Skip to content

Commit 81dd232

Browse files
committed
Update release notes to v2.7
Signed-off-by: Timur Rvachov <[email protected]>
1 parent 6195e56 commit 81dd232

File tree

2 files changed

+14
-14
lines changed

2 files changed

+14
-14
lines changed

VERSION

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1 +1 @@
1-
2.7rc1
1+
2.7

docs/docs/main/about/releasenotes-fw.md

Lines changed: 13 additions & 13 deletions
Original file line numberDiff line numberDiff line change
@@ -1,27 +1,27 @@
11
# Release Notes
22

3-
## BioNeMo Framework v2.7rc1
3+
## BioNeMo Framework v2.7
44

55
### Updates & Improvements
66

7-
- Adds a header to SCDL archives, providing improved provenance tracking and supporting future releases. Also adds tracking of the AnnData API coverage in SCDL tests.
8-
This header stores metadata about the archive and its composite arrays, including a version, the array lengths and data types, and information about the RowFeatureIndexes. This adds the features necessary to fix https://github.com/NVIDIA/bionemo-framework/issues/999 as well as implement simple bit-packing of the rowptr, colptr, and data arrays. It also should make SCDL more secure, enable strict compatibility checking, and open the door to more performance improvements. https://github.com/NVIDIA/bionemo-framework/pull/1030
7+
- Evo2 model improvements:
98

10-
- Spike-no-more support for potentially improved training stability: https://github.com/NVIDIA/bionemo-framework/pull/1011
9+
- LoRA fine-tuning by @gabenavarro: https://github.com/NVIDIA/bionemo-framework/pull/980. Note: internal CI coverage of LoRA convergence is still a work in progress; therefore, we cannot guarantee convergence.
10+
- Fix a 2x memory-usage issue during Evo2 generation: https://github.com/NVIDIA/NeMo/pull/14515
11+
- Add flash-decode support in inference: https://github.com/NVIDIA/bionemo-framework/pull/1000
12+
- Update Rotary Embedding and sequence-length defaults to address incorrect checkpoint conversion: https://github.com/NVIDIA/NeMo/pull/14514
13+
- Improvements to tag masking in the Evo2 loss: https://github.com/NVIDIA/bionemo-framework/pull/1008
14+
- Support for [Spike-no-more](https://arxiv.org/abs/2312.16903) to improve training stability: https://github.com/NVIDIA/bionemo-framework/pull/1011
1115

12-
- Improvements in tag masking in Evo2 loss https://github.com/NVIDIA/bionemo-framework/pull/1008
16+
- Added a header to SCDL archives, providing improved provenance tracking and supporting future releases. It also adds tracking of AnnData API coverage in SCDL tests.
17+
This header stores metadata about the archive and its composite arrays, including a version; the array lengths and data types; and information about the RowFeatureIndexes. This adds the features necessary to fix https://github.com/NVIDIA/bionemo-framework/issues/999 as well as to implement simple bit-packing of the rowptr, colptr, and data arrays. It should also make SCDL more secure, enable strict compatibility checking, and open the door to further performance improvements: https://github.com/NVIDIA/bionemo-framework/pull/1030
1318

14-
- Flash decode support in inference https://github.com/NVIDIA/bionemo-framework/pull/1000
15-
16-
- Evo2 Lora feature by @gabenavarro https://github.com/NVIDIA/bionemo-framework/pull/980
17-
18-
- Added partial-conv benchmark for Evo 2 finetune (from checkpoint) and Evo 2 LoRA finetuning https://github.com/NVIDIA/bionemo-framework/pull/1028
19+
- `bionemo-geometric` has been deprecated and removed. The molecular-featurization tooling in this package has moved to [cuik-molmaker](https://github.com/NVIDIA-Digital-Bio/cuik-molmaker).
1920

2021
### Known Issues
2122

22-
- Users have reported issues in the predict script at sequence lengths over 8192 https://github.com/NVIDIA/bionemo-framework/issues/910 and https://github.com/NVIDIA/bionemo-framework/issues/1048 for example.
23-
- PR in progress to update Rotary Embedding and sequence length defaults to support incorrect checkpoint conversion https://github.com/NVIDIA/NeMo/pull/14514
24-
- PR in progress to fix 2x memory usage issue during generation https://github.com/NVIDIA/NeMo/pull/14515
23+
- Users have reported issues with the Evo2 predict script at sequence lengths over 8192; see https://github.com/NVIDIA/bionemo-framework/issues/910 and https://github.com/NVIDIA/bionemo-framework/issues/1048.
24+
- We have removed `libtiff` from the container due to a known vulnerability, [CVE-2025-9900](https://ubuntu.com/security/CVE-2025-9900). `libtiff` isn't directly used in any BioNeMo code; however, users might face issues with e.g. Pillow or other common image-manipulation libraries inside this container.
2525

2626
## BioNeMo Framework v2.6.3
2727

0 commit comments

Comments
 (0)