Add starcoder support.

gpt4all-backend/CMakeLists.txt

@@ -129,6 +129,10 @@ foreach(BUILD_VARIANT IN LISTS BUILD_VARIANTS)
         add_library(bert-${BUILD_VARIANT} SHARED
             bert.cpp utils.h utils.cpp llmodel_shared.cpp llmodel_shared.h)
         prepare_target(bert llama-mainline)
+
+        add_library(starcoder-${BUILD_VARIANT} SHARED
+            starcoder.cpp utils.h utils.cpp llmodel_shared.cpp llmodel_shared.h)
+        prepare_target(starcoder llama-mainline)
     endif()
 endforeach()
 

gpt4all-backend/gptj.cpp

@@ -961,6 +961,11 @@ DLL_EXPORT const char *get_build_variant() {
 DLL_EXPORT bool magic_match(std::istream& f) {
     uint32_t magic = 0;
     f.read(reinterpret_cast<char*>(&magic), sizeof(magic));
+    gptj_hparams hparams;
+    f.read(reinterpret_cast<char*>(&hparams), sizeof(hparams));
+    if (!(hparams.n_vocab >= 50300 && hparams.n_vocab <= 50400)) {
+        return false; // not a gptj.
+    }
     return magic == 0x67676d6c;
 }
 

gpt4all-backend/starcoder_impl.h

@@ -0,0 +1,42 @@
+#ifndef STARCODER_H_I_KNOW_WHAT_I_AM_DOING_WHEN_INCLUDING_THIS_FILE
+#error This file is NOT meant to be included outside of starcoder.cpp. Doing so is DANGEROUS. Be sure to know what you are doing before proceeding to #define STARCODER_H_I_KNOW_WHAT_I_AM_DOING_WHEN_INCLUDING_THIS_FILE
+#endif
+#ifndef STARCODER_H
+#define STARCODER_H
+
+#include <string>
+#include <functional>
+#include <vector>
+#include <memory>
+#include "llmodel.h"
+
+struct StarcoderPrivate;
+class Starcoder : public LLModel {
+public:
+    Starcoder();
+    ~Starcoder();
+
+    bool supportsEmbedding() const override { return false; }
+    bool supportsCompletion() const override { return true; }
+    bool loadModel(const std::string &modelPath) override;
+    bool isModelLoaded() const override;
+    size_t requiredMem(const std::string &modelPath) override;
+    size_t stateSize() const override;
+    size_t saveState(uint8_t *dest) const override;
+    size_t restoreState(const uint8_t *src) override;
+    void setThreadCount(int32_t n_threads) override;
+    int32_t threadCount() const override;
+
+private:
+    std::unique_ptr<StarcoderPrivate> d_ptr;
+
+protected:
+    std::vector<Token> tokenize(PromptContext &, const std::string&) const override;
+    Token sampleToken(PromptContext &ctx) const override;
+    std::string tokenToString(Token) const override;
+    bool evalTokens(PromptContext &ctx, const std::vector<int32_t> &tokens) const override;
+    int32_t contextLength() const override;
+    const std::vector<Token>& endTokens() const override;
+};
+
+#endif // STARCODER_H

gpt4all-chat/CMakeLists.txt

@@ -208,6 +208,8 @@ install(TARGETS replit-mainline-metal DESTINATION lib COMPONENT ${COMPONENT_NAME
 endif()
 install(TARGETS bert-avxonly DESTINATION lib COMPONENT ${COMPONENT_NAME_MAIN})
 install(TARGETS bert-default DESTINATION lib COMPONENT ${COMPONENT_NAME_MAIN})
+install(TARGETS starcoder-avxonly DESTINATION lib COMPONENT ${COMPONENT_NAME_MAIN})
+install(TARGETS starcoder-default DESTINATION lib COMPONENT ${COMPONENT_NAME_MAIN})
 
 set(CPACK_GENERATOR "IFW")
 set(CPACK_VERBATIM_VARIABLES YES)

gpt4all-chat/chatllm.cpp

@@ -15,6 +15,7 @@
 #define LLAMA_INTERNAL_STATE_VERSION 0
 #define FALCON_INTERNAL_STATE_VERSION 0
 #define BERT_INTERNAL_STATE_VERSION 0
+#define STARCODER_INTERNAL_STATE_VERSION 0
 
 class LLModelStore {
 public:
@@ -266,6 +267,7 @@ bool ChatLLM::loadModel(const ModelInfo &modelInfo)
                     case 'R': m_llModelType = LLModelType::REPLIT_; break;
                     case 'F': m_llModelType = LLModelType::FALCON_; break;
                     case 'B': m_llModelType = LLModelType::BERT_; break;
+                    case 'S': m_llModelType = LLModelType::STARCODER_; break;
                     default:
                         {
                             delete std::exchange(m_llModelInfo.model, nullptr);
@@ -673,6 +675,7 @@ bool ChatLLM::serialize(QDataStream &stream, int version)
         case LLAMA_: stream << LLAMA_INTERNAL_STATE_VERSION; break;
         case FALCON_: stream << FALCON_INTERNAL_STATE_VERSION; break;
         case BERT_: stream << BERT_INTERNAL_STATE_VERSION; break;
+        case STARCODER_: stream << STARCODER_INTERNAL_STATE_VERSION; break;
         default: Q_UNREACHABLE();
         }
     }

gpt4all-chat/chatllm.h

@@ -16,7 +16,8 @@ enum LLModelType {
     CHATGPT_,
     REPLIT_,
     FALCON_,
-    BERT_
+    BERT_,
+    STARCODER_
 };
 
 struct LLModelInfo {

gpt4all-chat/cmake/deploy-qt-mac.cmake.in

@@ -8,6 +8,7 @@ file(GLOB MYLLAMALIBS ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_
 file(GLOB MYREPLITLIBS ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/lib/libreplit*)
 file(GLOB MYFALCONLLIBS ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/lib/libfalcon*)
 file(GLOB MYBERTLLIBS ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/lib/libbert*)
+file(GLOB MYSTARCODERLLIBS ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/lib/libstarcoder*)
 file(GLOB MYLLMODELLIBS ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/lib/libllmodel.*)
 file(COPY ${MYGPTJLIBS}
      DESTINATION ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/bin/gpt4all.app/Contents/Frameworks)
@@ -21,6 +22,10 @@ file(COPY ${MYFALCONLLIBS}
      DESTINATION ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/bin/gpt4all.app/Contents/Frameworks)
 file(COPY ${MYBERTLLIBS}
      DESTINATION ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/bin/gpt4all.app/Contents/Frameworks)
+file(COPY ${MYSTARCODERLLIBS}
+     DESTINATION ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/bin/gpt4all.app/Contents/Frameworks)
+file(COPY ${MYLLAMALIBS}
+     DESTINATION ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/bin/gpt4all.app/Contents/Frameworks)
 file(COPY ${MYLLMODELLIBS}
      DESTINATION ${CPACK_TEMPORARY_INSTALL_DIRECTORY}/packages/${COMPONENT_NAME_MAIN}/data/bin/gpt4all.app/Contents/Frameworks)
 file(COPY "${CMAKE_CURRENT_SOURCE_DIR}/icons/favicon.icns"