Snapshot 22-01-2024

What's Changed

• IPEX fix diffusers 0.21 lazy load by @Disty0 in #465
• Updated Kobold Lite to v63 by @LostRuins in #466
• GPTQ support for IPEX by @Disty0 in #468
• updated kobold lite to v66 by @LostRuins in #470
• [gptq_hf_torch] Fix typo in model type check by @nkpz in #471
• updated lite to v70 by @LostRuins in #472
• IPEX Optimizations by @Disty0 in #477
• Update Kobold Lite to v76 by @LostRuins in #478
• Typo. by @FunkEngine2023 in #479
• Abort code moved to model classes by @ebolam in #473
• Add epsilon and eta sampling by @StefanGliga in #475
• Updated Kobold Lite to v83 by @LostRuins in #480
• Image gen Enhancements by @ebolam in #476
• Update IPEX engine by @Disty0 in #487
• IPEX Torch 2.1 support and upgrade to Python 3.10 by @Disty0 in #488
• Add BigDL LLM backend by @Disty0 in #491
• Update BigDL and add commandline-ipex.sh by @Disty0 in #492
• Update Kobold Lite to v101 by @LostRuins in #496
• Disable model load by @ebolam in #499
• fix(load_json): use int action id by @nkpz in #483
• TTS update and double click to load story by @ebolam in #474

New Contributors

• @FunkEngine2023 made their first contribution in #479
• @StefanGliga made their first contribution in #475

Full Changelog: Snapshot-11-09-23...Snapshot-23-01-24

Snapshot 11-09-23

What's Changed

• Fix most of the API by @one-some in #411
• More api fixes by @one-some in #438
• switch to AI Horde Worker Scribe by @db0 in #441
• fix: Avoid conflicting args by @db0 in #445
• Create api_example.py by @YellowRoseCx in #448
• Updated Kobold Lite to v60 by @LostRuins in #449
• Initial Intel ARC support with IPEX by @Disty0 in #450
• Fix for mobile touch on setting sliders by @ebolam in #451
• Fix for options not showing up on refresh by @ebolam in #454
• Add new models by @AlpinDale in #453
• IPEX fixes by @Disty0 in #457
• UI2 visual fixes by @ebolam in #458
• update lite to v62 (instruct placeholder handling change) by @LostRuins in #460
• IPEX fix SDPA and linalg solve by @Disty0 in #461
• Added password transmit block when on --host mode by @ebolam in #459
• Merge exllama backend into united. by @pi6am in #447

New Contributors

• @Disty0 made their first contribution in #450
• @AlpinDale made their first contribution in #453

Full Changelog: Snapshot-11-08-23...Snapshot-11-09-23

Snapshot 11-08-2023

This latest snapshot features better AutoGPTQ support, the AutoGPTQ version in the installer is the official release without exllama kernels.
Attached is an experimental build compiled by us that does include them for those that want to test them, if this test is successful we will begin automatically distributing this file trough the updater (If unsure don't install it and just keep using our own updates).

What's Changed

• Configurable quantization level, fix for broken toggles in model settings by @nkpz in #409
• Fix WI comment editing by @one-some in #413
• updated lite to v50 by @LostRuins in #417
• Fixes forever by @one-some in #415
• Add AdventureMode stopper by @YellowRoseCx in #334
• GPTQ module by @0cc4m in #367
• index prioritization by @one-some in #418
• Rework single-gen streaming by @one-some in #419
• Fix prioritization (probably) by @one-some in #422
• gptq is soooo lazy by @one-some in #424
• Fix streaming bugs by @one-some in #423
• Move pickle stuff into modeling/pickling.py by @one-some in #431
• Workaround for option panel popping up too much by @one-some in #428
• Submit context menu by @one-some in #414
• UI: Fix the thingey modal by @one-some in #432
• (mostly) wi fixes and polish by @one-some in #416
• UI: Replace shift_down code with builtin event.shiftKey by @one-some in #426
• updated lite to v54 by @LostRuins in #433
• UI: Fix token streaming gobbling trailing whitespace by @one-some in #435
• UI: Change mobile aspect ratio threshold from 7/5 to 5/6 by @one-some in #434

New Contributors

• @0cc4m made their first contribution in #367

Full Changelog: Snapshot-18-07-23...Snapshot-11-08-23

Snapshot 18-07-23

This is a development snapshot of KoboldAI United meant for Windows users using the full offline installer.

Since the last snapshot the backend has been overhauled completely and now supports BitsandBytes 4-bit.
Note: In this release you can no longer use --trust_remote_code and the command line options for model loading have changed.

All the hard work since last release

• fix: tpu tokenizers errors by @Zurnaz in #353
• Fix TPU API errors by @one-some in #357
• GPT-J fix by @one-some in #359
• Fix memory->genre bug in context viewer bar tooltip by @one-some in #358
• Updated embedded Kobold Lite to v32 by @LostRuins in #363
• Implement modular model backends Phase 1 by @ebolam in #362
• Anti-malware, nobreakmodel, and css fix by @ebolam in #366
• Reverting the RTE by @LightSaveUs in #368
• Fix for --nobreakmodel forcing CPU by @ebolam in #369
• fix: tpu support models saved in bfloat16 by @Zurnaz in #370
• updated kobold lite to v37 by @LostRuins in #371
• Making model backends respond to a specific type in the aiserver menu for now by @ebolam in #373
• Updated Kobold Lite to v41 by @LostRuins in #376
• Fix for model backends that use toggles always returning true by @ebolam in #377
• Fix WI for UI1 by @jojorne in #378
• Fix move and delete WI item for UI1 by @jojorne in #382
• Accelerate offloading by @one-some in #385
• Use VE's patched load_from_state_dict on TPU for loading empty weights by @one-some in #386
• Updated Kobold Lite to v46 by @LostRuins in #387
• Stub seek_offset for cache sorting in load loop by @one-some in #389
• Fix bleeding edge model loads and add lazyload fallback by @one-some in #390
• Fix for UI2 model loading not showing progress by @ebolam in #392
• Basic HF backend by @one-some in #391
• Fixes for model backend UI by @ebolam in #397
• Speeds up bnb 4bit with a custom BitsAndBytesConfig by @Alephrin in #398
• Update to the upload_file function by @ebolam in #399
• missed the elif by @ebolam in #400
• Patches: Make lazyload work with quantization by @one-some in #402
• Delete basic 4bit by @one-some in #404
• Save the 4-bit flag to the model settings. by @ebolam in #401
• Clarified message on what's required for model backend parameters by @ebolam in #406

New Contributors

• @Alephrin made their first contribution in #398

Full Changelog: Snapshot-7-5-2023...Snapshot-18-07-23

Snapshot 7-5-2023

This is a development snapshot of KoboldAI United meant for Windows users using the full offline installer.

KoboldRT-BNB.zip is included for historical reasons but should no longer be used by anyone, KoboldAI will automatically download and install a newer version when you run the updater. It is also already included in the offline installer.