v0.6.1

What's Changed

• fix: disable loss exporting for medium training job by @RobotSail in #347
• build(deps): Bump DavidAnson/markdownlint-cli2-action from 17.0.0 to 18.0.0 by @dependabot in #348
• Update Dependencies to Move DeepSpeed to CUDA Extras by @Maxusmusti in #350

Full Changelog: v0.6.0...v0.6.1

v0.6.0

v0.6.0 - Granite 3.0 and HF Padding-Free Support

What's Changed

• Bump rhysd/actionlint from 1.7.2 to 1.7.3 in /.github/workflows by @dependabot in #234
• feat: automatically select LoRA modules when none are provided by @RobotSail in #166
• add jupyter notebook stuff to requirements dev for experimental work by @RobotSail in #163
• Make data processing optional in run_training() by @MichaelClifford in #220
• Bump pypa/gh-action-pypi-publish from 1.10.2 to 1.10.3 by @dependabot in #249
• Bump sigstore/gh-action-sigstore-python from 2.1.1 to 3.0.0 by @dependabot in #139
• Update pre-commit requirement from <4.0,>=3.0.4 to >=3.0.4,<5.0 by @dependabot in #251
• fix: Add explicit flash_attn requirement for ROCm by @fabiendupont in #255
• Add rocm extra to pyproject.toml by @Maxusmusti in #259
• build(deps): Bump actions/checkout from 4.2.0 to 4.2.1 by @dependabot in #261
• chore: rename 'basic-workflow-tests' to 'e2e-custom' by @nathan-weinberg in #270
• e2e: replace old small job with new medium job by @nathan-weinberg in #277
• fix: incorrect label for AWS medium runner by @nathan-weinberg in #279
• chore: add exit code & tox fix by @RobotSail in #217
• ci: grant HF_TOKEN access to the medium-size E2E CI job by @nathan-weinberg in #285
• ci: add large-size E2E CI job by @nathan-weinberg in #290
• fix: add working directory config to steps in large E2E CI job by @nathan-weinberg in #298
• fix: add remaining missing working directory configs by @nathan-weinberg in #299
• ci: use org variable for AWS EC2 AMI in E2E CI jobs by @nathan-weinberg in #302
• ci: convert med E2E CI job to L4 GPU by @nathan-weinberg in #305
• Implementing HF Padding-Free and GraniteLM Support by @aldopareja @Maxusmusti in #257
• adds basic smoketests for main_ds and data_process CLI args by @JamesKunstle in #280
• build(deps): Bump actions/checkout from 4.2.1 to 4.2.2 by @dependabot in #301
• build(deps): Bump actions/setup-python from 5.2.0 to 5.3.0 by @dependabot in #304
• build(deps): Bump pypa/gh-action-pypi-publish from 1.10.3 to 1.11.0 by @dependabot in #311
• Adding Dolomite Support and Bringing HF Padding-Free into Performance Parity by @Maxusmusti in #312
• build(deps): Bump hynek/build-and-inspect-python-package from 2.9.0 to 2.10.0 by @dependabot in #313
• build(deps): Bump machulav/ec2-github-runner from 2.3.6 to 2.3.7 by @dependabot in #314
• build(deps): Bump rhysd/actionlint from 1.7.3 to 1.7.4 in /.github/workflows by @dependabot in #318
• build(deps): Bump pypa/gh-action-pypi-publish from 1.11.0 to 1.12.0 by @dependabot in #321
• enhancement: enhances bash script with proper syntax by @RobotSail in #323
• Update default chat template to Granite 3.0 template and update token processing by @Maxusmusti in #319
• build(deps): Bump pypa/gh-action-pypi-publish from 1.12.0 to 1.12.2 by @dependabot in #322
• Documentation: Added CHANGELOG.md by @Lo-Mein in #315
• chore: move token classes into chat templates by @jaideepr97 in #324
• feat: add total_samples as a field to logs being emitted by @RobotSail in #326
• Added some safeguards when the necessary imports are not available by @Harthi7 in #291
• fix: incorrect enum selection by @RobotSail in #331
• feat: add toggle to pick legacy chat tmpl for granite by @jaideepr97 in #336
• feat: add log exporting to e2e tests by @RobotSail in #308
• Implement saving FSDP with LoRA by @RobotSail in #295
• deps: change hf accelerate bound, add requirements-hpu.txt by @JamesKunstle in #320

New Contributors

• @MichaelClifford made their first contribution in #220
• @fabiendupont made their first contribution in #255
• @Lo-Mein made their first contribution in #315
• @Harthi7 made their first contribution in #291

Full Changelog: v0.5.5...v0.6.0

v0.6.0a1

v0.6.0 Alpha 1 - New Granite 3.0 Template Support (Preview)

What's Changed

• build(deps): Bump hynek/build-and-inspect-python-package from 2.9.0 to 2.10.0 by @dependabot in #313
• build(deps): Bump machulav/ec2-github-runner from 2.3.6 to 2.3.7 by @dependabot in #314
• build(deps): Bump rhysd/actionlint from 1.7.3 to 1.7.4 in /.github/workflows by @dependabot in #318
• build(deps): Bump pypa/gh-action-pypi-publish from 1.11.0 to 1.12.0 by @dependabot in #321
• enhancement: enhances bash script with proper syntax by @RobotSail in #323
• Update default chat template to Granite 3.0 template and update token processing by @Maxusmusti in #319
• build(deps): Bump pypa/gh-action-pypi-publish from 1.12.0 to 1.12.2 by @dependabot in #322
• Documentation: Added CHANGELOG.md by @Lo-Mein in #315
• chore: move token classes into chat templates by @jaideepr97 in #324

New Contributors

• @Lo-Mein made their first contribution in #315

Full Changelog: v0.6.0a0...v0.6.0a1

v0.6.0a0

v0.6.0 Alpha 0 - Granite 3.0 8B and HF Padding-Free Support (Preview)

What's Changed

• Bump rhysd/actionlint from 1.7.2 to 1.7.3 in /.github/workflows by @dependabot in #234
• feat: automatically select LoRA modules when none are provided by @RobotSail in #166
• add jupyter notebook stuff to requirements dev for experimental work by @RobotSail in #163
• Make data processing optional in run_training() by @MichaelClifford in #220
• Bump pypa/gh-action-pypi-publish from 1.10.2 to 1.10.3 by @dependabot in #249
• Bump sigstore/gh-action-sigstore-python from 2.1.1 to 3.0.0 by @dependabot in #139
• Update pre-commit requirement from <4.0,>=3.0.4 to >=3.0.4,<5.0 by @dependabot in #251
• fix: Add explicit flash_attn requirement for ROCm by @fabiendupont in #255
• Add rocm extra to pyproject.toml by @Maxusmusti in #259
• build(deps): Bump actions/checkout from 4.2.0 to 4.2.1 by @dependabot in #261
• chore: rename 'basic-workflow-tests' to 'e2e-custom' by @nathan-weinberg in #270
• e2e: replace old small job with new medium job by @nathan-weinberg in #277
• fix: incorrect label for AWS medium runner by @nathan-weinberg in #279
• chore: add exit code & tox fix by @RobotSail in #217
• ci: grant HF_TOKEN access to the medium-size E2E CI job by @nathan-weinberg in #285
• ci: add large-size E2E CI job by @nathan-weinberg in #290
• fix: add working directory config to steps in large E2E CI job by @nathan-weinberg in #298
• fix: add remaining missing working directory configs by @nathan-weinberg in #299
• ci: use org variable for AWS EC2 AMI in E2E CI jobs by @nathan-weinberg in #302
• ci: convert med E2E CI job to L4 GPU by @nathan-weinberg in #305
• Implementing HF Padding-Free and GraniteLM Support by @aldopareja @Maxusmusti in #257
• adds basic smoketests for main_ds and data_process CLI args by @JamesKunstle in #280
• build(deps): Bump actions/checkout from 4.2.1 to 4.2.2 by @dependabot in #301
• build(deps): Bump actions/setup-python from 5.2.0 to 5.3.0 by @dependabot in #304
• build(deps): Bump pypa/gh-action-pypi-publish from 1.10.3 to 1.11.0 by @dependabot in #311
• Adding Dolomite Support and Bringing HF Padding-Free into Performance Parity by @Maxusmusti in #312

New Contributors

• @MichaelClifford made their first contribution in #220
• @fabiendupont made their first contribution in #255

Full Changelog: v0.5.5...v0.6.0a0

v0.5.5

What's Changed

• e2e: replace old small job with new medium job (backport #277) by @mergify in #283
• fix: incorrect label for AWS medium runner (backport #279) by @mergify in #284
• ci: grant HF_TOKEN access to the medium-size E2E CI job (backport #285) by @mergify in #286
• chore: add exit code & tox fix (backport #217) by @mergify in #282

Full Changelog: v0.5.4...v0.5.5

v0.5.4

What's Changed

• Add rocm extra to pyproject.toml (backport #259) by @mergify in #260

Full Changelog: v0.5.3...v0.5.4

v0.5.3

What's Changed

• fix: Add explicit flash_attn requirement for ROCm (backport #255) by @mergify in #258

New Contributors

• @mergify made their first contribution in #258

Full Changelog: v0.5.2...v0.5.3

v0.5.2 - Fix Pretraining Masking

What's Changed

• fix: improve linting and automation by @RobotSail in #172
• Fix pretrain token list->int for masking by @Maxusmusti in #244

Full Changelog: v0.5.1...v0.5.2

v0.5.1

What's Changed

• fix: updates sorting logic to correctly compare numbers by @RobotSail in #238

Full Changelog: v0.5.0...v0.5.1

v0.5.0 FSDP and Full-State Checkpoint Resuming

What's Changed

• fix lint error to make CI happy by @RobotSail in #173
• Add license identifiers by @danmcp in #157
• Fix typos by @danmcp in #158
• feat: add e2e test for instructlab CI by @RobotSail in #174
• Ap/fix multipack for non granite models by @aldopareja in #169
• feat: add mergify by @RobotSail in #187
• ci: update runner labels to uniquely identify instance sizes by @nathan-weinberg in #180
• ci: minor cleanup of E2E job by @nathan-weinberg in #189
• Fixing e2e to use relative path for working-directory by @danmcp in #191
• Fix linting error and missing quote by @danmcp in #192
• Fix typos by @danmcp in #190
• Bump step-security/harden-runner from 2.8.1 to 2.9.1 by @dependabot in #159
• ci: add AWS tags to show github ref and PR num for all jobs by @nathan-weinberg in #196
• Bump pypa/gh-action-pypi-publish from 1.9.0 to 1.10.2 by @dependabot in #219
• Adding FSDP Support to Training Library by @aldopareja @Maxusmusti @RobotSail in #213
• Bump actions/setup-python from 5.1.0 to 5.2.0 by @dependabot in #193
• Bump rhysd/actionlint from 1.7.1 to 1.7.2 in /.github/workflows by @dependabot in #221
• ci: fix lint action by @nathan-weinberg in #227
• Bump hynek/build-and-inspect-python-package from 2.6.0 to 2.9.0 by @dependabot in #198
• Bump DavidAnson/markdownlint-cli2-action from 16.0.0 to 17.0.0 by @dependabot in #199
• Bump step-security/harden-runner from 2.9.1 to 2.10.1 by @dependabot in #224
• Docs: Update training README by @kelbrown20 in #218
• changes StreamablePopen to return a process and implement listening by @JamesKunstle in #229
• switch -T to -a by @cdoern in #222
• github: add stale bot to training repo by @nathan-weinberg in #226
• fix: markdown lint error and mergify bug by @nathan-weinberg in #231
• Bump actions/checkout from 4.1.7 to 4.2.0 by @dependabot in #230
• Fix generic chat template saved to tokenizer for generation by @Maxusmusti in #232
• adds Accelerate full-state (opt, lr_sched, params) by @JamesKunstle in #235

New Contributors

• @danmcp made their first contribution in #157
• @kelbrown20 made their first contribution in #218

Full Changelog: v0.4.2...v0.5.0