Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / linkedin/Liger-Kernel issues and pull requests

#461 - Qwen VL Convergence Test Fails for Transformers >= 4.47.0

Issue - State: closed - Opened by ByronHsu 2 months ago - 2 comments

#460 - Add dynamic dependency management for CUDA and ROCm

Pull Request - State: closed - Opened by hebiao064 2 months ago

#459 - Fix liger orpo trainer import error

Pull Request - State: closed - Opened by ByronHsu 2 months ago

#457 - add sponsorship and collab

Pull Request - State: closed - Opened by ByronHsu 2 months ago

#456 - Add HIP (ROCm) and Liger Kernel to env report

Pull Request - State: closed - Opened by Comet0322 2 months ago

#455 - version bump to 0.5.0

Pull Request - State: closed - Opened by shivam15s 2 months ago

#454 - change chunked readme

Pull Request - State: closed - Opened by shivam15s 2 months ago

#453 - add chunked loss to readme

Pull Request - State: closed - Opened by shivam15s 2 months ago

#452 - add eng blog

Pull Request - State: closed - Opened by ByronHsu 2 months ago

#451 - Make pyproject.toml detect dependencies based on the platform

Issue - State: closed - Opened by ByronHsu 2 months ago - 1 comment

#450 - Make kernel doc lean

Pull Request - State: closed - Opened by ByronHsu 2 months ago

#449 - Add paper link and formula for preference loss

Pull Request - State: closed - Opened by ByronHsu 2 months ago

#448 - improve code quality for chunk loss

Pull Request - State: closed - Opened by ByronHsu 2 months ago

#447 - Reference model bug (no way to input ref_inputs)

Issue - State: closed - Opened by shivam15s 2 months ago

#446 - Specify scheduled CI in AMD badge

Pull Request - State: closed - Opened by ByronHsu 2 months ago

#445 - Show liger-kernel and rocm version in env_report

Issue - State: closed - Opened by ByronHsu 2 months ago - 1 comment

#444 - Refactor Temperature Scaling in Distillation Loss

Pull Request - State: closed - Opened by austin362667 2 months ago - 3 comments

#443 - Add xpu in env report

Pull Request - State: closed - Opened by abhilash1910 2 months ago

#442 - Fix Normalization Term in Distillation Loss

Pull Request - State: closed - Opened by austin362667 2 months ago - 1 comment

#441 - Support offline `logits` for teacher model

Issue - State: open - Opened by austin362667 2 months ago

#440 - update ci icon on readme

Pull Request - State: closed - Opened by bboyleonp666 2 months ago

#439 - Bug in chunking?

Issue - State: open - Opened by cinjon 2 months ago - 10 comments

#438 - Fix All `chunked_loss` Benchmark Scripts

Pull Request - State: closed - Opened by austin362667 2 months ago

#437 - [WIP] Add softcapping to preference based fused linear

Pull Request - State: closed - Opened by ryankert01 2 months ago - 1 comment

#436 - [AMD] [CI] Clean up `amd-ci`

Pull Request - State: closed - Opened by tjtanaa 2 months ago

#435 - Fix LigerCrossEntropyLoss Reduction Behavior for "None" Mode

Pull Request - State: closed - Opened by hebiao064 2 months ago - 1 comment

#434 - [CI] shorten ci name

Pull Request - State: closed - Opened by ByronHsu 2 months ago

#433 - [CI] rename ci and add cron job for amd

Pull Request - State: closed - Opened by ByronHsu 2 months ago

#432 - Introduce Knowledge Distillation Base

Pull Request - State: closed - Opened by austin362667 2 months ago - 2 comments

#431 - Add Build Success/Fail Badge

Pull Request - State: closed - Opened by hebiao064 2 months ago

#430 - [AMD] [CI] Added Dockerfile and AMD-CI test workflow

Pull Request - State: closed - Opened by tjtanaa 3 months ago

#428 - Switch amd-ci to use MI300X runner.

Pull Request - State: closed - Opened by saienduri 3 months ago - 11 comments
Labels: AMD

#427 - Switch amd-ci to use MI300X runner

Pull Request - State: closed - Opened by saienduri 3 months ago - 3 comments

#426 - PreferenceBase with Softcapping?

Issue - State: open - Opened by cinjon 3 months ago - 2 comments

#425 - Add JSD Loss for Distillation

Pull Request - State: open - Opened by austin362667 3 months ago - 3 comments

#424 - [tiny] Add QwQ to readme (same arch as Qwen2)

Pull Request - State: closed - Opened by tyler-romero 3 months ago

#423 - Enhance Cross Entropy Softcap Unit Test

Pull Request - State: closed - Opened by austin362667 3 months ago

#422 - Enhance Cross Entropy Softcap Unit Test

Pull Request - State: closed - Opened by austin362667 3 months ago - 1 comment

#421 - CrossEntropyLoss return single value when reduction is "none"

Issue - State: closed - Opened by leng-yue 3 months ago - 5 comments

#420 - Add weight support for LigerCrossEntropy

Pull Request - State: closed - Opened by Tcc0403 3 months ago - 7 comments

#419 - Add weight support for LigerCrossEntropy

Pull Request - State: closed - Opened by Tcc0403 3 months ago - 1 comment

#418 - No gradient test in softcap unit test for LigerCrossEntropy

Issue - State: closed - Opened by Tcc0403 3 months ago - 1 comment
Labels: good first issue

#417 - Introduce Knowledge Distillation Base

Pull Request - State: closed - Opened by austin362667 3 months ago - 3 comments

#416 - Fix os env

Pull Request - State: closed - Opened by ByronHsu 3 months ago

#416 - Fix os env

Pull Request - State: closed - Opened by ByronHsu 3 months ago

#415 - Add rebuild to CI

Pull Request - State: closed - Opened by ByronHsu 3 months ago

#415 - Add rebuild to CI

Pull Request - State: closed - Opened by ByronHsu 3 months ago

#414 - The accuracy is misaligned when using bf16

Issue - State: open - Opened by starstream 3 months ago

#414 - The accuracy is misaligned when using bf16

Issue - State: open - Opened by starstream 3 months ago

#413 - Fix `get_batch_loss_metrics` comments

Pull Request - State: closed - Opened by austin362667 3 months ago

#413 - Fix `get_batch_loss_metrics` comments

Pull Request - State: closed - Opened by austin362667 3 months ago

#412 - Adjust QWEN2 VL Loss `rtol`

Pull Request - State: closed - Opened by austin362667 3 months ago

#412 - Adjust QWEN2 VL Loss `rtol`

Pull Request - State: closed - Opened by austin362667 3 months ago

#411 - QWEN2 VL doesn't converge

Issue - State: closed - Opened by austin362667 3 months ago

#411 - QWEN2 VL doesn't converge

Issue - State: closed - Opened by austin362667 3 months ago

#410 - KTO loss

Pull Request - State: open - Opened by vulkomilev 3 months ago - 3 comments

#409 - No Significant Improvement Observed in Model Training Speed

Issue - State: open - Opened by lianghsun 3 months ago - 5 comments

#409 - No Significant Improvement Observed in Model Training Speed

Issue - State: open - Opened by lianghsun 3 months ago - 4 comments

#408 - Introduce Distillation with a Chunked, Fused Linear JS-divergence Loss

Pull Request - State: closed - Opened by austin362667 3 months ago - 4 comments

#407 - Xpu support

Pull Request - State: closed - Opened by mgrabban 3 months ago - 2 comments

#406 - Optimize CE Loss by casting dtype to float32 inside kernel

Pull Request - State: closed - Opened by pramodith 3 months ago

#404 - Weighted Cross Entropy Loss

Issue - State: open - Opened by winglian 3 months ago - 4 comments

#403 - Implement softcapping in fused jsd

Pull Request - State: open - Opened by wheynelau 3 months ago - 3 comments

#402 - add nn.module support for chunked loss function

Pull Request - State: closed - Opened by shivam15s 3 months ago - 1 comment

#400 - Enable keyword arguments for liger functional

Pull Request - State: closed - Opened by hongpeng-guo 3 months ago - 3 comments

#399 - [1/2] Enable keyword arguments for liger functional

Pull Request - State: closed - Opened by hongpeng-guo 3 months ago - 4 comments

#398 - Is eager attention still required for Gemma2?

Issue - State: closed - Opened by dachenlian 3 months ago - 2 comments

#397 - Add script to reproducibly run examples on Modal

Pull Request - State: closed - Opened by tyler-romero 3 months ago - 1 comment

#396 - add xpu support

Pull Request - State: closed - Opened by mgrabban 3 months ago - 9 comments

#395 - [WIP]: Autotune Chunk Size

Pull Request - State: open - Opened by pramodith 3 months ago

#394 - fix qwen2 import failure in test

Pull Request - State: closed - Opened by ByronHsu 3 months ago

#393 - Generalize JSD to FKL/RKL

Pull Request - State: closed - Opened by yundai424 3 months ago

#392 - Fix incomplete RMSNorm patch

Pull Request - State: closed - Opened by Tcc0403 3 months ago

#391 - Apple's cross entropy computation

Issue - State: open - Opened by fzyzcjy 3 months ago - 8 comments

#390 - AttributeError: 'Qwen2RMSNorm' object has no attribute 'in_place'

Issue - State: closed - Opened by jdf-prog 3 months ago - 6 comments

#389 - Qwen2-VL Training Example w/ Liger

Pull Request - State: closed - Opened by tyler-romero 3 months ago - 1 comment

#388 - Qwen2-VL Bug / Incompatibility Fixes

Pull Request - State: closed - Opened by tyler-romero 3 months ago

#387 - Fix DPO with Reference Model

Pull Request - State: closed - Opened by austin362667 3 months ago - 1 comment

#386 - Add Chunked SimPO Loss

Pull Request - State: closed - Opened by pramodith 3 months ago - 2 comments

#385 - Fix flce not being patched after reverting in convergence test

Pull Request - State: closed - Opened by Tcc0403 3 months ago

#384 - Support Qwen2-VL's multimodal RoPE implementation

Pull Request - State: closed - Opened by li-plus 3 months ago - 5 comments

#383 - AttributeError: 'MistralRMSNorm' object has no attribute 'in_place'

Issue - State: closed - Opened by hrdxwandg 3 months ago - 6 comments

#382 - Adds the CPO Alignment Loss Function

Pull Request - State: closed - Opened by pramodith 3 months ago - 1 comment

#381 - Refactor `LigerFusedLinearPreferenceBase`

Pull Request - State: closed - Opened by pramodith 3 months ago - 1 comment

#380 - [test fix] test_mini_models_with_logits and test_mini_models are backwards

Pull Request - State: closed - Opened by tyler-romero 3 months ago - 1 comment

#379 - add xpu device support for `rms_norm`

Pull Request - State: closed - Opened by faaany 3 months ago - 7 comments

#378 - Support Chunked DPO Loss Kernel

Pull Request - State: closed - Opened by austin362667 3 months ago - 5 comments

#377 - modify readmes and create license/acknowledgement docs

Pull Request - State: closed - Opened by shivam15s 3 months ago

#376 - Support out-of-place RMSNorm to fix gemma2

Pull Request - State: closed - Opened by ByronHsu 3 months ago

#375 - Support CE after grad acc fix

Pull Request - State: closed - Opened by ByronHsu 3 months ago - 2 comments

#374 - Fix cross entropy patch for LLama

Pull Request - State: closed - Opened by pramodith 3 months ago - 1 comment

#373 - Fix release password

Pull Request - State: closed - Opened by ByronHsu 3 months ago

#372 - Rotate modal and pypi tokens

Pull Request - State: closed - Opened by ByronHsu 3 months ago

#371 - [RFC] Liger FlexChunkLoss: Alignment and Distillation loss

Issue - State: open - Opened by shivam15s 3 months ago - 22 comments

#370 - Convergence test for gemma2

Issue - State: closed - Opened by Tcc0403 3 months ago - 1 comment

#369 - LigerCrossEntropyLoss is not patched for latest transformers models

Issue - State: closed - Opened by Tcc0403 3 months ago - 1 comment