Issues

• upgrade to transformers v5

  foundation-model-stack/fms-acceleration#172 · Filter by author dushyantbehl opened on Oct 1, 2026
  

• Reduce overhead in ODM

  foundation-model-stack/fms-acceleration#163 · Filter by author romitjain opened on Nov 21, 2025
  1

• Online Data Mixing Plugin Checklist

  foundation-model-stack/fms-acceleration#153 · Filter by author kmehant opened on Sep 25, 2025

• test: regression tests show higher memory, throughput and loss values for some settings and models

  foundation-model-stack/fms-acceleration#147 · Filter by author kmehant opened on Jun 13, 2025

• Error in loading quantized models without quantize_config.json

  foundation-model-stack/fms-acceleration#144 · Filter by author srikumar003 opened on May 27, 2025
  1

• test: Regression tests show increased use of memory causing 1 additional experiment to OOM over last tests

  foundation-model-stack/fms-acceleration#128 · Filter by author kmehant opened on Feb 22, 2025
  10

• Logic Inconsistency In ScatterMoE during Expert Parallel

  bughelp wanted
  foundation-model-stack/fms-acceleration#121 · Filter by author fabianlim opened on Jan 24, 2025
  2

• Enable GPTQModel to handle GraniteMoeParallelExperts

  help wanted
  foundation-model-stack/fms-acceleration#112 · Filter by author fabianlim opened on Nov 20, 2024
  1

• Slow down observed for BigCode Santa Coder

  bughelp wantedquestion
  foundation-model-stack/fms-acceleration#110 · Filter by author fabianlim opened on Nov 13, 2024

• ScatterMoE Gradient Norm Needs to be Properly Computed When Used With FSDP

  bughelp wanted
  foundation-model-stack/fms-acceleration#109 · Filter by author fabianlim opened on Nov 11, 2024
  3

• Improve Documentation

  documentationhelp wanted
  foundation-model-stack/fms-acceleration#108 · Filter by author fabianlim opened on Nov 8, 2024

• Extract ScatterMoE TriTon Kernels from Kernel Hyperdrive Fork

  dependencyhelp wantedtriton
  foundation-model-stack/fms-acceleration#105 · Filter by author fabianlim opened on Nov 8, 2024

• Incorporate Liger

  dependencyhelp wantedtriton
  foundation-model-stack/fms-acceleration#104 · Filter by author fabianlim opened on Nov 8, 2024

• ScatterMoE to support LoRA Adapters

  help wantedquestion
  foundation-model-stack/fms-acceleration#103 · Filter by author fabianlim opened on Nov 6, 2024

• ScatterMoE to support Quantized PEFT

  help wantedquestion
  foundation-model-stack/fms-acceleration#101 · Filter by author fabianlim opened on Nov 6, 2024

• Numba JIT TypingErrors Thrown on Multipack Functions

  bugquestion
  foundation-model-stack/fms-acceleration#100 · Filter by author fabianlim opened on Nov 6, 2024

• Register Kernels as AutoGrad Ops

  futurehelp wanted
  foundation-model-stack/fms-acceleration#91 · Filter by author fabianlim opened on Oct 11, 2024
  1

• Support for Iterable Datasets

  foundation-model-stack/fms-acceleration#87 · Filter by author kmehant opened on Sep 30, 2024
  2

• Slowdown and Higher Memory Consumption for GPTQ-LoRA with Bfloat16

  question
  foundation-model-stack/fms-acceleration#84 · Filter by author achew010 opened on Sep 12, 2024
  1

• Ensure Model is Correctly Loaded for Augmentation Purposes

  question
  foundation-model-stack/fms-acceleration#77 · Filter by author fabianlim opened on Aug 29, 2024

• Introduce Liger Fused Cross Entropy Kernel to FOAK Plugin

  tritontuning
  foundation-model-stack/fms-acceleration#76 · Filter by author achew010 opened on Aug 29, 2024
  4

  

• Inconsistency in Padding-Free Benchmarks with Different Transformers Versions

  question
  foundation-model-stack/fms-acceleration#70 · Filter by author achew010 opened on Aug 19, 2024
  3

• Potentially Dead Code in FOAK

  foundation-model-stack/fms-acceleration#59 · Filter by author fabianlim opened on Jul 31, 2024
  

• Quantized Peft Benchmark Experiments Run Out of Memory with Non-Zero Lora Dropout

  question
  foundation-model-stack/fms-acceleration#50 · Filter by author achew010 opened on Jul 12, 2024
  1

• Model Patcher To Work with Generics

  foundation-model-stack/fms-acceleration#49 · Filter by author fabianlim opened on Jul 4, 2024