Ollama v0.1.21 — Release Notes & Engineering Impact
Ollama v0.1.21. Source highlight: Qwen: Qwen is a series of large language models by Alibaba Cloud spanning from 1.8B to 72B parameters.
AI Models & LLM Frameworks archive — Page 19
1172 tracked entries. Older history is imported progressively; see source coverage.
Dify: v0.4.9
Dify 0.4.9. Source highlight: As per the planned schedule, these deprecated envs are being removed:
Dify: v0.4.8
Dify 0.4.8. Source highlight: Jina AI `jina-embeddings-v2-base-zh` model supported.
Dify: v0.4.7
Dify 0.4.7. Source highlight: Deploying with Docker Compose:
Dify: v0.4.6
Dify 0.4.6. Source highlight: Support for Models Integration Provided by the `Ollama`, details: https://docs.dify.ai/advanced/model-configuration/ollama
Ollama v0.1.20 — Release Notes & Engineering Impact
Ollama v0.1.20. Source highlight: MegaDolphin: A new 120B version of the Dolphin model.
Dify: v0.4.5
Dify 0.4.5. Source highlight: correct xorbits spelling
LocalAI v2.5.1 — Release Notes & Engineering Impact
LocalAI v2.5.1. Source highlight: Patch release to create `/build/models` in the container images.
Ollama v0.1.19 — Release Notes & Engineering Impact
Ollama v0.1.19. Source highlight: This release focuses on performance and fixing a number issues and crashes relating to memory allocation.
LocalAI v2.5.0 — Release Notes & Engineering Impact
LocalAI v2.5.0. Source highlight: This release adds more embedded models, and shrink image sizes.
LocalAI v2.4.1 — Release Notes & Engineering Impact
LocalAI v2.4.1. Source highlight: Exciting New Features 🎉
Dify: v0.4.4
Dify 0.4.4. Source highlight: Add Together.ai model provider and fix some bugs.
LocalAI v2.4.0 — Release Notes & Engineering Impact
LocalAI v2.4.0. Source highlight: fix(entrypoint): cd to backend dir before start
Dify: v0.4.3
Dify 0.4.3. Source highlight: Optimize performance & fix few bugs.
vLLM v0.2.7 — Release Notes & Engineering Impact
vLLM v0.2.7. Source highlight: Up to 70% throughput improvement for distributed inference by removing serialization/deserialization overheads
Ollama v0.1.18 — Release Notes & Engineering Impact
Ollama v0.1.18. Source highlight: TinyLlama: a compact 1.1B Llama model on 3 trillion tokens
Dify: v0.4.2
Dify 0.4.2. Source highlight: fix: ciphertext error
Dify: v0.4.1
Dify 0.4.1. Source highlight: fix: agent strategy missing in app model config
Dify: v0.4.0
Dify 0.4.0. Source highlight: **🎉🎉 Dify's Version 0.4 is out now.**
LocalAI v2.3.1 — Release Notes & Engineering Impact
LocalAI v2.3.1. Source highlight: Fix: Set proper Homebrew install location for x86 Macs
LocalAI v2.3.0 — Release Notes & Engineering Impact
LocalAI v2.3.0. Source highlight: fix: exllama2 backend
LocalAI v2.2.0 — Release Notes & Engineering Impact
LocalAI v2.2.0. Source highlight: This release brings updates to the backends and includes a fix for recompilation of LocalAI with go-rwkv (https://github.com/mudler/LocalAI/issues/1473). To note, it also tries to reduce the image size by allowing some backends (transformers-based) to share the same environment.
Dify: v0.3.34
Dify 0.3.34. Source highlight: Annotation Reply, see details: Link
Ollama v0.1.17 — Release Notes & Engineering Impact
Ollama v0.1.17. Source highlight: This release adds support for the Phi-2 model by Microsoft.
Hugging Face Transformers: Patch release: v4.36.2
Hugging Face Transformers v4.36.2. Source highlight: Patch release to resolve some critical issues relating to the recent cache refactor, flash attention refactor and training in the multi-gpu and multi-node settings:
vLLM v0.2.6 — Release Notes & Engineering Impact
vLLM v0.2.6. Source highlight: Fast model execution with CUDA/HIP graph
LocalAI v2.1.0 — Release Notes & Engineering Impact
LocalAI v2.1.0. Source highlight: feat(alias): alias llama to llama-cpp, update docs
Ollama v0.1.16 — Release Notes & Engineering Impact
Ollama v0.1.16. Source highlight: This release adds support Mixtral and other models based on its Mixture of Experts (MoE) architecture. Note: this model requires at least 48GB of memory.
vLLM v0.2.5 — Release Notes & Engineering Impact
vLLM v0.2.5. Source highlight: Optimize Mixtral performance with expert parallelism (thanks to @Yard1)
Hugging Face Transformers: Patch release: v4.36.1
Hugging Face Transformers v4.36.1. Source highlight: A patch release for critical torch issues mostly:
Ollama v0.1.15 — Release Notes & Engineering Impact
Ollama v0.1.15. Source highlight: Multimodal model support
vLLM v0.2.4 — Release Notes & Engineering Impact
vLLM v0.2.4. Source highlight: Mixtral model support (officially from @mistralai)
Hugging Face Transformers: v4.36: Mixtral, Llava/BakLlava, SeamlessM4T v2, AMD ROCm, F.sdpa wide-spread support
Hugging Face Transformers v4.36.0. Source highlight: Mixtral is the new open-source model from Mistral AI announced by the blogpost Mixtral of Experts. The model has been proven to have comparable capabilities to Chat-GPT according to the benchmark results shared on the release blogpost.
Ollama v0.1.14 — Release Notes & Engineering Impact
Ollama v0.1.14. Source highlight: StableLM Zephyr: A lightweight chat model allowing accurate, and responsive output without requiring high-end hardware.
Dify: v0.3.33
Dify 0.3.33. Source highlight: feat: add jina embedding
LocalAI v2.0.0 — Release Notes & Engineering Impact
LocalAI v2.0.0. Source highlight: :fire: add LLaVA support and GPT vision API, Multiple requests for llama.cpp, return JSON types
vLLM v0.2.3 — Release Notes & Engineering Impact
vLLM v0.2.3. Source highlight: Refactoring on Worker, InputMetadata, and Attention
Ollama v0.1.13 — Release Notes & Engineering Impact
Ollama v0.1.13. Source highlight: Starling: a large language model trained by reinforcement learning from AI feedback focused on improving chatbot helpfulness.
Ollama v0.1.12 — Release Notes & Engineering Impact
Ollama v0.1.12. Source highlight: Yi Chat: the chat variant of the popular Yi 34b model is now available.
Dify: v0.3.32
Dify 0.3.32. Source highlight: Supported the `rerank` model of `xinference` for local deployment, such as: `bge-reranker-large`, `bge-reranker-base`.
Dify: v0.3.31-fix3
Dify 0.3.31-fix3. Source highlight: Fix: Hybrid search recall fails when the chosen dataset contains no valid documents.
Dify: v0.3.31-fix2
Dify 0.3.31-fix2. Source highlight: add anthropic `claude-2.1` support
Dify: v0.3.31-fix1
Dify 0.3.31-fix1. Source highlight: Releasing a new version due to some issues that made the previous one unavailable.
Dify: v0.3.31
Dify 0.3.31. Source highlight: Added support for multi-datasets retrieval, hybrid search, and result re-ranking.
Ollama v0.1.11 — Release Notes & Engineering Impact
Ollama v0.1.11. Source highlight: Orca 2: A fine-tuned version of Meta's Llama 2 model, designed to excel particularly in reasoning.
vLLM v0.2.2 — Release Notes & Engineering Impact
vLLM v0.2.2. Source highlight: Bump up to PyTorch v2.1 + CUDA 12.1 (vLLM+CUDA 11.8 is also provided)
Ollama v0.1.10 — Release Notes & Engineering Impact
Ollama v0.1.10. Source highlight: OpenChat: An open-source chat model trained on a wide variety of data, surpassing ChatGPT on various benchmarks.
Hugging Face Transformers: Patch release: v4.35.2
Hugging Face Transformers v4.35.2. Source highlight: A patch release was made for the following commit:
Hugging Face Transformers: Patch release: v4.35.1
Hugging Face Transformers v4.35.1. Source highlight: A patch release was made for the following three commits:
Dify: v0.3.30
Dify 0.3.30. Source highlight: App supports image uploading and calls the OpenAI GPT-4V or `gpt-4-vision-preview` which allows the model to take in images and answer questions about them.