495) Co-authored-by: Philip Petrakian <pgpetrak@gmail.com>
  • ddf691df57 [dcos] Update docs title to Megatron Core (#2729)
  • 7f471d720b Fix UVM argument for RL (#2722)
  • 703bc36727 [docs] Add developer section to docs (#2717)
  • Compare 6 commits »

3 days ago

ultranationalism synced commits to pull-request/2716 at ultranationalism/Megatron-LM from mirror

  • e25faa6361 Revert "[Dev] Use the latest Hybrid-EP (#2424)" This reverts commit e8a927578d0fdeb98db5d40ab7bdc81d123795f7.

3 days ago

ultranationalism synced commits to pull-request/2474 at ultranationalism/Megatron-LM from mirror

  • db50940d2c Fix copyright Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>
  • ea0858e4b8 Address comments Signed-off-by: Keshav Santhanam <ksanthanam@nvidia.com>
  • ecb53d165b Merge remote-tracking branch 'upstream/main' into mamba_cuda_graphs
  • 8ea3b8d04a ci: Fix copy-pr-bot update (#2736) Signed-off-by: oliver könig <okoenig@nvidia.com>
  • 1c67e7e093 Update copy-pr-bot.yaml [skip ci]
  • Compare 11 commits »

3 days ago

ultranationalism synced commits to pull-request/2363 at ultranationalism/Megatron-LM from mirror

3 days ago

ultranationalism synced commits to main at ultranationalism/Megatron-LM from mirror

  • 8ea3b8d04a ci: Fix copy-pr-bot update (#2736) Signed-off-by: oliver könig <okoenig@nvidia.com>
  • 1c67e7e093 Update copy-pr-bot.yaml [skip ci]
  • cfd980b4e7 Add options to improve data loader initialization time, especially at scale (#2445) Signed-off-by: Antoni-Joan Solergibert <asolergibert@nvidia.com> Co-authored-by: Deepak Narayanan <2724038+deepakn94@users.noreply.github.com>
  • a05766227d Improve ModelOpt paths & add more Nemotron/hybrid model support (#2131) Signed-off-by: Jennifer Chen <jennifchen@nvidia.com> Co-authored-by: Jared Casper <155158+jaredcasper@users.noreply.github.com>
  • 4193f3aa9b remove fp16 assert in moe_grouped_gemm & EP (#2495) Co-authored-by: Philip Petrakian <pgpetrak@gmail.com>
  • Compare 5 commits »

3 days ago

ultranationalism synced commits to dev at ultranationalism/Megatron-LM from mirror

  • 4665be4dec Revert "[Dev] Use the latest Hybrid-EP (#2424)" (#2732)
  • 2b1fc70891 Revert "[dev] Add assertion for mxfp8 params without dp overlap (#2270)" This reverts commit 7968d5f98f8457297d4a73f96d8a086d84a8fa67.
  • Compare 2 commits »

3 days ago

Please read the following content carefully:

Dear OpenI User

Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.

For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》

Agree and continue
Disagree, exit
Baidu
map