Releases: macrocosm-os/apex
Releases · macrocosm-os/apex
v2.19.7
What's Changed
- Fix logprob scoring when failed to connect to API by @dbobrenko in #765
- Bump max model len to max supported by @dbobrenko in #766
- v2.19.7 by @dbobrenko in #767
Full Changelog: v2.19.6...v2.19.7
v2.19.6
What's Changed
- SN1 validator containerization by @mpnowacki-reef in #720
- Default To Mistral for Model by @bkb2135 in #747
- Fix testnet validator run by @dbobrenko in #749
- API enhancements by @dbobrenko in #752
- Sn1 509 Fix Precommit, Remove Ruff by @richwardle in #751
- Depreciate Target Uids by @richwardle in #750
- Sn1 502 remove validator generation for each reference by @richwardle in #745
- Gather reliability stats by coldkeys by @dbobrenko in #753
- Log Organic Scores and Timing by @richwardle in #754
- Bump v2.19.5 by @dbobrenko in #757
- v2.19.5 by @dbobrenko in #758
- Fix/verify uid in weight syncer by @bkb2135 in #756
- Move embeddings into docker utils by @bkb2135 in #760
- Remove Model Zoo and Config by @bkb2135 in #762
- Fix errors, remove hardcode, add error handling by @dbobrenko in #763
- v2.19.6 by @bkb2135 in #764
New Contributors
- @mpnowacki-reef made their first contribution in #720
Full Changelog: v2.19.4...v2.19.6
v2.19.4
What's Changed
- v2.18.3: Check for full completion, smooth chunk stream by @dbobrenko in #683
- Sync staging branch with main by @dbobrenko in #684
- 0 sum v2 by @Hollyqui in #678
- Fix api by @bkb2135 in #686
- Revisit logprobs checks by @dbobrenko in #688
- Comment health process by @bkb2135 in #689
- v2.19.0 by @bkb2135 in #687
- Fix web retrieval route by @bkb2135 in #690
- Enhance logprobs checks by @dbobrenko in #692
- Remove LLama Model by @bkb2135 in #694
- Fix Organic Scoring by @bkb2135 in #691
- Fix Stream Collection by @bkb2135 in #693
- Sync staging with main by @dbobrenko in #697
- Fix/append left msr by @bkb2135 in #695
- Bump Version Number by @bkb2135 in #700
- Fix MSRv2 by @dbobrenko in #702
- Fix BOS by @dbobrenko in #699
- Adjust inference reward scales by @dbobrenko in #703
- Reduce similarity threshold by @dbobrenko in #704
- v2.19.2 by @bkb2135 in #698
- Sn1 464 deep research truncation fix by @richwardle in #701
- Add Tools to API by @bkb2135 in #706
- Fix/use stream in deep by @bkb2135 in #705
- Reduce wandb size by @dbobrenko in #707
- Update task_registry.py by @bkb2135 in #709
- Use sqlite for job_store.py by @gevagorou-mai in #714
- Add submit and poll endpoints by @gevagorou-mai in #712
- Fix and optimize API, add strict checks to web, etc by @dbobrenko in #715
- Restore task registry by @dbobrenko in #716
- If client disconnects score whatever has been sent by @bkb2135 in #710
- v2.19.3 by @dbobrenko in #718
- API and deep researcher fixes by @dbobrenko in #721
- readme update by @richwardle in #722
- Sync Staging and Main by @bkb2135 in #723
- GEN-1220: Return response in chunks in the poll endpoint by @gevagorou-mai in #726
- Sn1 481 enhance docs for investor scraping by @richwardle in #728
- Staging by @bkb2135 in #731
- Merge Main to Staging by @bkb2135 in #732
- Switch to multi-modal mistral by @bkb2135 in #708
- Sn1 480 gan logging issues by @richwardle in #717
- Update pyproject.toml by @bkb2135 in #733
- Update README.md by @bkb2135 in #734
- Sn1 479 precommit weirdness by @richwardle in #725
- Fix API; bump vllm 0.8.5 by @dbobrenko in #736
- Fix tests and API balance by @dbobrenko in #737
- Create Weight Synchronizer by @bkb2135 in #730
- Fix README typos by @richwardle in #738
- SN1-482-gan-fixes by @richwardle in #739
- Remove Send to Self Logic by @richwardle in #740
- Remove Mistral Model by @bkb2135 in #741
- Add Text models by @bkb2135 in #742
- v2.19.4 by @bkb2135 in #735
New Contributors
- @gevagorou-mai made their first contribution in #714
Full Changelog: v2.18.2...v2.19.4
v2.18.2
What's Changed
- Exit processes on shutdown, reduce subtensor calls, etc by @dbobrenko in #680
Full Changelog: v2.18.1...v2.18.2
v2.18.1
v2.18.0
What's Changed
- Increase connect timeout; improve web checks by @dbobrenko in #668
- SN1-447: add logits by @Hollyqui in #660
- Fix DDGS, HF sampling, adjust timeouts by @dbobrenko in #669
- Fix Model = None Inference by @bkb2135 in #670
- Fix max model len by @bkb2135 in #671
- Clean up logs by @dbobrenko in #673
- v2.18.0 by @dbobrenko in #672
Full Changelog: v2.17.11...v2.18.0
v2.17.11
v2.17.8
What's Changed
- Fix/improving validator performance by @Hollyqui in #648
- SN1-424 remove wiki answers by @Hollyqui in #639
- Add UIDs, availability and metagraph API fallbacks by @dbobrenko in #652
- Integrate deep research with the api by @bkb2135 in #655
- Bump Version Number by @bkb2135 in #656
- v2.17.8 by @bkb2135 in #657
Full Changelog: v2.17.7...v2.17.8
v2.17.7
v2.17.6
What's Changed
- Double the weight of organics by @bkb2135 in #598
- Randomize Target Results by @bkb2135 in #606
- Fix API availability and scoring loops by @dbobrenko in #608
- fix/improving msr by @Hollyqui in #609
- API dependencies hotfix; version bump v2.17.4 by @dbobrenko in #610
- Enforce API Key for Web Retrieval And Allow For Specified UIDS by @richwardle in #607
- Add API scoring for non-streaming completions by @dbobrenko in #612
- Hotfix: Wait for loop execution first by @dbobrenko in #615
- Bump v2.17.5 by @bkb2135 in #616
- v2.17.5 by @dbobrenko in #617
- Fix attr errors in web retrieval by @bkb2135 in #618
- Add API uid fallback, handle exceptions in loop runner by @dbobrenko in #620
- Remove past websites by @bkb2135 in #619
- Remove Notebooks by @richwardle in #623
- Default to random miners by @bkb2135 in #622
- Adding cache by @Hollyqui in #625
- Make WebRetrievalRewardModel hashable by @Hollyqui in #629
- Add API systemd start/stop scripts by @dbobrenko in #626
- Making Reward Function Async & Various Optimisations by @Hollyqui in #628
- SN1-406-improve-api-docs by @richwardle in #627
- Add Past Websites Files by @richwardle in #630
- Add Await To Inference Reward Model by @richwardle in #632
- Fix pydantic types; move autoawq to poetry by @dbobrenko in #635
- Separate Prompting, Remove TTI Endpoint, Add Json Flag by @richwardle in #633
- Add Async Timeout by @richwardle in #636
- Pass Timeout Through by @richwardle in #637
- Teardown wandb at exit by @bkb2135 in #624
- Hotfix HF model initialization by @dbobrenko in #638
- Allow langflow arguments to be properly parsed by completion endpoint by @bkb2135 in #640
- v2.17.6 by @bkb2135 in #621
Full Changelog: v2.17.4...v2.17.6