v0.1.14

What's Changed

• Record latency histograms in base inference engine by @nikg4 in #1702
• Feat: add falcon-e integration by @younesbelkada in #1705
• [tiny] Minor update to fix the failing pre-commit checks by @oelachqar in #1707
• Add collator kwargs field to DataParams by @oelachqar in #1708
• [vision] Add option to process images individually by @oelachqar in #1706
• Update dev_setup.md to correct the order of steps by @ryan-arman in #1709
• Add configs for molmo support by @oelachqar in #1710
• [tiny] fix pre-commits checks on a fresh install by @oelachqar in #1711
• Add config for the Molmo O variant by @oelachqar in #1712
• Add experimental molmo grpo config and train aliases by @oelachqar in #1713
• Update installation.md to fix subversion handling by adding required … by @ryan-arman in #1715
• Frontier: Fix -n param in launcher script by @nikg4 in #1720
• Fix Falcon H1 dependency setup by @wizeng23 in #1723
• letter count notebook improvements by @penfever in #1697
• [vision] Update vision feature generator to support training on completions only by @oelachqar in #1722
• [tiny] fix bug with vl collator by @oelachqar in #1725
• Add data synthesis config, params, and unit tests by @jgreer013 in #1700
• Add support for additional exception types for remote inference engine, as well as fast failing for non-retryable status codes. by @jgreer013 in #1704
• Adds DPO + QLoRA example for Falcon-H1 by @stefanwebb in #1719
• Update inference to always write intermediate results to file. by @jgreer013 in #1724
• Added doc for new QLoRA param by @stefanwebb in #1727
• Readme for Falcon-E and note on extra dependencies required by @stefanwebb in #1729
• Add generic vision dataset by @oelachqar in #1726
• [tiny][bug] make git cmd optional by @oelachqar in #1730
• [tiny][bug] Add missing molmo feature by @oelachqar in #1731
• [tiny] Update phi3-vision configs to use oumi trainer by @oelachqar in #1733
• Minor bugfixes for 2 clouds in launcher code by @nikg4 in #1728
• Update dev_setup.md to add additional instructions by @ryan-arman in #1736
• Update trl to 0.18 by @wizeng23 in #1693
• Update Verl trainer to export models in HF format by @nikg4 in #1714
• Add lmms-lab/multimodal-open-r1-8k-verified dataset by @oelachqar in #1732
• Add placeholders for DCVLR by @oelachqar in #1738
• add debug logging capabilities to collators by @aniruddh-alt in #1678
• [bug] update trainer to save processor when training with fsdp by @oelachqar in #1742
• Add model revision param by @oelachqar in #1740
• Add ability to customize HF model config via model.model_kwargs by @oelachqar in #1741
• Add docker release workflow by @oelachqar in #1743
• [bug] fix rank/local rank parsing for docker env by @oelachqar in #1747
• deps: Update vLLM to 0.8.3 by @wizeng23 in #1739
• [docs] update dcvlr readme by @oelachqar in #1748
• Dcvlr by @penfever in #1750

New Contributors

• @younesbelkada made their first contribution in #1705
• @ryan-arman made their first contribution in #1709
• @stefanwebb made their first contribution in #1719
• @aniruddh-alt made their first contribution in #1678

Full Changelog: v0.1.13...v0.1.14