Merge branch 'master' into disable_grammar_by_default

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

.github/bump_deps.sh

@@ -6,17 +6,4 @@ VAR=$3
 
 LAST_COMMIT=$(curl -s -H "Accept: application/vnd.github.VERSION.sha" "https://api.github.com/repos/$REPO/commits/$BRANCH")
 
-# Read $VAR from Makefile (only first match)
-set +e
-CURRENT_COMMIT="$(grep -m1 "^$VAR?=" Makefile | cut -d'=' -f2)"
-set -e
-
 sed -i Makefile -e "s/$VAR?=.*/$VAR?=$LAST_COMMIT/"
-
-if [ -z "$CURRENT_COMMIT" ]; then
-    echo "Could not find $VAR in Makefile."
-    exit 0
-fi
-
-echo "Updated $VAR from $CURRENT_COMMIT to $LAST_COMMIT." > "$REPO_message.txt"
-echo "https://github.com/$REPO/compare/$CURRENT_COMMIT..$LAST_COMMIT" >> "$REPO_message.txt"

.github/workflows/bump_deps.yaml

@@ -40,14 +40,8 @@ jobs:
     steps:
       - uses: actions/checkout@v4
       - name: Bump dependencies 🔧
-        id: bump
         run: |
           bash .github/bump_deps.sh ${{ matrix.repository }} ${{ matrix.branch }} ${{ matrix.variable }}
-          {
-            echo 'message<<EOF'
-            cat "${{ matrix.repository }}_message.txt"
-            echo EOF
-          } >> "$GITHUB_OUTPUT"
       - name: Create Pull Request
         uses: peter-evans/create-pull-request@v6
         with:
@@ -56,7 +50,7 @@ jobs:
           commit-message: ':arrow_up: Update ${{ matrix.repository }}'
           title: 'chore: :arrow_up: Update ${{ matrix.repository }}'
           branch: "update/${{ matrix.variable }}"
-          body:  ${{ steps.bump.outputs.message }}
+          body: Bump of ${{ matrix.repository }} version
           signoff: true
 
 

.github/workflows/checksum_checker.yaml

@@ -41,7 +41,7 @@ jobs:
           token: ${{ secrets.UPDATE_BOT_TOKEN }}
           push-to-fork: ci-forks/LocalAI
           commit-message: ':arrow_up: Checksum updates in gallery/index.yaml'
-          title: 'chore(model-gallery): :arrow_up: update checksum'
+          title: 'models(gallery): :arrow_up: update checksum'
           branch: "update/checksum"
           body: Updating checksums in gallery/index.yaml
           signoff: true

.github/workflows/image-pr.yml

@@ -47,7 +47,7 @@ jobs:
           #   makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
-            cuda-minor-version: "0"
+            cuda-minor-version: "4"
             platforms: 'linux/amd64'
             tag-latest: 'false'
             tag-suffix: '-cublas-cuda12-ffmpeg'
@@ -120,7 +120,7 @@ jobs:
           #   makeflags: "--jobs=3 --output-sync=target"
           # - build-type: 'cublas'
           #   cuda-major-version: "12"
-          #   cuda-minor-version: "0"
+          #   cuda-minor-version: "4"
           #   platforms: 'linux/amd64'
           #   tag-latest: 'false'
           #   tag-suffix: '-cublas-cuda12-ffmpeg-core'

.github/workflows/image.yml

@@ -75,7 +75,7 @@ jobs:
             makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
-            cuda-minor-version: "0"
+            cuda-minor-version: "4"
             platforms: 'linux/amd64'
             tag-latest: 'false'
             tag-suffix: '-cublas-cuda12'
@@ -100,7 +100,7 @@ jobs:
             makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
-            cuda-minor-version: "0"
+            cuda-minor-version: "4"
             platforms: 'linux/amd64'
             tag-latest: 'auto'
             tag-suffix: '-cublas-cuda12-ffmpeg'
@@ -285,7 +285,7 @@ jobs:
             makeflags: "--jobs=4 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
-            cuda-minor-version: "0"
+            cuda-minor-version: "4"
             platforms: 'linux/amd64'
             tag-latest: 'false'
             tag-suffix: '-cublas-cuda12-core'
@@ -307,7 +307,7 @@ jobs:
             makeflags: "--jobs=4 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
-            cuda-minor-version: "0"
+            cuda-minor-version: "4"
             platforms: 'linux/amd64'
             tag-latest: 'false'
             tag-suffix: '-cublas-cuda12-ffmpeg-core'

.github/workflows/release.yaml

@@ -4,8 +4,6 @@ on:
   push:
     branches:
       - master
-    tags:
-      - 'v*'
   pull_request:
 
 env:
@@ -31,10 +29,11 @@ jobs:
         with:
           go-version: '1.21.x'
           cache: false
+
       - name: Dependencies
         run: |
           sudo apt-get update
-          sudo apt-get install build-essential ffmpeg protobuf-compiler ccache upx-ucl gawk
+          sudo apt-get install build-essential ffmpeg protobuf-compiler ccache gawk
           sudo apt-get install -qy binutils-aarch64-linux-gnu gcc-aarch64-linux-gnu g++-aarch64-linux-gnu libgmock-dev
       - name: Install CUDA Dependencies
         run: |
@@ -150,7 +149,7 @@ jobs:
       - name: Dependencies
         run: |
           sudo apt-get update
-          sudo apt-get install -y wget curl build-essential ffmpeg protobuf-compiler ccache upx-ucl gawk cmake libgmock-dev
+          sudo apt-get install -y wget curl build-essential ffmpeg protobuf-compiler ccache gawk cmake libgmock-dev
       - name: Intel Dependencies
         run: |
           wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB | gpg --dearmor | sudo tee /usr/share/keyrings/oneapi-archive-keyring.gpg > /dev/null
@@ -251,7 +250,7 @@ jobs:
       - name: Dependencies
         run: |
           sudo apt-get update
-          sudo apt-get install -y --no-install-recommends libopencv-dev protobuf-compiler ccache upx-ucl
+          sudo apt-get install -y --no-install-recommends libopencv-dev protobuf-compiler ccache
           go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@1958fcbe2ca8bd93af633f11e97d44e567e945af
           go install google.golang.org/protobuf/cmd/protoc-gen-go@v1.34.2
       - name: Build stablediffusion

.github/workflows/test.yml

@@ -70,7 +70,7 @@ jobs:
       - name: Dependencies
         run: |
           sudo apt-get update
-          sudo apt-get install build-essential ccache upx-ucl curl ffmpeg
+          sudo apt-get install build-essential curl ffmpeg
           sudo apt-get install -y libgmock-dev
           curl https://repo.anaconda.com/pkgs/misc/gpgkeys/anaconda.asc | gpg --dearmor > conda.gpg && \
              sudo install -o root -g root -m 644 conda.gpg /usr/share/keyrings/conda-archive-keyring.gpg && \

Dockerfile

@@ -24,7 +24,7 @@ RUN apt-get update && \
         cmake \
         curl \
         git \
-        unzip upx-ucl && \
+        unzip && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
@@ -99,7 +99,7 @@ FROM requirements-${IMAGE_TYPE} AS requirements-drivers
 
 ARG BUILD_TYPE
 ARG CUDA_MAJOR_VERSION=12
-ARG CUDA_MINOR_VERSION=0
+ARG CUDA_MINOR_VERSION=4
 
 ENV BUILD_TYPE=${BUILD_TYPE}
 

Makefile

@@ -8,7 +8,7 @@ DETECT_LIBS?=true
 # llama.cpp versions
 GOLLAMA_REPO?=https://github.com/go-skynet/go-llama.cpp
 GOLLAMA_VERSION?=2b57a8ae43e4699d3dc5d1496a1ccd42922993be
-CPPLLAMA_VERSION?=0d6fb52be0c1b7e77eb855f3adc4952771c8ce4c
+CPPLLAMA_VERSION?=705b7ecf60e667ced57c15d67aa86865e3cc7aa7
 
 # gpt4all version
 GPT4ALL_REPO?=https://github.com/nomic-ai/gpt4all
@@ -20,7 +20,7 @@ RWKV_VERSION?=661e7ae26d442f5cfebd2a0881b44e8c55949ec6
 
 # whisper.cpp version
 WHISPER_REPO?=https://github.com/ggerganov/whisper.cpp
-WHISPER_CPP_VERSION?=fe36c909715e6751277ddb020e7892c7670b61d4
+WHISPER_CPP_VERSION?=f68298ce06ca3edd6e6f3f21c3d0bb5f073942c3
 
 # bert.cpp version
 BERT_REPO?=https://github.com/go-skynet/go-bert.cpp
@@ -58,7 +58,7 @@ RANDOM := $(shell bash -c 'echo $$RANDOM')
 
 VERSION?=$(shell git describe --always --tags || echo "dev" )
 # go tool nm ./local-ai | grep Commit
-LD_FLAGS?=-s -w
+LD_FLAGS?=
 override LD_FLAGS += -X "github.com/mudler/LocalAI/internal.Version=$(VERSION)"
 override LD_FLAGS += -X "github.com/mudler/LocalAI/internal.Commit=$(shell git rev-parse HEAD)"
 
@@ -72,14 +72,6 @@ WHITE  := $(shell tput -Txterm setaf 7)
 CYAN   := $(shell tput -Txterm setaf 6)
 RESET  := $(shell tput -Txterm sgr0)
 
-UPX?=
-# check if upx exists
-ifeq (, $(shell which upx))
-	UPX=
-else
-	UPX=$(shell which upx)
-endif
-
 # Default Docker bridge IP
 E2E_BRIDGE_IP?=172.17.0.1
 
@@ -385,7 +377,6 @@ build: prepare backend-assets grpcs ## Build the project
 	$(info ${GREEN}I BUILD_TYPE: ${YELLOW}$(BUILD_TYPE)${RESET})
 	$(info ${GREEN}I GO_TAGS: ${YELLOW}$(GO_TAGS)${RESET})
 	$(info ${GREEN}I LD_FLAGS: ${YELLOW}$(LD_FLAGS)${RESET})
-	$(info ${GREEN}I UPX: ${YELLOW}$(UPX)${RESET})
 ifneq ($(BACKEND_LIBS),)
 	$(MAKE) backend-assets/lib
 	cp -f $(BACKEND_LIBS) backend-assets/lib/
@@ -430,7 +421,7 @@ else
 endif
 
 dist-cross-linux-arm64:
-	CMAKE_ARGS="$(CMAKE_ARGS) -DGGML_NATIVE=off" GRPC_BACKENDS="backend-assets/grpc/llama-cpp-fallback backend-assets/grpc/llama-cpp-grpc backend-assets/util/llama-cpp-rpc-server" GO_TAGS="p2p" \
+	CMAKE_ARGS="$(CMAKE_ARGS) -DGGML_NATIVE=off" GRPC_BACKENDS="backend-assets/grpc/llama-cpp-fallback backend-assets/grpc/llama-cpp-grpc backend-assets/util/llama-cpp-rpc-server" \
 	STATIC=true $(MAKE) build
 	mkdir -p release
 # if BUILD_ID is empty, then we don't append it to the binary name
@@ -480,7 +471,7 @@ prepare-e2e:
 	mkdir -p $(TEST_DIR)
 	cp -rfv $(abspath ./tests/e2e-fixtures)/gpu.yaml $(TEST_DIR)/gpu.yaml
 	test -e $(TEST_DIR)/ggllm-test-model.bin || wget -q https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGUF/resolve/main/codellama-7b-instruct.Q2_K.gguf -O $(TEST_DIR)/ggllm-test-model.bin
-	docker build --build-arg GRPC_BACKENDS="$(GRPC_BACKENDS)" --build-arg IMAGE_TYPE=core --build-arg BUILD_TYPE=$(BUILD_TYPE) --build-arg CUDA_MAJOR_VERSION=12 --build-arg CUDA_MINOR_VERSION=0 --build-arg FFMPEG=true -t localai-tests .
+	docker build --build-arg GRPC_BACKENDS="$(GRPC_BACKENDS)" --build-arg IMAGE_TYPE=core --build-arg BUILD_TYPE=$(BUILD_TYPE) --build-arg CUDA_MAJOR_VERSION=12 --build-arg CUDA_MINOR_VERSION=4 --build-arg FFMPEG=true -t localai-tests .
 
 run-e2e-image:
 	ls -liah $(abspath ./tests/e2e-fixtures)
@@ -742,22 +733,13 @@ backend-assets/grpc: protogen-go replace
 backend-assets/grpc/bert-embeddings: sources/go-bert.cpp sources/go-bert.cpp/libgobert.a backend-assets/grpc
 	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-bert.cpp LIBRARY_PATH=$(CURDIR)/sources/go-bert.cpp \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/bert-embeddings ./backend/go/llm/bert/
-ifneq ($(UPX),)
-	$(UPX) backend-assets/grpc/bert-embeddings
-endif
 
 backend-assets/grpc/gpt4all: sources/gpt4all sources/gpt4all/gpt4all-bindings/golang/libgpt4all.a backend-assets/gpt4all backend-assets/grpc
 	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/gpt4all/gpt4all-bindings/golang/ LIBRARY_PATH=$(CURDIR)/sources/gpt4all/gpt4all-bindings/golang/ \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/gpt4all ./backend/go/llm/gpt4all/
-ifneq ($(UPX),)
-	$(UPX) backend-assets/grpc/gpt4all
-endif
 
 backend-assets/grpc/huggingface: backend-assets/grpc
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/huggingface ./backend/go/llm/langchain/
-ifneq ($(UPX),)
-	$(UPX) backend-assets/grpc/huggingface
-endif
 
 backend/cpp/llama/llama.cpp:
 	LLAMA_VERSION=$(CPPLLAMA_VERSION) $(MAKE) -C backend/cpp/llama llama.cpp
@@ -859,50 +841,29 @@ backend-assets/util/llama-cpp-rpc-server: backend-assets/grpc/llama-cpp-grpc
 backend-assets/grpc/llama-ggml: sources/go-llama.cpp sources/go-llama.cpp/libbinding.a backend-assets/grpc
 	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-llama.cpp LIBRARY_PATH=$(CURDIR)/sources/go-llama.cpp \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/llama-ggml ./backend/go/llm/llama-ggml/
-ifneq ($(UPX),)
-	$(UPX) backend-assets/grpc/llama-ggml
-endif
 
 backend-assets/grpc/piper: sources/go-piper sources/go-piper/libpiper_binding.a backend-assets/grpc backend-assets/espeak-ng-data
 	CGO_CXXFLAGS="$(PIPER_CGO_CXXFLAGS)" CGO_LDFLAGS="$(PIPER_CGO_LDFLAGS)" LIBRARY_PATH=$(CURDIR)/sources/go-piper \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/piper ./backend/go/tts/
-ifneq ($(UPX),)
-	$(UPX) backend-assets/grpc/piper
-endif
 
 backend-assets/grpc/rwkv: sources/go-rwkv.cpp sources/go-rwkv.cpp/librwkv.a backend-assets/grpc
 	CGO_LDFLAGS="$(CGO_LDFLAGS)" C_INCLUDE_PATH=$(CURDIR)/sources/go-rwkv.cpp LIBRARY_PATH=$(CURDIR)/sources/go-rwkv.cpp \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/rwkv ./backend/go/llm/rwkv
-ifneq ($(UPX),)
-	$(UPX) backend-assets/grpc/rwkv
-endif
 
 backend-assets/grpc/stablediffusion: sources/go-stable-diffusion sources/go-stable-diffusion/libstablediffusion.a backend-assets/grpc
 	CGO_LDFLAGS="$(CGO_LDFLAGS)" CPATH="$(CPATH):$(CURDIR)/sources/go-stable-diffusion/:/usr/include/opencv4" LIBRARY_PATH=$(CURDIR)/sources/go-stable-diffusion/ \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/stablediffusion ./backend/go/image/stablediffusion
-ifneq ($(UPX),)
-	$(UPX) backend-assets/grpc/stablediffusion
-endif
 
 backend-assets/grpc/tinydream: sources/go-tiny-dream sources/go-tiny-dream/libtinydream.a backend-assets/grpc
 	CGO_LDFLAGS="$(CGO_LDFLAGS)" LIBRARY_PATH=$(CURDIR)/go-tiny-dream \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/tinydream ./backend/go/image/tinydream
-ifneq ($(UPX),)
-	$(UPX) backend-assets/grpc/tinydream
-endif
 
 backend-assets/grpc/whisper: sources/whisper.cpp sources/whisper.cpp/libwhisper.a backend-assets/grpc
 	CGO_LDFLAGS="$(CGO_LDFLAGS) $(CGO_LDFLAGS_WHISPER)" C_INCLUDE_PATH="$(CURDIR)/sources/whisper.cpp/include:$(CURDIR)/sources/whisper.cpp/ggml/include" LIBRARY_PATH=$(CURDIR)/sources/whisper.cpp \
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/whisper ./backend/go/transcribe/
-ifneq ($(UPX),)
-	$(UPX) backend-assets/grpc/whisper
-endif
 
 backend-assets/grpc/local-store: backend-assets/grpc
 	$(GOCMD) build -ldflags "$(LD_FLAGS)" -tags "$(GO_TAGS)" -o backend-assets/grpc/local-store ./backend/go/stores/
-ifneq ($(UPX),)
-	$(UPX) backend-assets/grpc/local-store
-endif
 
 grpcs: prepare $(GRPC_BACKENDS)
 

README.md

@@ -84,7 +84,6 @@ docker run -ti --name local-ai -p 8080:8080 localai/localai:latest-aio-cpu
 
 Hot topics (looking for contributors):
 
-- 🔥🔥 Distributed, P2P Global community pools: https://github.com/mudler/LocalAI/issues/3113
 - WebUI improvements: https://github.com/mudler/LocalAI/issues/2156
 - Backends v2: https://github.com/mudler/LocalAI/issues/1126
 - Improving UX v2: https://github.com/mudler/LocalAI/issues/1373
@@ -151,7 +150,6 @@ Other:
 
 ## :book: 🎥 [Media, Blogs, Social](https://localai.io/basics/news/#media-blogs-social)
 
-- [Run Visual studio code with LocalAI (SUSE)](https://www.suse.com/c/running-ai-locally/)
 - 🆕 [Run LocalAI on Jetson Nano Devkit](https://mudler.pm/posts/local-ai-jetson-nano-devkit/)
 - [Run LocalAI on AWS EKS with Pulumi](https://www.pulumi.com/blog/low-code-llm-apps-with-local-ai-flowise-and-pulumi/)
 - [Run LocalAI on AWS](https://staleks.hashnode.dev/installing-localai-on-aws-ec2-instance)

backend/cpp/llama/grpc-server.cpp

@@ -2259,6 +2259,7 @@ static void params_parse(const backend::ModelOptions* request,
      // get the directory of modelfile
      std::string model_dir = params.model.substr(0, params.model.find_last_of("/\\"));
      params.lora_adapter.push_back(std::make_tuple(model_dir + "/"+request->loraadapter(), scale_factor));
+     params.lora_base  =  model_dir + "/"+request->lorabase();
     }
     params.use_mlock = request->mlock();
     params.use_mmap = request->mmap();

backend/python/autogptq/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/autogptq/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/autogptq/requirements-intel.txt

@@ -2,4 +2,4 @@
 intel-extension-for-pytorch
 torch
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406

backend/python/autogptq/requirements.txt

@@ -1,6 +1,7 @@
 accelerate
 auto-gptq==0.7.1
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf
+torch
 certifi
 transformers

backend/python/bark/requirements-cublas11.txt

@@ -1,3 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch
-torchaudio

backend/python/bark/requirements-cublas12.txt

@@ -1,2 +0,0 @@
-torch
-torchaudio

backend/python/bark/requirements.txt

@@ -1,6 +1,6 @@
 accelerate
 bark==0.1.5
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf
 certifi
 transformers

backend/python/common/libbackend.sh

@@ -122,13 +122,6 @@ function installRequirements() {
         requirementFiles+=("${MY_DIR}/requirements-${BUILD_PROFILE}.txt")
     fi
 
-    # if BUILD_TYPE is empty, we are a CPU build, so we should try to install the CPU requirements
-    if [ "x${BUILD_TYPE}" == "x" ]; then
-        requirementFiles+=("${MY_DIR}/requirements-cpu.txt")
-    fi
-
-    requirementFiles+=("${MY_DIR}/requirements-after.txt")
-
     for reqFile in ${requirementFiles[@]}; do
         if [ -f ${reqFile} ]; then
             echo "starting requirements install for ${reqFile}"

backend/python/common/template/requirements.txt

@@ -1,2 +1,2 @@
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf

backend/python/coqui/requirements-cublas11.txt

@@ -1,3 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch
-torchaudio

backend/python/coqui/requirements-cublas12.txt

@@ -1,2 +0,0 @@
-torch
-torchaudio

backend/python/coqui/requirements-intel.txt

@@ -3,4 +3,4 @@ intel-extension-for-pytorch
 torch
 torchaudio
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406

backend/python/coqui/requirements.txt

@@ -1,6 +1,6 @@
 accelerate
 TTS==0.22.0
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf
 certifi
 transformers

backend/python/diffusers/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/diffusers/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/diffusers/requirements.txt

@@ -3,10 +3,11 @@ accelerate
 compel
 peft
 diffusers
-grpcio==1.65.4
+grpcio==1.65.0
 opencv-python
 pillow
 protobuf
 sentencepiece
+torch
 transformers
 certifi

backend/python/exllama/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/exllama/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/exllama/requirements.txt

@@ -1,5 +1,6 @@
 grpcio==1.65.0
 protobuf
+torch
 transformers
 certifi
 setuptools

backend/python/exllama2/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/exllama2/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/exllama2/requirements.txt

@@ -1,6 +1,7 @@
 accelerate
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf
 certifi
+torch
 wheel
 setuptools

backend/python/mamba/requirements-after.txt

@@ -1,2 +0,0 @@
-causal-conv1d==1.4.0
-mamba-ssm==2.2.2

backend/python/mamba/requirements-cpu.txt

@@ -1 +0,0 @@
-torch

backend/python/mamba/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/mamba/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/mamba/requirements-install.txt

@@ -3,4 +3,5 @@
 # https://github.com/Dao-AILab/causal-conv1d/issues/24
 packaging
 setuptools
-wheel
+wheel
+torch==2.3.1

backend/python/mamba/requirements.txt

@@ -1,4 +1,6 @@
-grpcio==1.65.1
+causal-conv1d==1.4.0
+mamba-ssm==2.2.2
+grpcio==1.65.0
 protobuf
 certifi
 transformers

backend/python/openvoice/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/openvoice/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/openvoice/requirements-intel.txt

@@ -2,7 +2,7 @@
 intel-extension-for-pytorch
 torch
 optimum[openvino]
-grpcio==1.65.4
+grpcio==1.64.1
 protobuf
 librosa==0.9.1
 faster-whisper==1.0.3

backend/python/openvoice/requirements.txt

@@ -1,4 +1,4 @@
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf
 librosa
 faster-whisper

backend/python/openvoice/test.sh

@@ -5,7 +5,7 @@ source $(dirname $0)/../common/libbackend.sh
 
 # Download checkpoints if not present
 if [ ! -d "checkpoints_v2" ]; then
-    wget https://myshell-public-repo-host.s3.amazonaws.com/openvoice/checkpoints_v2_0417.zip -O checkpoints_v2.zip
+    wget https://myshell-public-repo-hosting.s3.amazonaws.com/openvoice/checkpoints_v2_0417.zip -O checkpoints_v2.zip
     unzip checkpoints_v2.zip
 fi
 

backend/python/parler-tts/requirements-cublas11.txt

@@ -1,3 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch
-torchaudio

backend/python/parler-tts/requirements-cublas12.txt

@@ -1,2 +0,0 @@
-torch
-torchaudio

backend/python/parler-tts/requirements-intel.txt

@@ -3,4 +3,4 @@ intel-extension-for-pytorch
 torch
 torchaudio
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406

backend/python/parler-tts/requirements.txt

@@ -1,6 +1,7 @@
 accelerate
-grpcio==1.65.1
+grpcio==1.65.0
 protobuf
+torch
 git+https://github.com/huggingface/parler-tts.git@10016fb0300c0dc31a0fb70e26f3affee7b62f16
 certifi
 transformers

backend/python/petals/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/petals/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/petals/requirements-intel.txt

@@ -2,4 +2,4 @@
 intel-extension-for-pytorch
 torch
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406

backend/python/rerankers/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/rerankers/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/rerankers/requirements-intel.txt

@@ -2,4 +2,4 @@
 intel-extension-for-pytorch
 torch
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406

backend/python/rerankers/requirements.txt

@@ -1,6 +1,6 @@
 accelerate
 rerankers[transformers]
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf
 certifi
 transformers

backend/python/sentencetransformers/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/sentencetransformers/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/sentencetransformers/requirements-intel.txt

@@ -2,4 +2,4 @@
 intel-extension-for-pytorch
 torch
 optimum[openvino]
-setuptools==69.5.1 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406

backend/python/sentencetransformers/requirements.txt

@@ -1,6 +1,6 @@
 accelerate
 sentence-transformers==3.0.1
 transformers
-grpcio==1.65.1
+grpcio==1.65.0
 protobuf
 certifi

backend/python/transformers-musicgen/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/transformers-musicgen/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/transformers-musicgen/requirements-intel.txt

@@ -2,4 +2,4 @@
 intel-extension-for-pytorch
 torch
 optimum[openvino]
-setuptools==69.5.1 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406

backend/python/transformers-musicgen/requirements.txt

@@ -1,6 +1,7 @@
 accelerate
 transformers
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf
+torch
 scipy==1.14.0
 certifi

backend/python/transformers/requirements-cublas11.txt

@@ -1,2 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch

backend/python/transformers/requirements-cublas12.txt

@@ -1 +0,0 @@
-torch

backend/python/transformers/requirements-intel.txt

@@ -2,3 +2,4 @@
 intel-extension-for-pytorch
 torch
 optimum[openvino]
+setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406

backend/python/transformers/requirements.txt

@@ -1,8 +1,9 @@
 accelerate
 transformers
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf
+torch
 certifi
 intel-extension-for-transformers
 bitsandbytes
-setuptools==69.5.1 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406

backend/python/vall-e-x/requirements-cublas11.txt

@@ -1,3 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch
-torchaudio

backend/python/vall-e-x/requirements-cublas12.txt

@@ -1,2 +0,0 @@
-torch
-torchaudio

backend/python/vall-e-x/requirements-intel.txt

@@ -3,4 +3,4 @@ intel-extension-for-pytorch
 torch
 torchaudio
 optimum[openvino]
-setuptools==72.1.0 # https://github.com/mudler/LocalAI/issues/2406
+setuptools==70.3.0 # https://github.com/mudler/LocalAI/issues/2406

backend/python/vall-e-x/requirements.txt

@@ -1,4 +1,4 @@
 accelerate
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf
 certifi

backend/python/vllm/requirements-cublas.txt

@@ -0,0 +1 @@
+flash-attn

backend/python/vllm/requirements-cublas11.txt

@@ -1,3 +0,0 @@
---extra-index-url https://download.pytorch.org/whl/cu118
-torch
-flash-attn

backend/python/vllm/requirements-cublas12.txt

@@ -1,2 +0,0 @@
-torch
-flash-attn

backend/python/vllm/requirements.txt

@@ -1,6 +1,6 @@
 accelerate
 vllm
-grpcio==1.65.4
+grpcio==1.65.0
 protobuf
 certifi
 transformers

core/cli/models.go

@@ -83,9 +83,7 @@ func (mi *ModelsInstall) Run(ctx *cliContext.Context) error {
 			return err
 		}
 
-		modelURI := downloader.URI(modelName)
-
-		if !modelURI.LooksLikeOCI() {
+		if !downloader.LooksLikeOCI(modelName) {
 			model := gallery.FindModel(models, modelName, mi.ModelsPath)
 			if model == nil {
 				log.Error().Str("model", modelName).Msg("model not found")

core/cli/util.go

@@ -86,8 +86,8 @@ func (hfscmd *HFScanCMD) Run(ctx *cliContext.Context) error {
 		var errs error = nil
 		for _, uri := range hfscmd.ToScan {
 			log.Info().Str("uri", uri).Msg("scanning specific uri")
-			scanResults, err := downloader.HuggingFaceScan(downloader.URI(uri))
-			if err != nil && errors.Is(err, downloader.ErrUnsafeFilesFound) {
+			scanResults, err := downloader.HuggingFaceScan(uri)
+			if err != nil && !errors.Is(err, downloader.ErrNonHuggingFaceFile) {
 				log.Error().Err(err).Strs("clamAV", scanResults.ClamAVInfectedFiles).Strs("pickles", scanResults.DangerousPickles).Msg("! WARNING ! A known-vulnerable model is included in this repo!")
 				errs = errors.Join(errs, err)
 			}

core/config/backend_config.go

@@ -8,6 +8,7 @@ import (
 	"github.com/mudler/LocalAI/core/schema"
 	"github.com/mudler/LocalAI/pkg/downloader"
 	"github.com/mudler/LocalAI/pkg/functions"
+	"github.com/mudler/LocalAI/pkg/utils"
 )
 
 const (
@@ -71,9 +72,9 @@ type BackendConfig struct {
 }
 
 type File struct {
-	Filename string         `yaml:"filename" json:"filename"`
-	SHA256   string         `yaml:"sha256" json:"sha256"`
-	URI      downloader.URI `yaml:"uri" json:"uri"`
+	Filename string `yaml:"filename" json:"filename"`
+	SHA256   string `yaml:"sha256" json:"sha256"`
+	URI      string `yaml:"uri" json:"uri"`
 }
 
 type VallE struct {
@@ -212,32 +213,28 @@ func (c *BackendConfig) ShouldCallSpecificFunction() bool {
 // MMProjFileName returns the filename of the MMProj file
 // If the MMProj is a URL, it will return the MD5 of the URL which is the filename
 func (c *BackendConfig) MMProjFileName() string {
-	uri := downloader.URI(c.MMProj)
-	if uri.LooksLikeURL() {
-		f, _ := uri.FilenameFromUrl()
-		return f
+	modelURL := downloader.ConvertURL(c.MMProj)
+	if downloader.LooksLikeURL(modelURL) {
+		return utils.MD5(modelURL)
 	}
 
 	return c.MMProj
 }
 
 func (c *BackendConfig) IsMMProjURL() bool {
-	uri := downloader.URI(c.MMProj)
-	return uri.LooksLikeURL()
+	return downloader.LooksLikeURL(downloader.ConvertURL(c.MMProj))
 }
 
 func (c *BackendConfig) IsModelURL() bool {
-	uri := downloader.URI(c.Model)
-	return uri.LooksLikeURL()
+	return downloader.LooksLikeURL(downloader.ConvertURL(c.Model))
 }
 
 // ModelFileName returns the filename of the model
 // If the model is a URL, it will return the MD5 of the URL which is the filename
 func (c *BackendConfig) ModelFileName() string {
-	uri := downloader.URI(c.Model)
-	if uri.LooksLikeURL() {
-		f, _ := uri.FilenameFromUrl()
-		return f
+	modelURL := downloader.ConvertURL(c.Model)
+	if downloader.LooksLikeURL(modelURL) {
+		return utils.MD5(modelURL)
 	}
 
 	return c.Model

core/config/backend_config_loader.go

@@ -244,7 +244,7 @@ func (bcl *BackendConfigLoader) Preload(modelPath string) error {
 			// Create file path
 			filePath := filepath.Join(modelPath, file.Filename)
 
-			if err := file.URI.DownloadFile(filePath, file.SHA256, i, len(config.DownloadFiles), status); err != nil {
+			if err := downloader.DownloadFile(file.URI, filePath, file.SHA256, i, len(config.DownloadFiles), status); err != nil {
 				return err
 			}
 		}
@@ -252,10 +252,10 @@ func (bcl *BackendConfigLoader) Preload(modelPath string) error {
 		// If the model is an URL, expand it, and download the file
 		if config.IsModelURL() {
 			modelFileName := config.ModelFileName()
-			uri := downloader.URI(config.Model)
+			modelURL := downloader.ConvertURL(config.Model)
 			// check if file exists
 			if _, err := os.Stat(filepath.Join(modelPath, modelFileName)); errors.Is(err, os.ErrNotExist) {
-				err := uri.DownloadFile(filepath.Join(modelPath, modelFileName), "", 0, 0, status)
+				err := downloader.DownloadFile(modelURL, filepath.Join(modelPath, modelFileName), "", 0, 0, status)
 				if err != nil {
 					return err
 				}
@@ -269,10 +269,10 @@ func (bcl *BackendConfigLoader) Preload(modelPath string) error {
 
 		if config.IsMMProjURL() {
 			modelFileName := config.MMProjFileName()
-			uri := downloader.URI(config.MMProj)
+			modelURL := downloader.ConvertURL(config.MMProj)
 			// check if file exists
 			if _, err := os.Stat(filepath.Join(modelPath, modelFileName)); errors.Is(err, os.ErrNotExist) {
-				err := uri.DownloadFile(filepath.Join(modelPath, modelFileName), "", 0, 0, status)
+				err := downloader.DownloadFile(modelURL, filepath.Join(modelPath, modelFileName), "", 0, 0, status)
 				if err != nil {
 					return err
 				}

core/config/guesser.go

@@ -26,17 +26,15 @@ const (
 type settingsConfig struct {
 	StopWords      []string
 	TemplateConfig TemplateConfig
-	RepeatPenalty float64
 }
 
 // default settings to adopt with a given model family
 var defaultsSettings map[familyType]settingsConfig = map[familyType]settingsConfig{
 	Gemma: {
-		RepeatPenalty: 1.0,
 		StopWords: []string{"<|im_end|>", "<end_of_turn>", "<start_of_turn>"},
 		TemplateConfig: TemplateConfig{
-			Chat:        "{{.Input }}\n<start_of_turn>model\n",
-			ChatMessage: "<start_of_turn>{{if eq .RoleName \"assistant\" }}model{{else}}{{ .RoleName }}{{end}}\n{{ if .Content -}}\n{{.Content -}}\n{{ end -}}<end_of_turn>",
+			Chat:        "{{.Input }}\n<|start_of_turn|>model\n",
+			ChatMessage: "<|start_of_turn|>{{if eq .RoleName \"assistant\" }}model{{else}}{{ .RoleName }}{{end}}\n{{ if .Content -}}\n{{.Content -}}\n{{ end -}}<|end_of_turn|>",
 			Completion:  "{{.Input}}",
 		},
 	},
@@ -194,9 +192,6 @@ func guessDefaultsFromFile(cfg *BackendConfig, modelPath string) {
 		if len(cfg.StopWords) == 0 {
 			cfg.StopWords = settings.StopWords
 		}
-		if cfg.RepeatPenalty == 0.0 {
-			cfg.RepeatPenalty = settings.RepeatPenalty
-		}
 	} else {
 		log.Debug().Any("family", family).Msgf("guessDefaultsFromFile: no template found for family")
 	}
@@ -224,7 +219,7 @@ func identifyFamily(f *gguf.GGUFFile) familyType {
 	commandR := arch == "command-r" && eosTokenID == 255001
 	qwen2 := arch == "qwen2"
 	phi3 := arch == "phi-3"
-	gemma := strings.HasPrefix(arch, "gemma") || strings.Contains(strings.ToLower(f.Model().Name), "gemma")
+	gemma := strings.HasPrefix(f.Model().Name, "gemma")
 	deepseek2 := arch == "deepseek2"
 
 	switch {

core/dependencies_manager/manager.go

@@ -37,8 +37,7 @@ func main() {
 
 	// download the assets
 	for _, asset := range assets {
-		uri := downloader.URI(asset.URL)
-		if err := uri.DownloadFile(filepath.Join(destPath, asset.FileName), asset.SHA, 1, 1, utils.DisplayDownloadFunction); err != nil {
+		if err := downloader.DownloadFile(asset.URL, filepath.Join(destPath, asset.FileName), asset.SHA, 1, 1, utils.DisplayDownloadFunction); err != nil {
 			panic(err)
 		}
 	}

core/gallery/gallery.go

@@ -131,8 +131,7 @@ func AvailableGalleryModels(galleries []config.Gallery, basePath string) ([]*Gal
 
 func findGalleryURLFromReferenceURL(url string, basePath string) (string, error) {
 	var refFile string
-	uri := downloader.URI(url)
-	err := uri.DownloadAndUnmarshal(basePath, func(url string, d []byte) error {
+	err := downloader.DownloadAndUnmarshal(url, basePath, func(url string, d []byte) error {
 		refFile = string(d)
 		if len(refFile) == 0 {
 			return fmt.Errorf("invalid reference file at url %s: %s", url, d)
@@ -154,9 +153,8 @@ func getGalleryModels(gallery config.Gallery, basePath string) ([]*GalleryModel,
 			return models, err
 		}
 	}
-	uri := downloader.URI(gallery.URL)
 
-	err := uri.DownloadAndUnmarshal(basePath, func(url string, d []byte) error {
+	err := downloader.DownloadAndUnmarshal(gallery.URL, basePath, func(url string, d []byte) error {
 		return yaml.Unmarshal(d, &models)
 	})
 	if err != nil {
@@ -206,34 +204,35 @@ func DeleteModelFromSystem(basePath string, name string, additionalFiles []strin
 		log.Error().Err(err).Msgf("failed to read gallery file %s", configFile)
 	}
 
-	var filesToRemove []string
-
 	// Remove additional files
 	if galleryconfig != nil {
 		for _, f := range galleryconfig.Files {
 			fullPath := filepath.Join(basePath, f.Filename)
-			filesToRemove = append(filesToRemove, fullPath)
+			log.Debug().Msgf("Removing file %s", fullPath)
+			if e := os.Remove(fullPath); e != nil {
+				err = errors.Join(err, fmt.Errorf("failed to remove file %s: %w", f.Filename, e))
+			}
 		}
 	}
 
 	for _, f := range additionalFiles {
 		fullPath := filepath.Join(filepath.Join(basePath, f))
-		filesToRemove = append(filesToRemove, fullPath)
-	}
-
-	filesToRemove = append(filesToRemove, configFile)
-	filesToRemove = append(filesToRemove, galleryFile)
-
-	// skip duplicates
-	filesToRemove = utils.Unique(filesToRemove)
-
-	// Removing files
-	for _, f := range filesToRemove {
-		if e := os.Remove(f); e != nil {
+		log.Debug().Msgf("Removing additional file %s", fullPath)
+		if e := os.Remove(fullPath); e != nil {
 			err = errors.Join(err, fmt.Errorf("failed to remove file %s: %w", f, e))
 		}
 	}
 
+	log.Debug().Msgf("Removing model config file %s", configFile)
+
+	// Delete the model config file
+	if e := os.Remove(configFile); e != nil {
+		err = errors.Join(err, fmt.Errorf("failed to remove file %s: %w", configFile, e))
+	}
+
+	// Delete gallery config file
+	os.Remove(galleryFile)
+
 	return err
 }
 
@@ -254,8 +253,8 @@ func SafetyScanGalleryModels(galleries []config.Gallery, basePath string) error
 
 func SafetyScanGalleryModel(galleryModel *GalleryModel) error {
 	for _, file := range galleryModel.AdditionalFiles {
-		scanResults, err := downloader.HuggingFaceScan(downloader.URI(file.URI))
-		if err != nil && errors.Is(err, downloader.ErrUnsafeFilesFound) {
+		scanResults, err := downloader.HuggingFaceScan(file.URI)
+		if err != nil && !errors.Is(err, downloader.ErrNonHuggingFaceFile) {
 			log.Error().Str("model", galleryModel.Name).Strs("clamAV", scanResults.ClamAVInfectedFiles).Strs("pickles", scanResults.DangerousPickles).Msg("Contains unsafe file(s)!")
 			return err
 		}

core/gallery/models.go

@@ -68,8 +68,7 @@ type PromptTemplate struct {
 
 func GetGalleryConfigFromURL(url string, basePath string) (Config, error) {
 	var config Config
-	uri := downloader.URI(url)
-	err := uri.DownloadAndUnmarshal(basePath, func(url string, d []byte) error {
+	err := downloader.DownloadAndUnmarshal(url, basePath, func(url string, d []byte) error {
 		return yaml.Unmarshal(d, &config)
 	})
 	if err != nil {
@@ -119,14 +118,14 @@ func InstallModel(basePath, nameOverride string, config *Config, configOverrides
 		filePath := filepath.Join(basePath, file.Filename)
 
 		if enforceScan {
-			scanResults, err := downloader.HuggingFaceScan(downloader.URI(file.URI))
-			if err != nil && errors.Is(err, downloader.ErrUnsafeFilesFound) {
+			scanResults, err := downloader.HuggingFaceScan(file.URI)
+			if err != nil && !errors.Is(err, downloader.ErrNonHuggingFaceFile) {
 				log.Error().Str("model", config.Name).Strs("clamAV", scanResults.ClamAVInfectedFiles).Strs("pickles", scanResults.DangerousPickles).Msg("Contains unsafe file(s)!")
 				return err
 			}
 		}
-		uri := downloader.URI(file.URI)
-		if err := uri.DownloadFile(filePath, file.SHA256, i, len(config.Files), downloadStatus); err != nil {
+
+		if err := downloader.DownloadFile(file.URI, filePath, file.SHA256, i, len(config.Files), downloadStatus); err != nil {
 			return err
 		}
 	}

core/http/app_test.go

@@ -73,9 +73,8 @@ func getModelStatus(url string) (response map[string]interface{}) {
 }
 
 func getModels(url string) (response []gallery.GalleryModel) {
-	uri := downloader.URI(url)
 	// TODO: No tests currently seem to exercise file:// urls. Fix?
-	uri.DownloadAndUnmarshal("", func(url string, i []byte) error {
+	downloader.DownloadAndUnmarshal(url, "", func(url string, i []byte) error {
 		// Unmarshal YAML data into a struct
 		return json.Unmarshal(i, &response)
 	})

core/http/elements/gallery.go

@@ -9,6 +9,7 @@ import (
 	"github.com/mudler/LocalAI/core/gallery"
 	"github.com/mudler/LocalAI/core/p2p"
 	"github.com/mudler/LocalAI/core/services"
+	"github.com/mudler/LocalAI/pkg/xsync"
 )
 
 const (
@@ -371,12 +372,7 @@ func dropBadChars(s string) string {
 	return strings.ReplaceAll(s, "@", "__")
 }
 
-type ProcessTracker interface {
-	Exists(string) bool
-	Get(string) string
-}
-
-func ListModels(models []*gallery.GalleryModel, processTracker ProcessTracker, galleryService *services.GalleryService) string {
+func ListModels(models []*gallery.GalleryModel, processing *xsync.SyncedMap[string, string], galleryService *services.GalleryService) string {
 	modelsElements := []elem.Node{}
 	descriptionDiv := func(m *gallery.GalleryModel) elem.Node {
 		return elem.Div(
@@ -400,7 +396,7 @@ func ListModels(models []*gallery.GalleryModel, processTracker ProcessTracker, g
 
 	actionDiv := func(m *gallery.GalleryModel) elem.Node {
 		galleryID := fmt.Sprintf("%s@%s", m.Gallery.Name, m.Name)
-		currentlyProcessing := processTracker.Exists(galleryID)
+		currentlyProcessing := processing.Exists(galleryID)
 		jobID := ""
 		isDeletionOp := false
 		if currentlyProcessing {
@@ -408,7 +404,7 @@ func ListModels(models []*gallery.GalleryModel, processTracker ProcessTracker, g
 			if status != nil && status.Deletion {
 				isDeletionOp = true
 			}
-			jobID = processTracker.Get(galleryID)
+			jobID = processing.Get(galleryID)
 			// TODO:
 			// case not handled, if status == nil : "Waiting"
 		}

core/http/endpoints/localai/welcome.go

@@ -17,10 +17,7 @@ func WelcomeEndpoint(appConfig *config.ApplicationConfig,
 		backendConfigs := cl.GetAllBackendConfigs()
 
 		galleryConfigs := map[string]*gallery.Config{}
-		modelsWithBackendConfig := map[string]interface{}{}
-
 		for _, m := range backendConfigs {
-			modelsWithBackendConfig[m.Name] = nil
 
 			cfg, err := gallery.GetLocalModelConfiguration(ml.ModelPath, m.Name)
 			if err != nil {
@@ -35,7 +32,7 @@ func WelcomeEndpoint(appConfig *config.ApplicationConfig,
 		modelsWithoutConfig := []string{}
 
 		for _, m := range models {
-			if _, ok := modelsWithBackendConfig[m]; !ok {
+			if _, ok := galleryConfigs[m]; !ok {
 				modelsWithoutConfig = append(modelsWithoutConfig, m)
 			}
 		}

core/http/endpoints/openai/chat.go

@@ -201,7 +201,7 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, startup
 		}
 
 		switch {
-		case !config.FunctionsConfig.GrammarConfig.NoGrammar && shouldUseFn:
+		case config.FunctionsConfig.GrammarConfig.EnableGrammar && shouldUseFn:
 			noActionGrammar := functions.Function{
 				Name:        noActionName,
 				Description: noActionDescription,
@@ -226,15 +226,9 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, startup
 
 			// Update input grammar
 			jsStruct := funcs.ToJSONStructure(config.FunctionsConfig.FunctionNameKey, config.FunctionsConfig.FunctionNameKey)
-			g, err := jsStruct.Grammar(config.FunctionsConfig.GrammarOptions()...)
-			if err == nil {
-				config.Grammar = g
-			}
+			config.Grammar = jsStruct.Grammar(config.FunctionsConfig.GrammarConfig.Options()...)
 		case input.JSONFunctionGrammarObject != nil:
-			g, err := input.JSONFunctionGrammarObject.Grammar(config.FunctionsConfig.GrammarOptions()...)
-			if err == nil {
-				config.Grammar = g
-			}
+			config.Grammar = input.JSONFunctionGrammarObject.Grammar(config.FunctionsConfig.GrammarConfig.Options()...)
 		default:
 			// Force picking one of the functions by the request
 			if config.FunctionToCall() != "" {

core/http/routes/ui.go

@@ -21,40 +21,6 @@ import (
 	"github.com/google/uuid"
 )
 
-type modelOpCache struct {
-	status *xsync.SyncedMap[string, string]
-}
-
-func NewModelOpCache() *modelOpCache {
-	return &modelOpCache{
-		status: xsync.NewSyncedMap[string, string](),
-	}
-}
-
-func (m *modelOpCache) Set(key string, value string) {
-	m.status.Set(key, value)
-}
-
-func (m *modelOpCache) Get(key string) string {
-	return m.status.Get(key)
-}
-
-func (m *modelOpCache) DeleteUUID(uuid string) {
-	for _, k := range m.status.Keys() {
-		if m.status.Get(k) == uuid {
-			m.status.Delete(k)
-		}
-	}
-}
-
-func (m *modelOpCache) Map() map[string]string {
-	return m.status.Map()
-}
-
-func (m *modelOpCache) Exists(key string) bool {
-	return m.status.Exists(key)
-}
-
 func RegisterUIRoutes(app *fiber.App,
 	cl *config.BackendConfigLoader,
 	ml *model.ModelLoader,
@@ -63,7 +29,7 @@ func RegisterUIRoutes(app *fiber.App,
 	auth func(*fiber.Ctx) error) {
 
 	// keeps the state of models that are being installed from the UI
-	var processingModels = NewModelOpCache()
+	var processingModels = xsync.NewSyncedMap[string, string]()
 
 	// modelStatus returns the current status of the models being processed (installation or deletion)
 	// it is called asynchonously from the UI
@@ -266,8 +232,6 @@ func RegisterUIRoutes(app *fiber.App,
 			return c.SendString(elements.ProgressBar("100"))
 		}
 		if status.Error != nil {
-			// TODO: instead of deleting the job, we should keep it in the cache and make it dismissable
-			processingModels.DeleteUUID(jobUID)
 			return c.SendString(elements.ErrorProgress(status.Error.Error(), status.GalleryModelName))
 		}
 
@@ -282,7 +246,12 @@ func RegisterUIRoutes(app *fiber.App,
 		status := galleryService.GetStatus(jobUID)
 
 		galleryID := ""
-		processingModels.DeleteUUID(jobUID)
+		for _, k := range processingModels.Keys() {
+			if processingModels.Get(k) == jobUID {
+				galleryID = k
+				processingModels.Delete(k)
+			}
+		}
 		if galleryID == "" {
 			log.Debug().Msgf("no processing model found for job : %+v\n", jobUID)
 		}

core/http/views/p2p.html

@@ -16,16 +16,7 @@
                 </a> 
             </h2> 
             <h5 class="mb-4 text-justify">LocalAI uses P2P technologies to enable distribution of work between peers. It is possible to share an instance with Federation and/or split the weights of a model across peers (only available with llama.cpp models). You can now share computational resources between your devices or your friends!</h5>
-            <!-- Warning box if p2p token is empty and p2p is enabled -->
-            {{ if and .IsP2PEnabled (eq .P2PToken "") }}
-            <div class="bg-red-500 p-4 rounded-lg shadow-lg mb-12 text-left">
-                <p class="text-xl font-semibold text-white"> <i class="fa-solid fa-exclamation-triangle"></i> Warning: P2P mode is disabled or no token was specified</p>
-                <p class="mb-4">You have to enable P2P mode by starting LocalAI with <code>--p2p</code>. Please restart the server with <code>--p2p</code> to generate a new token automatically that can be used to automatically discover other nodes. If you already have a token specify it with <code>export TOKEN=".."</code> <a href="https://localai.io/features/distribute/" target="_blank">
-                    Check out the documentation for more information.
-                </a> </p>
-            </div>
-            {{ else }}
-
+            
             <!-- Federation Box -->
             <div class="bg-gray-800 p-6 rounded-lg shadow-lg mb-12 text-left">
 
@@ -137,8 +128,7 @@
                     </div>
                 </div>
             </div>
-            <!-- Llama.cpp Box END -->    
-            {{ end }}   
+            <!-- Llama.cpp Box END -->       
         </div>
     </div>
 

docs/content/docs/features/distributed_inferencing.md

@@ -5,65 +5,17 @@ weight = 15
 url = "/features/distribute/"
 +++
 
-
-This functionality enables LocalAI to distribute inference requests across multiple worker nodes, improving efficiency and performance. Nodes are automatically discovered and connect via p2p by using a shared token which makes sure the communication is secure and private between the nodes of the network.
-
-LocalAI supports two modes of distributed inferencing via p2p:
-
-- **Federated Mode**: Requests are shared between the cluster and routed to a single worker node in the network based on the load balancer's decision.
-- **Worker Mode** (aka "model sharding" or "splitting weights"): Requests are processed by all the workers which contributes to the final inference result (by sharing the model weights).
-
-## Usage
-
-Starting LocalAI with `--p2p` generates a shared token for connecting multiple instances: and that's all you need to create AI clusters, eliminating the need for intricate network setups. 
-
-Simply navigate to the "Swarm" section in the WebUI and follow the on-screen instructions.
-
-For fully shared instances, initiate LocalAI with --p2p --federated and adhere to the Swarm section's guidance. This feature, while still experimental, offers a tech preview quality experience.
-
-### Federated mode
-
-Federated mode allows to launch multiple LocalAI instances and connect them together in a federated network. This mode is useful when you want to distribute the load of the inference across multiple nodes, but you want to have a single point of entry for the API. In the Swarm section of the WebUI, you can see the instructions to connect multiple instances together.
-
-![346663124-1d2324fd-8b55-4fa2-9856-721a467969c2](https://github.com/user-attachments/assets/19ebd44a-20ff-412c-b92f-cfb8efbe4b21)
-
-To start a LocalAI server in federated mode, run:
-
-```bash
-local-ai run --p2p --federated
-```
-
-This will generate a token that you can use to connect other LocalAI instances to the network or others can use to join the network. If you already have a token, you can specify it using the `TOKEN` environment variable.
-
-To start a load balanced server that routes the requests to the network, run with the `TOKEN`:
-
-```bash
-local-ai federated
-```
-
-To see all the available options, run `local-ai federated --help`.
-
-The instructions are displayed in the "Swarm" section of the WebUI, guiding you through the process of connecting multiple instances.
-
-### Workers mode
-
 {{% alert note %}}
 This feature is available exclusively with llama-cpp compatible models.
 
 This feature was introduced in [LocalAI pull request #2324](https://github.com/mudler/LocalAI/pull/2324) and is based on the upstream work in [llama.cpp pull request #6829](https://github.com/ggerganov/llama.cpp/pull/6829).
 {{% /alert %}}
 
-To connect multiple workers to a single LocalAI instance, start first a server in p2p mode:
+This functionality enables LocalAI to distribute inference requests across multiple worker nodes, improving efficiency and performance.
 
-```bash
-local-ai run --p2p
-```
+## Usage
 
-And navigate the WebUI to the "Swarm" section to see the instructions to connect multiple workers to the network.
-
-![346663124-1d2324fd-8b55-4fa2-9856-721a467969c2](https://github.com/user-attachments/assets/b8cadddf-a467-49cf-a1ed-8850de95366d)
-
-### Without P2P
+### Starting Workers
 
 To start workers for distributing the computational load, run:
 
@@ -71,27 +23,48 @@ To start workers for distributing the computational load, run:
 local-ai worker llama-cpp-rpc <listening_address> <listening_port>
 ```
 
-And you can specify the address of the workers when starting LocalAI with the `LLAMACPP_GRPC_SERVERS` environment variable:
+Alternatively, you can build the RPC server following the llama.cpp [README](https://github.com/ggerganov/llama.cpp/blob/master/examples/rpc/README.md), which is compatible with LocalAI.
+
+### Starting LocalAI
+
+To start the LocalAI server, which handles API requests, specify the worker addresses using the `LLAMACPP_GRPC_SERVERS` environment variable:
 
 ```bash
 LLAMACPP_GRPC_SERVERS="address1:port,address2:port" local-ai run
 ```
+
 The workload on the LocalAI server will then be distributed across the specified nodes.
 
-Alternatively, you can build the RPC workers/server following the llama.cpp [README](https://github.com/ggerganov/llama.cpp/blob/master/examples/rpc/README.md), which is compatible with LocalAI.
+## Peer-to-Peer Networking
 
-## Manual example (worker)
+![output](https://github.com/mudler/LocalAI/assets/2420543/8ca277cf-c208-4562-8929-808b2324b584)
 
-Use the WebUI to guide you in the process of starting new workers. This example shows the manual steps to highlight the process.
+Workers can also connect to each other in a peer-to-peer network, distributing the workload in a decentralized manner.
+
+A shared token between the server and the workers is required for communication within the peer-to-peer network. This feature supports both local network (using mDNS discovery) and DHT for communication across different networks.
+
+The token is automatically generated when starting the server with the `--p2p` flag. Workers can be started with the token using `local-ai worker p2p-llama-cpp-rpc` and specifying the token via the environment variable `TOKEN` or with the `--token` argument.
+
+A network is established between the server and workers using DHT and mDNS discovery protocols. The llama.cpp RPC server is automatically started and exposed to the peer-to-peer network, allowing the API server to connect.
+
+When the HTTP server starts, it discovers workers in the network and creates port forwards to the local service. Llama.cpp is configured to use these services. For more details on the implementation, refer to [LocalAI pull request #2343](https://github.com/mudler/LocalAI/pull/2343).
+
+### Usage
 
 1. Start the server with `--p2p`:
 
 ```bash
 ./local-ai run --p2p
-# Get the token in the Swarm section of the WebUI
+# 1:02AM INF loading environment variables from file envFile=.env
+# 1:02AM INF Setting logging to info
+# 1:02AM INF P2P mode enabled
+# 1:02AM INF No token provided, generating one
+# 1:02AM INF Generated Token:
+# XXXXXXXXXXX
+# 1:02AM INF Press a button to proceed
 ```
 
-Copy the token from the WebUI or via API call (e.g., `curl http://localhost:8000/p2p/token`) and save it for later use.
+Copy the displayed token and press Enter.
 
 To reuse the same token later, restart the server with `--p2ptoken` or `P2P_TOKEN`.
 
@@ -120,7 +93,11 @@ The server logs should indicate that new workers are being discovered.
 
 3. Start inference as usual on the server initiated in step 1.
 
-![output](https://github.com/mudler/LocalAI/assets/2420543/8ca277cf-c208-4562-8929-808b2324b584)
+## Notes
+
+- If running in p2p mode with container images, make sure you start the container with `--net host` or `network_mode: host` in the docker-compose file.
+- Only a single model is supported currently.
+- Ensure the server detects new workers before starting inference. Currently, additional workers cannot be added once inference has begun.
 
 
 ## Environment Variables
@@ -132,20 +109,3 @@ There are options that can be tweaked or parameters that can be set using enviro
 | **LOCALAI_P2P_DISABLE_DHT** | Set to "true" to disable DHT and enable p2p layer to be local only (mDNS) |
 | **LOCALAI_P2P_DISABLE_LIMITS** | Set to "true" to disable connection limits and resources management |
 | **LOCALAI_P2P_TOKEN** | Set the token for the p2p network |
-
-## Architecture
-
-LocalAI uses https://github.com/libp2p/go-libp2p under the hood, the same project powering IPFS. Differently from other frameworks, LocalAI uses peer2peer without a single master server, but rather it uses sub/gossip and ledger functionalities to achieve consensus across different peers. 
-
-[EdgeVPN](https://github.com/mudler/edgevpn) is used as a library to establish the network and expose the ledger functionality under a shared token to ease out automatic discovery and have separated, private peer2peer networks.
-
-The weights are split proportional to the memory when running into worker mode, when in federation mode each request is split to every node which have to load the model fully.
-
-## Notes
-
-- If running in p2p mode with container images, make sure you start the container with `--net host` or `network_mode: host` in the docker-compose file.
-- Only a single model is supported currently.
-- Ensure the server detects new workers before starting inference. Currently, additional workers cannot be added once inference has begun.
-- For more details on the implementation, refer to [LocalAI pull request #2343](https://github.com/mudler/LocalAI/pull/2343)
-
-

docs/data/version.json

@@ -1,3 +1,3 @@
 {
-  "version": "v2.19.4"
+  "version": "v2.18.1"
 }

docs/static/install.sh

@@ -194,7 +194,7 @@ install_container_toolkit_yum() {
     curl -s -L https://nvidia.github.io/libnvidia-container/stable/rpm/nvidia-container-toolkit.repo | \
     $SUDO  tee /etc/yum.repos.d/nvidia-container-toolkit.repo
 
-    if [ "$PACKAGE_MANAGER" = "dnf" ]; then
+    if [ "$PACKAGE_MANAGER" == "dnf" ]; then
         $SUDO $PACKAGE_MANAGER config-manager --enable nvidia-container-toolkit-experimental
     else 
         $SUDO $PACKAGE_MANAGER -y install yum-utils
@@ -629,7 +629,7 @@ case "$ARCH" in
     *) fatal "Unsupported architecture: $ARCH" ;;
 esac
 
-if [ "$OS" = "Darwin" ]; then
+if [ "$OS" == "Darwin" ]; then
     install_binary_darwin
     exit 0
 fi

embedded/embedded.go

@@ -38,8 +38,8 @@ func init() {
 
 func GetRemoteLibraryShorteners(url string, basePath string) (map[string]string, error) {
 	remoteLibrary := map[string]string{}
-	uri := downloader.URI(url)
-	err := uri.DownloadAndUnmarshal(basePath, func(_ string, i []byte) error {
+
+	err := downloader.DownloadAndUnmarshal(url, basePath, func(_ string, i []byte) error {
 		return yaml.Unmarshal(i, &remoteLibrary)
 	})
 	if err != nil {

examples/chainlit/requirements.txt

@@ -1,6 +1,6 @@
-llama_index==0.10.59
+llama_index==0.10.55
 requests==2.32.3
-weaviate_client==4.6.7
+weaviate_client==4.6.5
 transformers
 torch
 chainlit

examples/functions/requirements.txt

@@ -1,2 +1,2 @@
-langchain==0.2.12
-openai==1.37.0
+langchain==0.2.8
+openai==1.35.13

examples/langchain-chroma/requirements.txt

@@ -1,4 +1,4 @@
-langchain==0.2.12
-openai==1.37.0
-chromadb==0.5.5
-llama-index==0.10.56
+langchain==0.2.8
+openai==1.35.13
+chromadb==0.5.4
+llama-index==0.10.55

examples/langchain/langchainpy-localai-example/requirements.txt

@@ -10,21 +10,21 @@ debugpy==1.8.2
 frozenlist==1.4.1
 greenlet==3.0.3
 idna==3.7
-langchain==0.2.12
-langchain-community==0.2.9
+langchain==0.2.8
+langchain-community==0.2.7
 marshmallow==3.21.3
 marshmallow-enum==1.5.1
 multidict==6.0.5
 mypy-extensions==1.0.0
 numexpr==2.10.1
-numpy==2.0.1
-openai==1.37.1
+numpy==1.26.4
+openai==1.35.13
 openapi-schema-pydantic==1.2.4
 packaging>=23.2
 pydantic==2.8.2
 PyYAML==6.0.1
 requests==2.32.3
-SQLAlchemy==2.0.32
+SQLAlchemy==2.0.30
 tenacity==8.5.0
 tqdm==4.66.4
 typing-inspect==0.9.0

examples/streamlit-bot/requirements.txt

@@ -1,2 +1,2 @@
-streamlit==1.37.1
+streamlit==1.36.0
 requests