0.14.0 - Gemma2

Major Changes

• July 2024 binaries by @martindevans, @SignalRT & @m0nsky in #833, #843
• Vulkan support by @m0nsky in #794, #797, #799
• Generic Prompt Formatter by @patrick-hovsepian in #787
• Async implementation of LLamaExecutors by @asmirnov82 in #834
• Include debug symbols with NuGet packages by @dpmm99 in #851
• Allow continuation in Instruct and Interact executors by @dpmm99 in #852

Bug Fixes

• Fixes Typo in IContextParamsExtensions by @SerialKicked in #806
• Fix possible file locks due to race conditions during build & test by @m0nsky in #836
• Fixed CI Out Of Disk Space by @martindevans in #846
• Fix CUDA detection on WSL by @m0nsky in #847

Other Changes

• Remove Beam Search by @martindevans in #783
• Batched Beam Search by @martindevans in #796
• Remove OpenCL Backend. by @martindevans in #782
• BoolQ Benchmark by @martindevans in #802

New Contributors

• @patrick-hovsepian made their first contribution in #787
• @SerialKicked made their first contribution in #806
• @dpmm99 made their first contribution in #851

Full Changelog: v0.13.0...v0.14.0

0.13.0

Major Changes

• June 2024 Binary Update by @martindevans in #751
• Add Rosetta2 Binaries by @abhiaagarwal in #755
• Optimize the native library loading. by @AsakusaRinne in #688
• Add experimental auto-download support. by @AsakusaRinne in #692
• Upload native libraries to huggingface when release new version. by @AsakusaRinne in #749
• Batch Queueing by @martindevans in #761
• BatchedExecutor LLava by @martindevans in #770
• StreamingTextDecoder Special Token Support by @martindevans in #777

Documentation

• Updated READMEs. by @Lamothe in #742
• Add documentation for the change of native library loading. by @AsakusaRinne in #747
• Add OAI demo link by @sangyuxiaowu in #778

Bug Fixes

• KernelMemory bug fix by @zsogitbe in #726
• Fixes for website when running on case-sensitive filesystems. by @Lamothe in #753

Other Changes

• Exposed basic timing information from llama.cpp by @martindevans in #736
• Less Sampler Allocations by @martindevans in #735
• Several updates to web project by @Lamothe in #718
• Conversation Generate All Logits by @martindevans in #743
• Removed .NET 7 now that it's reached EOL. by @Lamothe in #741
• BatchedExecutor Double Buffering by @martindevans in #748
• Minor BatchedExecutor Example Improvements by @martindevans in #750
• LLava API Improvements by @martindevans in #757
• SemanticKernel HomeAutomation Example by @zsogitbe in #769

New Contributors

• @Lamothe made their first contribution in #718
• @sangyuxiaowu made their first contribution in #778
• @abhiaagarwal made their first contribution in #755

Full Changelog: v0.12.0...v0.13.0

0.12.0 - LLama3

Major Changes

• May 2024 Binary Update (Take 2) by @martindevans in #712
• Logging interceptor by @martindevans in #649
• fix: add cuda llava native libraries. by @AsakusaRinne in #647
• Added Whisper.net x LLamaSharp examples for Speech Chat by @Lyrcaxis in #656
• Extension LLava with in memory images by @zsogitbe in #653
• Llava Initial approach to clear images by @SignalRT in #664
• BatchedExecutor Save/Load by @martindevans in #681
• Interruptible Async Model Loading With Progress Monitoring by @martindevans in #702
• LLava Async Loading by @martindevans in #703
• Optional IHistoryTransform added to ChatSession.InitializeSessionFromHistoryAsync by @Norne9 in #711
• Implement context shifting in executor base by @ksanman in #714
• Llama Text Templater by @martindevans in #715
• Make LLamaKvCacheView Safe by @martindevans in #721

Documentation

• Removed ContextSize from most examples by @martindevans in #663
• Add issue templates. by @AsakusaRinne in #700
• Add LLaMA3 chat session example. by @AsakusaRinne in #708
• Improved Example Docs by @martindevans in #728

Bug Fixes

• fix: llava backend ignores avx and cuda. by @AsakusaRinne in #642
• Embeddings correction by @zsogitbe in #674
• Fix context params defaults by @dlyz in #729
• Made special tokens included in prompts tokenize as intended by @Lyrcaxis in #677
• Fixed Minor Issues With Model Loading by @martindevans in #697
• Empty batch check by @martindevans in #691
• Optimize Semantic Kernel LLamaSharpChatCompletion when running with StatefulExecutorBase models by @kidkych in #671
• SemanticKernel: Correcting non-standard way of working with PromptExecutionSettings by @zsogitbe in #689
• Fix cublas build action by @martindevans in #725

Other Changes

• build(deps): bump System.Text.Json from 8.0.2 to 8.0.3 by @dependabot in #636
• build(deps): bump Microsoft.AspNetCore.OpenApi from 8.0.2 to 8.0.3 by @dependabot in #638
• build(deps): bump Microsoft.Extensions.Logging.Abstractions from 8.0.0 to 8.0.1 by @dependabot in #637
• April 2024 Binary Update by @martindevans in #662
• Slightly Safer Quantize Params by @martindevans in #698
• Removed Unnecessary Constructor From Safe Handles by @martindevans in #696
• ci: add workflow to check the spellings. by @AsakusaRinne in #710
• ci: add a workflow to check code format. by @AsakusaRinne in #709
• Remove Conversation.Prompt(String) by @martindevans in #719
• ci: add benchmark test. by @AsakusaRinne in #720

New Contributors

• @Lyrcaxis made their first contribution in #656
• @Norne9 made their first contribution in #711
• @ksanman made their first contribution in #714
• @dlyz made their first contribution in #729
• @eltociear made their first contribution in #730

Full Changelog: 0.11.1...v0.12.0

0.11.1 - LLaVA support

🎏 Major Changes

• LLaVA Support by @SignalRT in #556, #563, #609
• Chat session state management by @eublefar in #560
• Classifier Free Guidance by @martindevans in #536
• March Binary Update by @martindevans in #565
• SetDllImportResolver based loading by @martindevans in #603

📖 Documentation

• The documentation has been improved and no longer outdated, see LLamaSharp Documentation.

🔧 Bug Fixes

• Added conditional compilation code to progress_callback (in LlamaModelParams struct) by @clovisribeiro in #593
• Memory Disposal Tests by @martindevans in #551
• Fixed Publish File paths by @martindevans in #561
• llama_decode lock by @martindevans in #595
• BatchedExecutor Fixed Forking by @martindevans in #621
• Fixed off by one error in LLamaBatch sampling position by @martindevans in #626
• [LLama.KernelMemory] Fixed System.ArgumentException: EmbeddingMode must be true & #617 by @ChengYen-Tang in #615
• fix: the missing of llava_shared library. by @AsakusaRinne in #633

📌 Other Changes

• Removed llama_eval() by @martindevans in #553
• ChatSession: improve exception message by @swharden in #523
• Improve "embeddings" example by @swharden in #525
• Add path to find llama.dll for MAUI by @evolcano in #631
• LLama.Examples: improve model path prompt by @swharden in #526
• NativeLibraryConfig.WithLogs() overload to set log level by @swharden in #529
• LLamaSharp.Examples: Document Q&A with local storage by @swharden in #532
• Used AnsiConsole in a few more places by @martindevans in #534
• ReadOnlySpan<float> in ISamplingPipeline by @martindevans in #538
• KernelMemory update with adding the use of already loaded model by @zsogitbe in #630
• Add Link To Blazor Demo by @alexhiggins732 in #539
• Removed Obsolete SamplingApi by @martindevans in #552
• update readme.md backends by @warquys in #587
• docs: update the example in readme. by @AsakusaRinne in #604
• Update Semantic Kernel & Kernel Memory Package by @xbotter in #612
• BatchedExecutor.Create() method by @martindevans in #613
• LLamaBatch Logit Tracking by @martindevans in #624

🙌 New Contributors

• @swharden made their first contribution in #523
• @alexhiggins732 made their first contribution in #539
• @clovisribeiro made their first contribution in #593
• @warquys made their first contribution in #587
• @eublefar made their first contribution in #560
• @ChengYen-Tang made their first contribution in #615
• @evolcano made their first contribution in #631

Full Changelog: v0.10.0...0.11.0

0.10.0 - Phi2

Major Changes

• Update binaries feb 2024 by @martindevans in #479
• Add CLBLAST native library to native libraries build by @jasoncouture in #468
• Introduced a new BatchedExecutor by @martindevans in #503
• Swapped StatelessExecutor to use llama_decode! by @martindevans in #445
• LLamaToken Struct by @martindevans in #404

Bug Fixes

• KernelMemory EmbeddingMode bug correction by @zsogitbe in #485
• Normalize Embeddings by @martindevans in #507
• StreamingTextDecoder Fix & Tests by @martindevans in #428
• Tokenizer Fixes For Issue 430 by @martindevans in #433

Other Changes

• Use llama instead of libllama in [DllImport] by @jasoncouture in #465
• Updated Examples by @vikramvee in #502
• Added new file types to quantisation by @martindevans in #495
• Smaller Unit Test Model by @martindevans in #496
• Using AddRange in LLamaEmbedder by @martindevans in #499
• Small KV Cache Handling Improvements by @martindevans in #500
• Added increment and decrement operators to LLamaPos by @martindevans in #501
• Swapped GetEmbeddings to llama_decode by @martindevans in #474
• kv_cache_instance_methods by @martindevans in #454
• Removed IModelParams and IContextParams setters. by @martindevans in #472
• Managed LLamaBatch by @martindevans in #442
• Check Model Path Exists by @martindevans in #437
• Model Metadata Loading Cleanup by @martindevans in #438
• Added a check for EOS token in LLamaStatelessExecutor by @martindevans in #434
• Update README.md by @Oceania2018 in #427
• Gpu layer count change by @Kaotic3 in #424
• Improved exceptions in IModelParams for unknown KV override types. by @martindevans in #416

New Contributors

• @Kaotic3 made their first contribution in #424
• @Oceania2018 made their first contribution in #427
• @jasoncouture made their first contribution in #465
• @zsogitbe made their first contribution in #485
• @vikramvee made their first contribution in #502

Full Changelog: 0.9.1...v0.10.0

0.9.1 - Mixtral!

Major Changes

• Rebuilt ChatSession class by @philippjbauer in #344
• Custom Sampling Pipelines by @martindevans in #348
• Updated Binaries December 2023 by @martindevans in #361
• Added LLamaWeights.Metadata property by @martindevans in #380

Bug Fixes

• Fix documentation to reflect changes in ChatSession API by @asmirnov82 in #366
• Added missing field to LLamaModelQuantizeParams by @martindevans in #367
• Fix broken references in docs by @asmirnov82 in #378
• Updated & Fixed WebAPI by @scotmcc in #377
• Fixed loading of very large metadata values by @martindevans in #384
• Update compile.yml to fix not building for windows by @edgett in #386
• Metadata Fixes by @martindevans in #385
• Fix typos in SemanticKernel README file by @asmirnov82 in #408

Other Changes

• Context Set Seed by @martindevans in #368
• Update README.md by @martindevans in #335
• ci: fix error in auto-release. by @AsakusaRinne in #334
• Update README.md by @markvantilburg in #339
• 🔧 Refactor Semantic Kernel chat completion implementation by @xbotter in #341
• build(deps): bump xunit.runner.visualstudio from 2.5.4 to 2.5.5 by @dependabot in #353
• build(deps): bump xunit from 2.6.2 to 2.6.3 by @dependabot in #352
• Added AVX and AVX2 to MacOS x86_64 builds by @martindevans in #360
• Upgrade unittest target framework to .NET 8.0 by @xbotter in #358
• Clone Grammar by @martindevans in #370
• Renamed llama_sample_temperature to llama_sample_temp by @martindevans in #369
• Reset Custom Sampling Pipeline by @martindevans in #372
• Improved support for AVX512 by @martindevans in #373
• bump sk to 1.0.1 & km to 0.18 by @xbotter in #356
• build(deps): bump xunit from 2.6.3 to 2.6.4 by @dependabot in #389
• build(deps): bump xunit.runner.visualstudio from 2.5.5 to 2.5.6 by @dependabot in #391
• build(deps): bump Swashbuckle.AspNetCore from 6.4.0 to 6.5.0 by @dependabot in #388
• build(deps): bump Microsoft.KernelMemory.Abstractions from 0.18.231209.1-preview to 0.24.231228.5 by @dependabot in #397
• build(deps): bump Microsoft.KernelMemory.Core and Microsoft.KernelMemory.Abstractions by @dependabot in #396
• Code cleanup driven by R# suggestions by @martindevans in #400
• Removed some unnecessary uses of unsafe by @martindevans in #401
• Safer Model Handle Creation by @martindevans in #402
• Extra ModelParams Checking by @martindevans in #403

New Contributors

• @markvantilburg made their first contribution in #339
• @asmirnov82 made their first contribution in #366
• @scotmcc made their first contribution in #377
• @edgett made their first contribution in #386

Thank you so much for all the contributions! 😻

Full Changelog: v0.8.1...0.9.1

v0.8.1 - Major BUG fix and better feature detection

Break changes

• Change NativeLibraryConfig.Default to NativeLibraryConfig.Instance.

Major features and fix

• MinP Sampler by @martindevans in #277
• CPU Feature Detection 2 by @martindevans in #281
• AntipromptProcessor access by @saddam213 in #288
• StreamingTextDecoder in LLamaExecutorBase by @martindevans in #293
• November Binary Update by @martindevans in #316
• Update KernelMemory Package by @xbotter in #325
• fix: Chinese encoding error with gb2312. by @AsakusaRinne in #326
• feat: allow customized search path for native library loading. by @AsakusaRinne in #333

Other changes

• Add targets in Web project by @SignalRT in #286
• Update examples by @xbotter in #295
• dotnet8.0 by @martindevans in #292
• progress_callback in LLamaModelParams by @martindevans in #303
• Added Obsolete markings to all Eval overloads by @martindevans in #304
• Improved test coverage. by @martindevans in #311
• Removed Obsolete ModelParams Constructor by @martindevans in #312
• Better TensorSplitsCollection Initialisation by @martindevans in #310
• Add DefaultInferenceParams to Kernel Memory by @xbotter in #307
• Added a converter similar to the Open AI one by @futzy314 in #315

New Contributors

• @futzy314 made their first contribution in #315

Thank you so much for all the contributions!

Full Changelog: v0.8.0...v0.8.1

v0.8.0: performance improvement, cuda feature detection and kernel-memory integration

What's Changed

• fix: binary not copied on MAC platform. by @AsakusaRinne in #238
• docs: add related repos. by @AsakusaRinne in #240
• docs: add example models for v0.7.0. by @AsakusaRinne in #243
• Adapts to SK Kernel Memory by @xbotter in #226
• CodeQL Pointer Arithmetic by @martindevans in #246
• build(deps): bump xunit from 2.5.0 to 2.6.1 by @dependabot in #233
• build(deps): bump xunit.runner.visualstudio from 2.5.0 to 2.5.3 by @dependabot in #234
• build(deps): bump Swashbuckle.AspNetCore from 6.2.3 to 6.5.0 by @dependabot in #235
• build(deps): bump Microsoft.SemanticKernel from 1.0.0-beta1 to 1.0.0-beta4 by @dependabot in #231
• feat(kernel-memory): avoid loading model twice. by @AsakusaRinne in #248
• GitHub Action Pipeline Improvements by @martindevans in #245
• Update README.md by @hswlab in #252
• Removed some CI targets by @martindevans in #253
• Removed Old Targets From CI matrix by @martindevans in #254
• Align with llama.cpp b1488 by @SignalRT in #249
• Enhance framework compatibility by @Uralstech in #259
• Update LLama.Examples using Spectre.Console by @xbotter in #255
• Context Size Autodetect by @martindevans in #263
• Prevent duplication of user prompts / chat history in ChatSession. by @philippjbauer in #266
• build: add package for kernel-memory integration. by @AsakusaRinne in #244
• Exposed YaRN scaling parameters in IContextParams by @martindevans in #257
• Update ToLLamaSharpChatHistory extension method to be public and support semantic-kernel author roles by @kidkych in #274
• Runtime detection MacOS by @SignalRT in #258
• feat: cuda feature detection. by @AsakusaRinne in #275

New Contributors

• @dependabot made their first contribution in #233
• @hswlab made their first contribution in #252
• @Uralstech made their first contribution in #259
• @philippjbauer made their first contribution in #266
• @kidkych made their first contribution in #274

Full Changelog: v0.7.0...v0.8.0

v0.7.0 - improve performance

This release fixes the performance problem in v0.6.0, so that it's strongly recommended to upgraded to this version. Many thanks for the catch of this problem by @lexxsoft and the fix from @martindevans !

What's Changed

• RoundTrip Tokenization Errors by @martindevans in #205
• Fixed Broken Text Decoding by @martindevans in #219
• Multi GPU by @martindevans in #202
• New Binaries & Improved Sampling API by @martindevans in #223

Full Changelog: v0.6.0...v0.7.0

v0.6.0 - follow major llama.cpp changes

What's Changed

• Better Antiprompt Testing by @martindevans in #150
• Simplified LLamaInteractExecutor antiprompt matching by @martindevans in #152
• Changed OpenOrCreate to Create by @martindevans in #153
• Beam Search by @martindevans in #155
• ILogger implementation by @saddam213 in #158
• Removed GenerateResult by @martindevans in #159
• GetState() fix by @martindevans in #160
• llama_get_kv_cache_token_count by @martindevans in #164
• better_instruct_antiprompt_checking by @martindevans in #165
• skip_empty_tokenization by @martindevans in #167
• SemanticKernel API Update by @drasticactions in #169
• Removed unused properties of InferenceParams & ModelParams by @martindevans in #149
• Coding assistent example by @Regenhardt in #172
• Remove non-async by @martindevans in #173
• MacOS default build now is metal llama.cpp #2901 by @SignalRT in #163
• CPU Feature Detection by @martindevans in #65
• make InferenceParams a record so we can use with by @redthing1 in #175
• fix opaque GetState (fixes #176) by @redthing1 in #177
• Extensions Method Unit Tests by @martindevans in #179
• Async Stateless Executor by @martindevans in #182
• Fixed GitHub Action by @martindevans in #190
• GrammarRule Tests by @martindevans in #192
• More Tests by @martindevans in #194
• Support SemanticKernel 1.0.0-beta1 by @DVaughan in #193
• Major llama.cpp API Change by @martindevans in #185
• Cleanup by @martindevans in #196
• Update WebUI inline with v5.0.x by @saddam213 in #197
• More Logging by @martindevans in #198
• chore: Update LLama.Examples and LLama.SemanticKernel by @xbotter in #201
• ci: add auto release workflow. by @AsakusaRinne in #204

New Contributors

• @Regenhardt made their first contribution in #172
• @redthing1 made their first contribution in #175
• @DVaughan made their first contribution in #193

Full Changelog: v0.5.1...v0.6.0