[Roadmap] SparseML Roadmap Q2 2024

[mgoin]

Short-term Work

Issues or PRs that the NM team are planning to tackle for this quarter:

• Dependencies

  • [x] Forkless Transformers https://github.com/neuralmagic/sparseml/pull/2199
  • [x] Upgrade Transformers to latest version https://github.com/neuralmagic/sparseml/pull/2204
  • [x] Upgrade PyTorch to latest version https://github.com/neuralmagic/sparseml/pull/2247 https://github.com/neuralmagic/sparseml/pull/2259
  • [x] Upgrade Pydantic to v2 https://github.com/neuralmagic/sparseml/pull/2248

• SparseGPT one-shot pruning for Transformers

  • [ ] General model support AutoModel https://github.com/neuralmagic/sparseml/pull/2003

• Saving compressed models

  • [x] Export sparse models w/ sparsity_config to Transformers https://github.com/neuralmagic/sparseml/pull/2190
  • [ ] Export weight-quantized models w/ quant_config to Transformers
  • [ ] Export W8A8-quantized models w/ quant_config to Transformers

Long-term Tracked Feature Requests/Issues