NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #1737 most downloaded on PyPI
Library for utilization of compressed safetensors of neural network models
Last release 2 days ago
15 Sep 2026
Ships on a steady schedule
a new release about every 9 days
Most releases are documented
notes for 28 of 32 stable releases
Nothing withdrawn
no release was ever pulled
2 years old
208 releases · first in 2024
Nothing published for this version
Nothing published for this version
Nothing published for this version
Remove upper limit on torch version in setup.py by @dsikka in https://github.com/vllm-project/compressed-tensors/pull/674
One column per month.
Full Changelog: https://github.com/vllm-project/compressed-tensors/compare/0.15.0...0.15.0.1
[Bugfix] Support N-dimensional tensors in pack_to_int32 and unpack_from_int32 by @Yatimai in https://github.com/vllm-project/compressed-tensors/pull/6
update_offload by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/617offload_folder when performing disk offloading by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/602style and test claude skills by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/616generate_gparam calculation to handle NaNs and Infs by @dsikka in https://github.com/vllm-project/compressed-tensors/pull/637update_offload_parameter by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/633Full Changelog: https://github.com/vllm-project/compressed-tensors/compare/0.14.0.1...0.15.0
update_offload by @kylesayrs in #617offload_folder when performing disk offloading by @kylesayrs in #602style and test claude skills by @kylesayrs in #616generate_gparam calculation to handle NaNs and Infs by @dsikka in #637update_offload_parameter by @kylesayrs in #633Full Changelog: 0.14.0.1...0.15.0
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Patch release to add support for distributed disk cache and fix bugs related to file writing. For more information see #617
Full Changelog: https://github.com/vllm-project/compressed-tensors/compare/0.14.0...0.14.0.1
Full Changelog: 0.14.0...0.14.0.1
[Deprecation] Add deprecation warning for marlin24 format by @Etelis in https://github.com/vllm-project/compressed-tensors/pull/544
delete_offload_parameter, add clear_quantization by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/539DistributedCPUCache by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/534DistributedDeviceCache by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/568DiskCache, DistributedDiskCache by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/535update_offload_parameter more async and direct (2) by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/576update_parameter_data by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/588clear_quantization by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/596load_offloaded_model by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/605Full Changelog: https://github.com/vllm-project/compressed-tensors/compare/0.13.0...0.14.0
delete_offload_parameter, add clear_quantization by @kylesayrs in #539DistributedCPUCache by @kylesayrs in #534DistributedDeviceCache by @kylesayrs in #568DiskCache, DistributedDiskCache by @kylesayrs in #535update_offload_parameter more async and direct (2) by @kylesayrs in #576update_parameter_data by @kylesayrs in #588clear_quantization by @kylesayrs in #596load_offloaded_model by @kylesayrs in #605Full Changelog: 0.13.0...0.14.0
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
[Quantization] Refactor initialize for activation shape inference by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/476
find_name_or_class_matches util by @kylesayrs in https://github.com/vllm-project/compressed-tensors/pull/488Full Changelog: https://github.com/vllm-project/compressed-tensors/compare/0.12.2...0.13.0
find_name_or_class_matches util by @kylesayrs in #488Full Changelog: 0.12.2...0.13.0
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
remove deprecated safe_permute by @brian-dellabetta in https://github.com/neuralmagic/compressed-tensors/pull/471
Full Changelog: https://github.com/neuralmagic/compressed-tensors/compare/0.12.1...0.12.2
Full Changelog: 0.12.1...0.12.2
Nothing published for this version
Nothing published for this version
[Patch Fix] Add get_missing_module_keys to support transformers lower bound by @dsikka in https://github.com/neuralmagic/compressed-tensors/pull/479
get_missing_module_keys to support transformers lower bound by @dsikka in https://github.com/neuralmagic/compressed-tensors/pull/479Full Changelog: https://github.com/neuralmagic/compressed-tensors/compare/0.12.0...0.12.1
Nothing published for this version
[Utils] Deprecate safe_permute by @kylesayrs in https://github.com/neuralmagic/compressed-tensors/pull/464
apply_quantization_config by @kylesayrs in https://github.com/neuralmagic/compressed-tensors/pull/433llmcompressor.transformers.oneshot in examples by @brian-dellabetta in https://github.com/neuralmagic/compressed-tensors/pull/422is_module_offloaded and update_prefix_dict by @kylesayrs in https://github.com/neuralmagic/compressed-tensors/pull/366safe_permute by @kylesayrs in https://github.com/neuralmagic/compressed-tensors/pull/464frozendict dependency, use types.MappingProxyType instead by @brian-dellabetta in https://github.com/neuralmagic/compressed-tensors/pull/469TransformScheme.head_dim for compatibility with vllm by @brian-dellabetta in https://github.com/neuralmagic/compressed-tensors/pull/472Full Changelog: https://github.com/neuralmagic/compressed-tensors/compare/0.11.0...0.12.0
Your coding agent can read these notes before it upgrades. Set up the MCP server →