composer
composer copied to clipboard
mosaicml
Supercharge Your Model Training
how can i plot the eval loss during finetuning using composer? I am able to get f1,acc scores, but by default it's not plotting eval loss in wandb
When I try to save a ResNet18 model to ONNX with the `export_for_inference` function while providing the `apply_blurpool` surgery algorithm as well as the dynamic axes I get the following...
Updates the requirements on [torchmetrics](https://github.com/Lightning-AI/torchmetrics) to permit the latest version. Release notes Sourced from torchmetrics's releases. Minor patch release [1.7.3] - 2025-06-13 Fixed Fixed: ensure WrapperMetric resets wrapped_metric state (#3123)...
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot] avatar"){kind=avatar}
Updates the requirements on [transformers](https://github.com/huggingface/transformers) to permit the latest version. Release notes Sourced from transformers's releases. Release v4.53.0 Gemma3n Gemma 3n models are designed for efficient execution on low-resource devices....
Bumps [junitparser](https://github.com/weiwei/junitparser) from 3.1.2 to 4.0.2. Changelog Sourced from junitparser's changelog. [4.0.2] - 2025-06-23 Fixed write_xml - now supports Path from pathlib. [4.0.1] - 2025-06-23 Fixed write_xml - Remove encoding...
Bumps [sphinxcontrib-katex](https://github.com/hagenw/sphinxcontrib-katex) from 0.9.10 to 0.9.11. Release notes Sourced from sphinxcontrib-katex's releases. Release v0.9.11 Added: support for Python 3.13 Changed: use KaTeX 0.16.22 Removed: support for Python 3.8 Fixed: equation...
Draft PR for supporting automicrobatching on FSDP2 This isn't added yet because we ran into some hiccups with how FSDP2 handles state transitions. As FSDP2 is stateful, it expects the...
*Pretraining a large modernbert model on 1.1 trillion tokens* I am training on a multi-node setup with the following yaml. And it works fine, except that when I run it...
## 🚀 Feature Request Will composer be supporting FSDP2 parallelism available in latest version of pytorch (https://pytorch.org/docs/stable/distributed.fsdp.fully_shard.html)? ## Motivation Pytorch FSDP2 allows better handling of frozen vs non-frozen params which...
Make it possible to define the root temporary directory for composer # What does this PR do? Checkpoints can become very large, and sometimes the temporary file system doesn't have...
