Dify: v0.7.2
Dify 0.7.2. Source highlight: ✨ What’s New in v0.7.2? ✨
1172 tracked entries. Older history is imported progressively; see source coverage.
Dify 0.7.2. Source highlight: ✨ What’s New in v0.7.2? ✨
vLLM v0.5.5. Source highlight: We introduced a new mode that schedule multiple GPU steps in advance, reducing CPU overhead (#7000, #7387, #7452, #7703). Initial result shows 20% improvements in QPS for a single GPU running 8B and 30B models. You can set `--num-scheduler-steps 8` as a parameter to the API server (via `vllm serve`)
LocalAI v2.20.1. Source highlight: It's that time again—I’m excited (and honestly, a bit proud) to announce the release of **LocalAI v2.20**! This one’s a biggie, with some of the most requested features and enhancements, all designed to make your self-hosted AI journey even smoother and more powerful.
LocalAI v2.20.0. Source highlight: 🌍 **Explorer & Community:** Explore global community pools at explorer.localai.io
Hugging Face Transformers v4.44.2. Source highlight: Patch release v4.44.2, mostly 2 regressions that were not caught for Jamba and for processors!
Open WebUI v0.3.15. Source highlight: [0.3.15] - 2024-08-21
Open WebUI v0.3.14. Source highlight: [0.3.14] - 2024-08-21
Hugging Face Transformers v4.44.1. Source highlight: Here are the different fixes, mostly Gemma2 context length, nits here and there, and generation issues
Ollama v0.3.7. Source highlight: Hermes 3: Hermes 3 is the latest version of the flagship Hermes series of LLMs by Nous Research, which includes support for tool calling.
Dify 0.7.1. Source highlight: ✨ What’s New in v0.7.1? ✨
Open WebUI v0.3.13. Source highlight: [0.3.13] - 2024-08-14
Dify 0.7.0. Source highlight: ✨ What’s New in v0.7.0? ✨
Ollama v0.3.6. Source highlight: Fixed issue where `/api/embed` would return an error instead of loading the model when the `input` field was not provided.
Ollama v0.3.5. Source highlight: Fixed `Incorrect function` error when downloading models on Windows
Open WebUI v0.3.12. Source highlight: [0.3.12] - 2024-08-07
Hugging Face Transformers v4.44.0. Source highlight: Release v4.44.0: End to end compile generation!!! Gemma2 (with assisted decoding), Codestral (Mistral for code), Nemotron, Efficient SFT training, CPU Offloaded KVCache, torch export for static cache
Ollama v0.3.4. Source highlight: New embedding models
Dify 0.6.16. Source highlight: ✨ What’s New in v0.6.16? ✨
vLLM v0.5.4. Source highlight: Enhanced pipeline parallelism support for DeepSeek v2 (#6519), Qwen (#6974), Qwen2 (#6924), and Nemotron (#6863)
Hugging Face Transformers v4.43.4. Source highlight: Patch Release v4.43.4
Open WebUI v0.3.11. Source highlight: [0.3.11] - 2024-08-02
Ollama v0.3.3. Source highlight: The `/api/embed` endpoint now returns statistics: `total_duration`, `load_duration`, and `prompt_eval_count`
LocalAI v2.19.4. Source highlight: chore(model-gallery): :arrow_up: update checksum
Ollama v0.3.2. Source highlight: Fixed issue where `ollama pull` would not resume download progress
Ollama v0.3.1. Source highlight: Google's Gemma 2 has a new 2B parameter small model!
LocalAI v2.19.3. Source highlight: fix(gallery): do not attempt to delete duplicate files
Hugging Face Transformers v4.43.3. Source highlight: Patch release v4.43.3:
Ollama v0.3.0. Source highlight: Ollama now supports tool calling with popular models such as Llama 3.1. This enables a model to answer a given prompt using tool(s) it knows about, making it possible for models to perform more complex tasks or interact with the outside world.
LocalAI v2.19.2. Source highlight: This release is a patch release to fix well known issues from 2.19.x
Hugging Face Transformers v4.43.2. Source highlight: Fix float8_e4m3fn in modeling_utils (#32193)
vLLM v0.5.3.post1. Source highlight: We fixed an configuration incompatibility between vLLM (which tested against pre-released version) and the published Meta Llama 3.1 weights (#6693)
Hugging Face Transformers v4.43.1 is published, but the available notes are too limited to assess upgrade impact. Review the official source before upgrading.
Hugging Face Transformers v4.43.0. Source highlight: The Llama 3.1 models are released by Meta and come in three flavours: 8B, 70B, and 405B.
Dify 0.6.15. Source highlight: ✨ What’s New in v0.6.15? ✨
vLLM v0.5.3. Source highlight: vLLM now supports Meta Llama 3.1! Please checkout our blog here for initial details on running the model.
Ollama v0.2.8. Source highlight: Mistral Nemo: A state-of-the-art 12B model with 128k context length, built by Mistral AI in collaboration with NVIDIA.
LocalAI v2.19.1. Source highlight: LocalAI 2.19.1 is out! :mega:
LocalAI v2.19.0. Source highlight: LocalAI 2.19.0 is out! :mega:
Ollama v0.2.7. Source highlight: Fixed issue where last message when streaming would omit the `content` response field
Ollama v0.2.6. Source highlight: Mathstral: MathΣtral is a 7B model designed for math reasoning and scientific discovery by Mistral AI.
Open WebUI v0.3.10. Source highlight: [0.3.10] - 2024-07-17
Open WebUI v0.3.9. Source highlight: [0.3.9] - 2024-07-17
vLLM v0.5.2. Source highlight: ❗Planned breaking change ❗: we plan to remove beam search (see more in #6226) in the next few releases. This release come with a warning when beam search is enabled for the request. Please voice your concern in the RFC if you do have a valid use case for beam search in vLLM
Dify 0.6.14. Source highlight: ✨ What’s New in v0.6.14? ✨
Ollama v0.2.5. Source highlight: Fixed issue where a model's `SYSTEM` message not be applied
Ollama v0.2.4. Source highlight: Fixed issue where `context`, `load_duration` and `total_duration` fields would not be set in the `/api/generate` endpoint.
Ollama v0.2.3. Source highlight: Fix issue where system prompt would not be applied
Hugging Face Transformers v4.42.4. Source highlight: Mostly gemma2 support FA2 softcapping!
Ollama v0.2.2. Source highlight: Fixed errors that occurred when using Ollama with Nvidia V100 GPUs
Open WebUI v0.3.8. Source highlight: [0.3.8] - 2024-07-09