Release v0.17.2

• feat: rename to generative #354
• fix prompt texts + timeout #353
• fix: default low conf (without PR)
• fix: default low conf for universal (without PR)

Release v0.17.1

• Fix/hot fixes #351

Release v0.17.0

• Fix/add midas backend #341
• fix: get port
• Add component cards, update pipelines #287
• Feat/bloomz7b example #342
• feat: utilize gpt-3.5 in demo dist #343
• Feat/chatgpt working example #344
• fix: n hypotheses and proxies for transformers LMs #346
• fix: no restrictions selector #347
• fix: output by chatgpt #349
• feat: deepy assistant distr #348
• Feat/universal prompting #345

Release v0.16.0

• Feat/openai api lm #302
• Feat/demo distributions full #328
• Feat/unused services #337
• Fix/memory limit #338
• Feat/service ports names #339
• fix: remove midas cls as it is in combined #336

Release v0.15.0

• fix: added CUDA_VISIBLE_DEVICES for entity-linking #329
• Book skill fix #326
• property extraction annotator #202
• Russian Fact Retrieval and Text QA #275
• fixL move el to another GPU #331
• fix: generative skill timeout #330

Release v0.14.2

• update entity detection #309
• fix: utilize new entity-detection #325

Release v0.14.1

• Fix/prompted skill #322
• Feat/debug deepy #321

Release v0.14.0

• Fix/refactor transformers lm #317
• fix: persona extractor tests #318
• Updated torch version for combined to work on A100 #319
• Fix/refactor names #310

Release v0.13.4

• feat: no use of GiB and MiB #306
• Wikidata parsing files #307
• fix: spacy requirements #312
• fix: russian_toxic #315
• faster entity linking and wiki parser #308

Release v0.13.3

• Fix/move env to args (#304)