llama-cpp-python 0.1.8__tar.gz → 0.1.10__tar.gz

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: llama_cpp_python
-Version: 0.1.8
+Version: 0.1.10
 Summary: A Python wrapper for llama.cpp
 Author: Andrei Betlen
 Author-email: abetlen@gmail.com

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/llama_cpp/llama.py

@@ -105,6 +105,20 @@ class Llama:
             output += llama_cpp.llama_token_to_str(self.ctx, token)
         return output
 
+    def embed(self, text: str):
+        """Embed a string.
+
+        Args:
+            text: The utf-8 encoded string to embed.
+
+        Returns:
+            A list of embeddings.
+        """
+        tokens = self.tokenize(text.encode("utf-8"))
+        self._eval(tokens, 0)
+        embeddings = llama_cpp.llama_get_embeddings(self.ctx)
+        return embeddings[:llama_cpp.llama_n_embd(self.ctx)]
+
     def _eval(self, tokens: List[int], n_past):
         rc = llama_cpp.llama_eval(
             self.ctx,

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/llama_cpp/llama_cpp.py

@@ -3,7 +3,6 @@ import ctypes
 from ctypes import (
     c_int,
     c_float,
-    c_double,
     c_char_p,
     c_void_p,
     c_bool,
@@ -40,7 +39,7 @@ class llama_token_data(Structure):
 
 llama_token_data_p = POINTER(llama_token_data)
 
-llama_progress_callback = ctypes.CFUNCTYPE(None, c_double, c_void_p)
+llama_progress_callback = ctypes.CFUNCTYPE(None, c_float, c_void_p)
 
 
 class llama_context_params(Structure):
@@ -48,6 +47,7 @@ class llama_context_params(Structure):
         ("n_ctx", c_int),  # text context
         ("n_parts", c_int),  # -1 for default
         ("seed", c_int),  # RNG seed, 0 for random
+
         ("f16_kv", c_bool),  # use fp16 for KV cache
         (
             "logits_all",
@@ -56,6 +56,7 @@ class llama_context_params(Structure):
         ("vocab_only", c_bool),  # only load the vocabulary, no weights
         ("use_mlock", c_bool),  # force system to keep model in RAM
         ("embedding", c_bool),  # embedding mode only
+
         # called with a progress value between 0 and 1, pass NULL to disable
         ("progress_callback", llama_progress_callback),
         # context pointer passed to the progress callback
@@ -70,8 +71,7 @@ llama_context_params_p = POINTER(llama_context_params)
 
 
 def llama_context_default_params() -> llama_context_params:
-    params = _lib.llama_context_default_params()
-    return params
+    return _lib.llama_context_default_params()
 
 
 _lib.llama_context_default_params.argtypes = []
@@ -229,9 +229,9 @@ def llama_sample_top_p_top_k(
     last_n_tokens_data: llama_token_p,
     last_n_tokens_size: c_int,
     top_k: c_int,
-    top_p: c_double,
-    temp: c_double,
-    repeat_penalty: c_double,
+    top_p: c_float,
+    temp: c_float,
+    repeat_penalty: c_float,
 ) -> llama_token:
     return _lib.llama_sample_top_p_top_k(
         ctx, last_n_tokens_data, last_n_tokens_size, top_k, top_p, temp, repeat_penalty
@@ -243,9 +243,9 @@ _lib.llama_sample_top_p_top_k.argtypes = [
     llama_token_p,
     c_int,
     c_int,
-    c_double,
-    c_double,
-    c_double,
+    c_float,
+    c_float,
+    c_float,
 ]
 _lib.llama_sample_top_p_top_k.restype = llama_token
 

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/llama_cpp_python.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: llama-cpp-python
-Version: 0.1.8
+Version: 0.1.10
 Summary: A Python wrapper for llama.cpp
 Author: Andrei Betlen
 Author-email: abetlen@gmail.com

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "llama_cpp"
-version = "0.1.8"
+version = "0.1.10"
 description = "Python bindings for the llama.cpp library"
 authors = ["Andrei Betlen <abetlen@gmail.com>"]
 license = "MIT"

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/setup.py

@@ -3,7 +3,7 @@ from skbuild import setup
 setup(
     name="llama_cpp_python",
     description="A Python wrapper for llama.cpp",
-    version="0.1.8",
+    version="0.1.10",
     author="Andrei Betlen",
     author_email="abetlen@gmail.com",
     license="MIT",

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/vendor/llama.cpp/.github/workflows/build.yml

@@ -8,10 +8,10 @@ on:
         required: true
         type: boolean
   push:
-    paths: ['.github/workflows/**', 'CMakeLists.txt', 'Makefile', '**.h', '*.c', '**.cpp']
+    paths: ['.github/workflows/**', '**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.c', '**/*.cpp']
   pull_request:
     types: [opened, synchronize, edited, reopened, review_requested, ready_for_review]
-    paths: ['CMakeLists.txt', 'Makefile', '**.h', '*.c', '**.cpp']
+    paths: ['**/CMakeLists.txt', '**/Makefile', '**/*.h', '**/*.c', '**/*.cpp']
 
 env:
  BRANCH_NAME: ${{ github.head_ref || github.ref_name }}
@@ -62,7 +62,7 @@ jobs:
         id: cmake_test
         run: |
           cd build
-          ctest --output-on-failure
+          ctest --verbose
 
   ubuntu-latest-cmake-sanitizer:
     runs-on: ubuntu-latest
@@ -98,7 +98,7 @@ jobs:
         id: cmake_test
         run: |
           cd build
-          ctest --output-on-failure
+          ctest --verbose
 
   macOS-latest-make:
     runs-on: macos-latest
@@ -143,7 +143,7 @@ jobs:
         id: cmake_test
         run: |
           cd build
-          ctest --output-on-failure
+          ctest --verbose
 
   windows-latest-cmake:
     runs-on: windows-latest
@@ -176,16 +176,14 @@ jobs:
         if: ${{ matrix.build == 'avx512' }}
         continue-on-error: true
         run: |
-          cd build
-          Set-Content -Path .\avx512f.exe -Value ([Convert]::FromBase64String('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')) -AsByteStream
-          .\avx512f.exe && echo " AVX512F: YES" && ( echo HAS_AVX512F=1 >> $env:GITHUB_ENV ) || echo " AVX512F: NO"
+          echo "TODO: check avx512f"
 
       - name: Test
         id: cmake_test
         if: ${{ matrix.build != 'avx512' || env.HAS_AVX512F == '1' }} # Test AVX-512 only when possible
         run: |
           cd build
-          ctest -C Release --output-on-failure
+          ctest -C Release --verbose
 
       - name: Get commit hash
         id: commit

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/vendor/llama.cpp/.gitignore

@@ -5,6 +5,7 @@
 .vscode/
 .DS_Store
 
+.build/
 build/
 build-em/
 build-debug/
@@ -20,9 +21,14 @@ models/*
 /quantize
 /result
 /perplexity
+/embedding
 
 arm_neon.h
 compile_commands.json
 
 .envrc
 .direnv/
+
+.venv
+__pycache__
+.swiftpm

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/vendor/llama.cpp/CMakeLists.txt

@@ -124,8 +124,9 @@ if (LLAMA_ALL_WARNINGS)
             -Wall
             -Wextra
             -Wpedantic
-            -Wshadow
             -Wcast-qual
+            -Wdouble-promotion
+            -Wshadow
             -Wstrict-prototypes
             -Wpointer-arith
             -Wno-unused-function
@@ -135,6 +136,7 @@ if (LLAMA_ALL_WARNINGS)
             -Wextra
             -Wpedantic
             -Wcast-qual
+            -Wno-unused-function
         )
     else()
         # todo : msvc

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/vendor/llama.cpp/Makefile

@@ -35,6 +35,10 @@ CFLAGS   = -I.              -O3 -DNDEBUG -std=c11   -fPIC
 CXXFLAGS = -I. -I./examples -O3 -DNDEBUG -std=c++11 -fPIC
 LDFLAGS  =
 
+# warnings
+CFLAGS   += -Wall -Wextra -Wpedantic -Wcast-qual -Wdouble-promotion -Wshadow -Wstrict-prototypes -Wpointer-arith -Wno-unused-function
+CXXFLAGS += -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function
+
 # OS specific
 # TODO: support Windows
 ifeq ($(UNAME_S),Linux)
@@ -212,7 +216,7 @@ $(info I CC:       $(CCV))
 $(info I CXX:      $(CXXV))
 $(info )
 
-default: main quantize perplexity
+default: main quantize perplexity embedding
 
 #
 # Build library
@@ -228,7 +232,7 @@ common.o: examples/common.cpp examples/common.h
 	$(CXX) $(CXXFLAGS) -c examples/common.cpp -o common.o
 
 clean:
-	rm -vf *.o main quantize perplexity
+	rm -vf *.o main quantize perplexity embedding
 
 main: examples/main/main.cpp ggml.o llama.o common.o
 	$(CXX) $(CXXFLAGS) examples/main/main.cpp ggml.o llama.o common.o -o main $(LDFLAGS)
@@ -242,6 +246,9 @@ quantize: examples/quantize/quantize.cpp ggml.o llama.o
 perplexity: examples/perplexity/perplexity.cpp ggml.o llama.o common.o
 	$(CXX) $(CXXFLAGS) examples/perplexity/perplexity.cpp ggml.o llama.o common.o -o perplexity $(LDFLAGS)
 
+embedding: examples/embedding/embedding.cpp ggml.o llama.o common.o
+	$(CXX) $(CXXFLAGS) examples/embedding/embedding.cpp ggml.o llama.o common.o -o embedding $(LDFLAGS)
+
 #
 # Tests
 #

{llama_cpp_python-0.1.8 → llama_cpp_python-0.1.10}/vendor/llama.cpp/convert-pth-to-ggml.py

@@ -145,13 +145,11 @@ def main():
 
         print(f"Extracting only the vocab from '{fname_model}'\n")
 
-        model = torch.load(fname_model, map_location="cpu")
 
         with open(fname_out, "wb") as fout:
             write_header(fout, hparams, ftype)
             write_tokens(fout, tokenizer)
 
-        del model
 
         print(f"Done. Output file: {fname_out}\n")