v0.0.3-alpha36

What's Changed

• fixed "TypeError: 'NoneType' object is not iterable" for reward model… by @egg-west in #3587
• Use rng_seed param when creating custom dataset sampler by @andreaskoepf in #3592
• Update Flash Attention forward for Llama 2: by @jordiclive in #3595
• 3517 - Send reports to discord via celery task by @nablabits in #3566
• Bump tailwindcss from 3.3.2 to 3.3.3 in /website by @dependabot in #3601
• Bump next from 13.4.8 to 13.4.12 in /website by @dependabot in #3600
• Fix Rope scaling by @shahules786 in #3598
• Cleanup handle_worker() in preparation for #2815 (Stop generation early) by @0xfacade in #3573
• Add dataset loader for MegaCodeTraining112k & Evol-Instruct-Code-80k-v1 by @andreaskoepf in #3605
• Dolphin gpt 3.5 data mix by @shahules786 in #3606
• Update: Pydantic version from 1.10.4 to 1.10.7 by @Ali1858 in #3615
• Changes for orcacode experiment by @andreaskoepf in #3612
• Added support for LLongMA by @shahules786 in #3616
• Custom instructions feature by @draganjovanovich in #3597
• update ru-translation by @0x22almostEvil in #3619
• Bump @storybook/testing-library from 0.1.0 to 0.2.0 in /website by @dependabot in #3620
• Bump pino-pretty from 10.0.0 to 10.2.0 in /website by @dependabot in #3622
• Fix Typo by @pauloaferreiraa in #3610
• Add Wikipedia Persian Dataset by @pourmand1376 in #3629
• Add Isna Persian Dataset by @pourmand1376 in #3631

New Contributors

• @egg-west made their first contribution in #3587
• @nablabits made their first contribution in #3566
• @0xfacade made their first contribution in #3573
• @Ali1858 made their first contribution in #3615
• @pauloaferreiraa made their first contribution in #3610
• @pourmand1376 made their first contribution in #3629

Full Changelog: v0.0.3-alpha35...v0.0.3-alpha36