Ecosyste.ms: Timeline
Browse the timeline of events for every public repo on GitHub. Data updated hourly from GH Archive.
shuttie pushed 1 commit to main nixiesearch/onnx-convert
- add skip_shape_inference to work with bge-m3 07a1ace
shuttie opened an issue on nixiesearch/nixiesearch
add comments to all code snippets in mkdocs
like in pydantic docs: https://fastapi.tiangolo.com/it/python-types/
shuttie opened a pull request on nixiesearch/nixiesearch
fix int overflow in download progress logs
fixes #371
shuttie created a branch on nixiesearch/nixiesearch
fix/log-size-overflow - Hybrid search engine, combining best features of text and semantic search worlds
shuttie opened an issue on nixiesearch/nixiesearch
Downloading large GGUFs overflows the progess counter
``` nixiesearch-1 | 18:55:43.638 INFO ai.nixiesearch.core.PrintProgress$ - processed -1748834219 bytes, rate: 9 MB/s nixiesearch-1 | 18:55:44.644 INFO ai.nixiesearch.core.PrintProgress$ - p...
shuttie opened an issue on nixiesearch/nixiesearch
Lucene index back-compat is broken
We wrap the exact version-specific codec in NixieCodec, but should rely on Lucene logic for that.
shuttie created a tag on nixiesearch/nixiesearch
0.3.4 - Hybrid search engine, combining best features of text and semantic search worlds
shuttie published a release on nixiesearch/nixiesearch
0.3.4
## A bugfix release * fix DirectoryStateClient spamming logs with index refresh events by @shuttie in https://github.com/nixiesearch/nixiesearch/pull/316 * fix broken links in github readme by @s...
shuttie pushed 2 commits to master nixiesearch/nixiesearch
shuttie closed an issue on nixiesearch/nixiesearch
Alias support is broken
Aliases defined in config file are rn ignored.
shuttie closed a pull request on nixiesearch/nixiesearch
fix alias support
fixes #272 now all routes correctly handle aliased paths
shuttie opened a pull request on nixiesearch/nixiesearch
fix alias support
fixes #272 now all routes correctly handle aliased paths
shuttie created a branch on nixiesearch/nixiesearch
fix/alias-support - Hybrid search engine, combining best features of text and semantic search worlds
shuttie pushed 2 commits to master nixiesearch/nixiesearch
shuttie closed an issue on nixiesearch/nixiesearch
Model cache consistency is not checked on load
So if you interrupt model download, cache can be inconsistent. We should at least check the file size when loading.
shuttie closed a pull request on nixiesearch/nixiesearch
make model file cache atomic
fixes: https://github.com/nixiesearch/nixiesearch/issues/364 Now the cache saving is atomic: happens only on successful model download.
shuttie closed an issue on nixiesearch/nixiesearch
Loading e5-multilingual-large results in OOM: Java Heap Space on default docker container
Should we have dynamic -Xmx settings?
shuttie pushed 2 commits to master nixiesearch/nixiesearch
shuttie closed a pull request on nixiesearch/nixiesearch
use mmap for onnx model loading
Should fix https://github.com/nixiesearch/nixiesearch/issues/363 As we're now not loading the model itself in jvm heap, and just mmapping it directly into the onnx runtime.
shuttie created a branch on nixiesearch/nixiesearch
fix/atomic-model-cache - Hybrid search engine, combining best features of text and semantic search worlds
shuttie opened a pull request on nixiesearch/nixiesearch
use mmap for onnx model loading
Should fix https://github.com/nixiesearch/nixiesearch/issues/363 As we're now not loading the model itself in jvm heap, and just mmapping it directly into the onnx runtime.
shuttie created a branch on nixiesearch/nixiesearch
fix/use-mmap-onnx - Hybrid search engine, combining best features of text and semantic search worlds
shuttie pushed 3 commits to master nixiesearch/nixiesearch
shuttie pushed 3 commits to master nixiesearch/nixiesearch
shuttie closed an issue on nixiesearch/nixiesearch
Wrong number of onnx inputs for XLM-based e5-multilingual
``` nixiesearch-1 | 15:12:06.676 ERROR ai.nixiesearch.api.API$ - ai.onnxruntime.OrtException: Unexpected number of inputs, expected [1,2) found 3 nixiesearch-1 | ai.onnxruntime.OrtException: Un...