Unclaimed project
Are you a maintainer of burn? Claim this project to take control of your public changelog and roadmap.
Changelog
burn
Burn is a next generation tensor library and Deep Learning Framework that doesn't compromise on flexibility, efficiency and portability.
Back to changelogNew
v0.20.0-pre.6
What's Changed
- doc warning fix by @crutcher in https://github.com/tracel-ai/burn/pull/4130
- Fix tch bf16 into_data by @laggui in https://github.com/tracel-ai/burn/pull/4142
- Update raspberry-pi-pico example to use the Pico 2, and burnpack by @BjornTheProgrammer in https://github.com/tracel-ai/burn/pull/4132
- Unify all_reduce
LocalCollectiveClient operation handling. by @crutcher in https://github.com/tracel-ai/burn/pull/4125
- Add direct tensor snapshot retrieval API to ModuleStore by @antimora in https://github.com/tracel-ai/burn/pull/4131
- Fix outer-scope variable references in ONNX subgraphs (If/Loop/Scan) by @antimora in https://github.com/tracel-ai/burn/pull/4119
- Add removed docs for tensor equal_elem by @laggui in https://github.com/tracel-ai/burn/pull/4145
- Add ceil_mode support to pooling operations (MaxPool, AvgPool) by @antimora in https://github.com/tracel-ai/burn/pull/4112
- chore: Update cubecl by @wingertge in https://github.com/tracel-ai/burn/pull/4134
- Implement Slice iterator and utility methods. by @crutcher in https://github.com/tracel-ai/burn/pull/4042
- Bump peter-evans/create-pull-request from 7 to 8 by @dependabot[bot] in https://github.com/tracel-ai/burn/pull/4148
Add slice_dyn, slice_assign_dyn, and slice_fill_dyn variants. by @crutcher in https://github.com/tracel-ai/burn/pull/4127Add Reshape scalar optimization and Gather scalar input support by @antimora in https://github.com/tracel-ai/burn/pull/4146Shape FromStr/ToString by @crutcher in https://github.com/tracel-ai/burn/pull/4143Add contiguous reindexing for non-contiguous layer indices by @antimora in https://github.com/tracel-ai/burn/pull/4150Add warmup epochs to MetricEarlyStoppingStrategy. (#3970) by @crutcher in https://github.com/tracel-ai/burn/pull/4041fix(onnx): Use activation function for GELU codegen instead of non-existent tensor method by @antimora in https://github.com/tracel-ai/burn/pull/4161Refactor more basic ops by @laggui in https://github.com/tracel-ai/burn/pull/4156Refactor LocalCollectiveServer for improved clarity and error handling by @crutcher in https://github.com/tracel-ai/burn/pull/4126Fix typo in comment for logger_task function by @crutcher in https://github.com/tracel-ai/burn/pull/4159Refactor configurable backend tests (no more testgen macros) by @laggui in https://github.com/tracel-ai/burn/pull/4129Zero-copy loading for embedded burnpack weights by @antimora in https://github.com/tracel-ai/burn/pull/4154Fix candle cuda imports by @laggui in https://github.com/tracel-ai/burn/pull/4171Backends no longer depend on burn-tensor, but strictly burn-backend by @laggui in https://github.com/tracel-ai/burn/pull/4169Chore/update cubek cubecl by @nathanielsimard in https://github.com/tracel-ai/burn/pull/4172Add ONNX CumSum operator support by @antimora in https://github.com/tracel-ai/burn/pull/4162Add backend supports_dtype by @laggui in https://github.com/tracel-ai/burn/pull/4155Fix attention shapes and out rank by @laggui in https://github.com/tracel-ai/burn/pull/4192Fix matmul & reduce execute fuse no autotune by @laggui in https://github.com/tracel-ai/burn/pull/4193Fix output dtype for argmin / argmax by @laggui in https://github.com/tracel-ai/burn/pull/4195Add flatten_dims method to Shape and refactor tensor flattening API by @crutcher in https://github.com/tracel-ai/burn/pull/4189Return slice for each dimension in shape by @laggui in https://github.com/tracel-ai/burn/pull/4152Make xtask validate run no-std checks first. by @crutcher in https://github.com/tracel-ai/burn/pull/4198Fix: CubeCL Reduce by @nathanielsimard in https://github.com/tracel-ai/burn/pull/4197Reorganize and tracing::instrument collective operations. by @crutcher in https://github.com/tracel-ai/burn/pull/4157Log running values by @Charles23R in https://github.com/tracel-ai/burn/pull/4199Remove global ONNX opset version restriction, recommend opset 16 by @antimora in https://github.com/tracel-ai/burn/pull/4168Fix dtype preservation when loading tensors in burn-store by @antimora in https://github.com/tracel-ai/burn/pull/4194Fix TchTensor::from_data bf16 by @laggui in https://github.com/tracel-ai/burn/pull/4203Perf/reduce cpu + Fix OOB by @nathanielsimard in https://github.com/tracel-ai/burn/pull/4204feat: Implicit GEMM weight gradients for convolution by @wingertge in https://github.com/tracel-ai/burn/pull/4182Fix checkpoint and summary log level by @J-F-Liu in https://github.com/tracel-ai/burn/pull/4201fix: handle 1D slope when importing prelu from onnx by @mertalev in https://github.com/tracel-ai/burn/pull/4205Zero-copy tensor loading for NdArray backend by @antimora in https://github.com/tracel-ai/burn/pull/4178Fix quantized tensor storage data length calculation by @antimora in https://github.com/tracel-ai/burn/pull/4180Fix handling scalar scan outputs in ONNX loop nodes by @antimora in https://github.com/tracel-ai/burn/pull/4210Perf/improve reduce autotuning + plane non uniform control flow check by @nathanielsimard in https://github.com/tracel-ai/burn/pull/4208Add ONNX external data support for models >2GB by @antimora in https://github.com/tracel-ai/burn/pull/4158Update/cubek by @louisfd in https://github.com/tracel-ai/burn/pull/4214Refactor: Replace canonicalize_dim with expect_dim by @crutcher in https://github.com/tracel-ai/burn/pull/4196fix: handle negative indices in onnx gather op by @mertalev in https://github.com/tracel-ai/burn/pull/4207Refactor/cube dim by @nathanielsimard in https://github.com/tracel-ai/burn/pull/4217Refactor: Consolidate shape and slice error handling into ExpressionError by @crutcher in https://github.com/tracel-ai/burn/pull/4218Update: CubeK by @louisfd in https://github.com/tracel-ai/burn/pull/4222feat: Accelerated convolution data gradient by @wingertge in https://github.com/tracel-ai/burn/pull/4220Fix repeat 0 times by @laggui in https://github.com/tracel-ai/burn/pull/4216Burn train api refactor by @Charles23R in https://github.com/tracel-ai/burn/pull/4223Chore/pre release 6 by @nathanielsimard in https://github.com/tracel-ai/burn/pull/4224