04baf7ff27 Sanitize profile filename (#21395) * sanitize filename * add testing * changelog * fix comment --------- Co-authored-by: Bhimraj Yadav <bhimrajyadav977@gmail.com>
716c2c61bb build(deps): update jsonargparse[jsonnet,signatures] requirement from <4.44.0,>=4.39.0 to >=4.39.0,<4.45.0 in /requirements (#21392) build(deps): update jsonargparse[jsonnet,signatures] requirement Updates the requirements on [jsonargparse[jsonnet,signatures]](https://github.com/omni-us/jsonargparse) to permit the latest version. - [Changelog](https://github.com/omni-us/jsonargparse/blob/main/CHANGELOG.rst) - [Commits](https://github.com/omni-us/jsonargparse/compare/v4.39.0...v4.44.0) --- updated-dependencies: - dependency-name: jsonargparse[jsonnet,signatures] dependency-version: 4.44.0 dependency-type: direct:production ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Nicki Skafte Detlefsen <skaftenicki@gmail.com> Co-authored-by: Bhimraj Yadav <bhimrajyadav977@gmail.com>
- Compare 3 commits »
Xin Yan
thomas-yanxin
- Shanghai, China, CHN
- https://thomas-yanxin.github.io/
- OpenI 核心体验官
- Joined on Sep 05, 2021
- Organization
-
-
-
25
-
8
-
-
57
-
48
-
-
2
-
1
-
-
12
-
13
-
-
2
-
1
-
-
2
-
4
-
-
36
-
45
-
-
1
-
6
-
-
6
-
0
-
-
17
-
11
-
Loading Heatmap…
thomas-yanxin synced commits to master at thomas-yanxin/lightning from mirror
1 hour ago
thomas-yanxin synced commits to dependabot-pip-requirements-scikit-learn-gt-0.22.1-and-lt-1.9.0 at thomas-yanxin/lightning from mirror
88de68b4aa Merge branch 'master' into dependabot-pip-requirements-scikit-learn-gt-0.22.1-and-lt-1.9.0
- Compare 7 commits »
1 hour ago
thomas-yanxin synced commits to dependabot-pip-requirements-pytest-9.0.2 at thomas-yanxin/lightning from mirror
- Compare 11 commits »
1 hour ago
thomas-yanxin synced commits to dependabot-pip-requirements-coverage-7.13.0 at thomas-yanxin/lightning from mirror
- Compare 7 commits »
1 hour ago
thomas-yanxin synced commits to dependabot-github_actions-actions-cache-5 at thomas-yanxin/lightning from mirror
fc39230f96 fix torch-tensorrt dependency in testing (#21413) fix test
- Compare 5 commits »
1 hour ago
thomas-yanxin synced commits to master at thomas-yanxin/apex from mirror
67f496870a [pre-commit.ci] pre-commit autoupdate (#1971) updates: - [github.com/astral-sh/ruff-pre-commit: v0.14.8 → v0.14.9](https://github.com/astral-sh/ruff-pre-commit/compare/v0.14.8...v0.14.9) Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2 hours ago
thomas-yanxin synced commits to main at thomas-yanxin/lmdeploy from mirror
c45970c3d1 [ci] add mllm eval (#4194) * update * update * update * update * update * update * update * update * update * update * update * update * fix lint * updata * add more models * update * update * update * update * add more models * update score * update score format * update * fix lint * update * fix typo
9756373638 Add test for "generate" endpoint (#4181) * TEST: add generate test * update test_input_ids_mode * fix lint * update rl test * fix mm processor args * comment * optimize * add test case * add comments * fix typo * fix generate * warning * remove * check parameters for /generate endpoint * update generate test * update toolkit --------- Co-authored-by: zxy <zhou0493@e.ntu.edu.sg> Co-authored-by: lvhan028 <lvhan_028@163.com>
- Compare 2 commits »
2 hours ago
thomas-yanxin synced commits to main at thomas-yanxin/LLaMA-Factory from mirror
9fd4b094d4 [model] support VibeThinker models (#9616)
18c21bce5a [test] add allreduce test on npu (#9619) Co-authored-by: frozenleaves <frozen@Mac.local>
- Compare 2 commits »
2 hours ago
thomas-yanxin synced commits to restart-failed-nodes at thomas-yanxin/dora from mirror
95d6c27f8f Use atomic to store and update pid
b00b4adee5 chore: Update Cargo.lock (#1263) Co-authored-by: Dora Bot <dora-bot@phil-opp.com>
3b49a44d1e feat: implement dora node list command with metrics and filtering (#1202) (#1262)
e5a11b1892 Fix serialization of `LogLevelOrStdout` to match deserialization (#1258) In https://github.com/dora-rs/dora/commit/40de6a27fdc7c1212477aed4b3d8edbc86a7da38#diff-9503d86f1f775fbb8657d29e9e692c96d5227828367fdd3e05ca08fb6becb28fR89-R111 , the derived `Deserialize` implementation was replaced with a custom one, which doesn't match the `Serialize` implementation. This led to deserialization errors when sending log messages from the daemon to the coordinator. This commit fixes the issue by removing the enum tagging for the `LogLevel` variant, thereby matching the `Deserialize` implementation again. Serialization before: ```json "level": { "LogLevel": "INFO" } ``` Serialization after: ```json "level": "INFO" ```
- Compare 35 commits »
2 hours ago
thomas-yanxin synced commits to main at thomas-yanxin/dora from mirror
- Compare 2 commits »
2 hours ago
thomas-yanxin synced commits to main at thomas-yanxin/flash-attention from mirror
0a5339f4cb [FIRST] Fix softcap scoremod kwargs typo. (#2072)
179f793bbc [CUTE] Seeing if tvvm reduces cpu overhead (#2042)
fd8d5eb363 [Cute,Fwd] Extend score_mod to variable sequence length (#2043) * rebase to main * varlen support for score mod * interface change for varlen score mod * implement varlen support for score mod * varlen score mod working; updated tests * modify varlen score mod to use fastdiv_mods updated per sequence * updated test suite * current working state of varlen score mod * refactor varlen score mod tests * fix to transpose * refactor varlen score mod tests; fix bug; clean up varlen score mod application in kernel * refactor test_score_mod.py to use external score mod definition file * update flash_fwd.py for varlen score mod * sm90 varlen score mod working; test revisions * enable packgqa for varlen score mod; set up fastdiv_mod recomputation * update flash_fwd_sm100.py for recomputing fastdiv_mods & format varlen score mod test * Overwrite pack_gqa.py, tile_scheduler.py, and test_flash_attn.py with origin/main versions * rebase to main * fix test rebase artifacts * fix floor_if_packed redundancy * correct sm90 divmods mismatch * revert test_flash_attn to main * add varlen score mod benchmark script * packgqa for varlen (independent of score mod) * rm benchmark from PR * move score mod arg wrapping to utils.py * format with ruff * major refactor: change score_mod signature to accept seqlen_info and update all tests accordingly * reinstate varlen packgqa exclusion checks * move fastdiv_mods recomputation out of apply_score_mod in prep for varlen mask_mod support * remove duplicate fastdiv_mod recomputation * [Fix] fastdiv_mods for paged attn and seqused_* * clean up PR; fix paged_kv varlen for sm90 * update to varlen score mod test script (paged kv) * remove premature seqlen arguments from sm90 apply_mask_mod
- Compare 3 commits »
3 hours ago
thomas-yanxin synced commits to v5-test_tensor_parallel_moe at thomas-yanxin/transformers from mirror
48c69f7f68 Merge branch 'main' into v5-test_tensor_parallel_moe
4d6516e256 Simplify tie weights logic (#42895) * fix * let's not use source backup, clearer to use original name imo * fix * use a set * simplify * style * add comment
24b311eead fix FastSpeech2ConformerTokenizer crash in tokenize (#42888) Signed-off-by: Wang, Yi <yi.a.wang@intel.com>
0f89661972 Added kernels from kernel hub for Bamba model (#41540) * Added kernels from kernel hub for Bamba model * Updated kernel loading Signed-off-by: romit <romit@ibm.com> * Remove einops Signed-off-by: romit <romit@ibm.com> * Removed global vars Signed-off-by: romit <romit@ibm.com> * Fixed make style Signed-off-by: romit <romit@ibm.com> * Nit Signed-off-by: romit <romit@ibm.com> * Added modeling files Signed-off-by: romit <romit@ibm.com> * Fixed merge conflict Signed-off-by: romit <romit@ibm.com> * fixed lint Signed-off-by: romitjain <romit@ibm.com> * Removed global import * Small updates * Updated * Resolved merge conflicts * Fixed the nested import Signed-off-by: romit <romit@ibm.com> * Moved imports inside mixer Signed-off-by: romit <romit@ibm.com> * CI CD fix Signed-off-by: romit <romit@ibm.com> --------- Signed-off-by: romit <romit@ibm.com> Signed-off-by: romitjain <romit@ibm.com> Co-authored-by: Mohamed Mekkouri <93391238+MekkCyber@users.noreply.github.com>
5d2f82b530 Fix GraniteMoeHybrid in transformers v5 (#42872) * apply_rotary_pos_emb should be called * fix position_embeddings usage in granitemoehybrid * setting `self.rotary_emb` to None only in hybrid models. Safer, since all modules are highly modular. * minor * adding `position_embedding_type` to the config. * review cleanup * modeling too * rewrite conditionally applying rope * resolve rotary_emb issue
- Compare 54 commits »
3 hours ago
thomas-yanxin synced commits to update_v5_guide_toks at thomas-yanxin/transformers from mirror
5ff1f0ea8d Merge branch 'main' into update_v5_guide_toks
31de95ef71 [docs] optimizations quickstart (#42538) * quickstart * feedback * feedback
23394cc491 Simplify using custom resolution for sam3 and sam3_video inference (#42787) * simplify using custom resolution for sam3 and sam3_video inference * revert auto format * use setters and properties * Fix docstring * update dict to correctly save image_size to file for backward compatibility
06378d40e6 fix: Initialize ApertusMLP's xielu activation using `torch_dtype` (#42864) * Fix Apertus model crash on float16 hardware Initialize XIELU activation with correct dtype from config (using config.dtype instead of default bfloat16) to prevent promotion to float32 and subsequent crashes on Turing/float16 GPUs. * refactor: Move `ACT2CLS` import to top-level in Apertus models.
- Compare 26 commits »
3 hours ago
thomas-yanxin synced commits to main at thomas-yanxin/transformers from mirror
b61da25169 Remove duplicated processor class from config (#42806) * remove duplicated processor class from config * adjust the test cases * check public and private attr, both were used in the past
24275124c6 Add local kernel loading support to KernelConfig(). (#42800) * add add_to_mapping_local for KernelConfig * refactor kernel_mapping format * lint code * specify the kernel path * fix `abs/path` Co-authored-by: Mohamed Mekkouri <93391238+MekkCyber@users.noreply.github.com> * rename check_kernel_from_local to use_local_kernel --------- Co-authored-by: Mohamed Mekkouri <93391238+MekkCyber@users.noreply.github.com>
- Compare 15 commits »
3 hours ago
thomas-yanxin synced commits to init-full-meta at thomas-yanxin/transformers from mirror
b19ed98405 until lfm2_moe
- Compare 2 commits »
3 hours ago
thomas-yanxin synced commits to fix-dtype-quantizer at thomas-yanxin/transformers from mirror
32df9bc4dc fix
4b84c54bcd Update src/transformers/modeling_utils.py Co-authored-by: Mohamed Mekkouri <93391238+MekkCyber@users.noreply.github.com>
- Compare 5 commits »
3 hours ago
thomas-yanxin synced commits to export-friendly at thomas-yanxin/transformers from mirror
14113861f1 Merge branch 'main' into export-friendly
7f52a2a4ea Add `.on_push_begin()` callback to Trainer and implement for `TrackioCallback` (#42850) * changes * changes * Update src/transformers/integrations/integration_utils.py Co-authored-by: Quentin Gallouédec <45557362+qgallouedec@users.noreply.github.com> * changes * changes * changes * changes * changes * changes * changes --------- Co-authored-by: Quentin Gallouédec <45557362+qgallouedec@users.noreply.github.com>
- Compare 145 commits »
3 hours ago
thomas-yanxin synced commits to gh-pages at thomas-yanxin/guidellm from mirror
e52635b510 deploy: 7666c658460bc34abe3cc821d3ca072cfd39074a
4 hours ago
thomas-yanxin synced commits to refactor-dpo at thomas-yanxin/trl from mirror
3fabfa4705 sft loss + deprecated max prompt/completion length
- Compare 13 commits »
4 hours ago
thomas-yanxin synced commits to push-generation-with-tiny at thomas-yanxin/trl from mirror
70b9360292 Merge branch 'main' into push-generation-with-tiny
036ae820b3 [docs] Adds GRPO, RSO and LoRA to Paper Index (#4441) Co-authored-by: Quentin Gallouédec <45557362+qgallouedec@users.noreply.github.com> Co-authored-by: Quentin Gallouédec <gallouedec.quentin@gmail.com>
- Compare 6 commits »
4 hours ago