Notice: Upcoming Default Server Port Change The latest llama.cpp release (build b10240) introduces a notice about a planned change to the default server port. Currently set to 8080, the port will move to 9931 in a future update. Self-hosted users who rely on the default configuration should be aware that scripts, clients, and firewall rules […]
llama.cpp Update: Upcoming Server Port Change and New Builds (b10240) Read More »
The latest release of llama.cpp, build b10227, introduces a specialized parser for the Qwen3 model family, enhancing how the chat interface handles tagged thinking tools. This update refactors internal chat processing and adds a permute helper, streamlining the codebase for future expansions. Self-hosted users will benefit from more accurate tool calling, as the update adds
llama.cpp Build b10227 Brings Qwen3 Parser and Improved Tool Calling Support Read More »
The latest llama.cpp release, build b10217, enables tool calling during the thinking phase for DeepSeek (DS4) models in chat mode. This enhancement allows the model to seamlessly invoke external tools while it is still reasoning, producing more coherent and effective multi-step interactions. For self-hosted users running DeepSeek, this means smarter assistant behavior and improved reliability
Llama.cpp Build b10217 Introduces Tool Calling in Thinking Mode for DeepSeek (DS4) Read More »
The langchain-core library has been updated to version 1.5.3, delivering a small but crucial fix for self-hosted users. A previous issue could cause gateway authentication failures when an API key wasn’t explicitly configured in certain environments. This release patches that behavior by automatically falling back to the LANGSMITH_API_KEY environment variable, ensuring seamless connectivity and reducing
LangChain Core 1.5.3 Fixes Gateway Authentication Fallback Read More »
New Feature: SYCL q2_0 Mul_Mat The latest build of llama.cpp, version b10203, introduces support for the q2_0 quantization scheme in the matrix multiplication (mul_mat) operation when using the SYCL backend. This enables efficient inference with 2-bit quantized models on Intel GPUs and other SYCL-compatible hardware, reducing memory footprint and potentially boosting throughput. Both the basic
llama.cpp Build b10203 Adds SYCL q2_0 Mul_Mat Support and Wide Platform Binaries Read More »
The latest build of llama.cpp brings a notable refinement to its computation graph. A custom CPU operation that was previously hard-coded for the M3 graph has been eliminated, and its functionality is now expressed using standard, stock operations. This change simplifies the codebase, making it more maintainable and potentially improving compatibility across different hardware backends.
llama.cpp Release Cleans Up M3 Graph by Replacing Custom CPU Op with Standard Operations Read More »
LangChain Core version 1.5.2 is now available, bringing a targeted bug fix and several routine dependency updates. The most notable change resolves an issue where empty string values in gateway environment variables were not handled correctly, which could lead to unexpected behavior in certain configurations. This fix ensures more robust handling of environment settings for
LangChain Core v1.5.2 Released with Bug Fix and Dependency Updates Read More »
The latest llama.cpp build, b10176, introduces a valuable new feature for distributed and remote inference scenarios: the tensor_memset remote procedure call. This addition allows for efficient initialization of tensors across RPC connections, which is particularly beneficial for self-hosted setups that distribute model layers across multiple machines. By streamlining how memory is set up remotely, it
llama.cpp Build b10176 Brings Tensor Memset RPC and Expanded Platform Support Read More »
The langchain-fireworks package has been bumped to version 1.5.2, bringing a series of behind-the-scenes improvements that matter for anyone running their own AI stack. This release is all about keeping model profile data accurate and current. For self-hosted users, up-to-date profiles mean your integration can correctly leverage the latest capabilities and pricing of models from
LangChain-Fireworks 1.5.2: Keeping Model Profiles Fresh for Self-Hosted Setups Read More »
We’re excited to announce a new build of llama.cpp (b10159), which includes a performance-boosting addition for Apple Silicon users. This release introduces the Fast Walsh-Hadamard Transform (FWHT) kernel into the Metal backend, improving the efficiency of certain mathematical operations during model inference. While the change is low-level, it helps streamline computation on macOS and iOS
llama.cpp Release b10159 Brings FWHT Kernel to Metal Backend Read More »
