onnxruntime 0.9.4-aarch64-linux → 0.10.0-aarch64-linux

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9544eb4f8df1fd611ba42ce3b3687af10d226434202e42caac7e03dd3926162d
-  data.tar.gz: 9176e3659227243408ed5f5eab6508b719ccfbdddb37a2d173f6ebcb26ce490d
+  metadata.gz: 63c3880cb27601251c1e5f4ba8da03ba2a81aa3b7336205c0d1bfe85ecce3b54
+  data.tar.gz: '086e5ab987918620bfc17e5be5c65136a11ad0a46b2602c8266448ba4149701e'
 SHA512:
-  metadata.gz: 255f60ce409099e1ef8b4bbf820360903785b0690c26fd540511c839185ba8abfde9eee1a67ffb3ff1c38d1a36c1910b9b05626da98eb1b6119eddc1b18036ee
-  data.tar.gz: 177f73d9d97856c34c9f5dccb19a9428776421a54edf2b65349411eae556783ce24d5375f06b1a6ff12449c5c894ce12f0ffd566880e0263ae9ea67aa474ac3b
+  metadata.gz: 37f07766244ec17bb4cd0ff8bc2711118fb1b4eae08f933b7428a923df02128504d5639306a1e6b5a425e77e824f8e2c4a8aca9544086f2aab0b0f2e795e225c
+  data.tar.gz: d46924ba93cae57f6f0f3bda864dad8f59cd9d1d119f669d8e40afe038c43201f3712a33bb8b903e537091b99b5e241f35d0d831a69dec323bee4262e6203da4

data/CHANGELOG.md

@@ -1,3 +1,8 @@
+## 0.10.0 (2025-05-11)
+
+- Updated ONNX Runtime to 1.22.0
+- Dropped support for Ruby < 3.2
+
 ## 0.9.4 (2025-03-08)
 
 - Updated ONNX Runtime to 1.21.0

data/LICENSE.txt

@@ -1,7 +1,7 @@
 MIT License
 
 Copyright (c) Microsoft Corporation
-Copyright (c) 2019-2024 Andrew Kane
+Copyright (c) 2019-2025 Andrew Kane
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

data/lib/onnxruntime/ffi.rb

@@ -147,7 +147,7 @@ module OnnxRuntime
         :FillStringTensorElement, callback(%i[pointer string size_t], :pointer),
         :AddSessionConfigEntry, callback(%i[pointer string string], :pointer),
         :CreateAllocator, callback(%i[], :pointer),
-        :ReleaseAllocator, callback(%i[], :pointer),
+        :ReleaseAllocator, callback(%i[pointer], :pointer),
         :RunWithBinding, callback(%i[], :pointer),
         :CreateIoBinding, callback(%i[], :pointer),
         :ReleaseIoBinding, callback(%i[], :pointer),
@@ -249,7 +249,7 @@ module OnnxRuntime
     def self.api
       @api ||= begin
         api = self.OrtGetApiBase[:GetApi].call(ORT_API_VERSION)
-        api = Api.by_ref.from_native(api, nil) if RUBY_PLATFORM == "java"
+        api = Api.by_ref.from_native(api, nil) if RUBY_ENGINE == "jruby"
         api
       end
     end

data/lib/onnxruntime/inference_session.rb

@@ -6,52 +6,54 @@ module OnnxRuntime
       # session options
       session_options = ::FFI::MemoryPointer.new(:pointer)
       check_status api[:CreateSessionOptions].call(session_options)
+      session_options = ::FFI::AutoPointer.new(session_options.read_pointer, api[:ReleaseSessionOptions])
+
       if enable_cpu_mem_arena
-        check_status api[:EnableCpuMemArena].call(session_options.read_pointer)
+        check_status api[:EnableCpuMemArena].call(session_options)
       else
-        check_status api[:DisableCpuMemArena].call(session_options.read_pointer)
+        check_status api[:DisableCpuMemArena].call(session_options)
       end
       if enable_mem_pattern
-        check_status api[:EnableMemPattern].call(session_options.read_pointer)
+        check_status api[:EnableMemPattern].call(session_options)
       else
-        check_status api[:DisableMemPattern].call(session_options.read_pointer)
+        check_status api[:DisableMemPattern].call(session_options)
       end
       if enable_profiling
-        check_status api[:EnableProfiling].call(session_options.read_pointer, ort_string(profile_file_prefix || "onnxruntime_profile_"))
+        check_status api[:EnableProfiling].call(session_options, ort_string(profile_file_prefix || "onnxruntime_profile_"))
       else
-        check_status api[:DisableProfiling].call(session_options.read_pointer)
+        check_status api[:DisableProfiling].call(session_options)
       end
       if execution_mode
         execution_modes = {sequential: 0, parallel: 1}
         mode = execution_modes[execution_mode]
         raise ArgumentError, "Invalid execution mode" unless mode
-        check_status api[:SetSessionExecutionMode].call(session_options.read_pointer, mode)
+        check_status api[:SetSessionExecutionMode].call(session_options, mode)
       end
       if free_dimension_overrides_by_denotation
         free_dimension_overrides_by_denotation.each do |k, v|
-          check_status api[:AddFreeDimensionOverride].call(session_options.read_pointer, k.to_s, v)
+          check_status api[:AddFreeDimensionOverride].call(session_options, k.to_s, v)
         end
       end
       if free_dimension_overrides_by_name
         free_dimension_overrides_by_name.each do |k, v|
-          check_status api[:AddFreeDimensionOverrideByName].call(session_options.read_pointer, k.to_s, v)
+          check_status api[:AddFreeDimensionOverrideByName].call(session_options, k.to_s, v)
         end
       end
       if graph_optimization_level
         optimization_levels = {none: 0, basic: 1, extended: 2, all: 99}
         level = optimization_levels[graph_optimization_level]
         raise ArgumentError, "Invalid graph optimization level" unless level
-        check_status api[:SetSessionGraphOptimizationLevel].call(session_options.read_pointer, level)
+        check_status api[:SetSessionGraphOptimizationLevel].call(session_options, level)
       end
-      check_status api[:SetInterOpNumThreads].call(session_options.read_pointer, inter_op_num_threads) if inter_op_num_threads
-      check_status api[:SetIntraOpNumThreads].call(session_options.read_pointer, intra_op_num_threads) if intra_op_num_threads
-      check_status api[:SetSessionLogSeverityLevel].call(session_options.read_pointer, log_severity_level) if log_severity_level
-      check_status api[:SetSessionLogVerbosityLevel].call(session_options.read_pointer, log_verbosity_level) if log_verbosity_level
-      check_status api[:SetSessionLogId].call(session_options.read_pointer, logid) if logid
-      check_status api[:SetOptimizedModelFilePath].call(session_options.read_pointer, ort_string(optimized_model_filepath)) if optimized_model_filepath
+      check_status api[:SetInterOpNumThreads].call(session_options, inter_op_num_threads) if inter_op_num_threads
+      check_status api[:SetIntraOpNumThreads].call(session_options, intra_op_num_threads) if intra_op_num_threads
+      check_status api[:SetSessionLogSeverityLevel].call(session_options, log_severity_level) if log_severity_level
+      check_status api[:SetSessionLogVerbosityLevel].call(session_options, log_verbosity_level) if log_verbosity_level
+      check_status api[:SetSessionLogId].call(session_options, logid) if logid
+      check_status api[:SetOptimizedModelFilePath].call(session_options, ort_string(optimized_model_filepath)) if optimized_model_filepath
       if session_config_entries
         session_config_entries.each do |k, v|
-          check_status api[:AddSessionConfigEntry].call(session_options.read_pointer, k.to_s, v.to_s)
+          check_status api[:AddSessionConfigEntry].call(session_options, k.to_s, v.to_s)
         end
       end
       providers.each do |provider|
@@ -64,15 +66,15 @@ module OnnxRuntime
         when "CUDAExecutionProvider"
           cuda_options = ::FFI::MemoryPointer.new(:pointer)
           check_status api[:CreateCUDAProviderOptions].call(cuda_options)
-          check_status api[:SessionOptionsAppendExecutionProvider_CUDA_V2].call(session_options.read_pointer, cuda_options.read_pointer)
-          release :CUDAProviderOptions, cuda_options
+          cuda_options = ::FFI::AutoPointer.new(cuda_options.read_pointer, api[:ReleaseCUDAProviderOptions])
+          check_status api[:SessionOptionsAppendExecutionProvider_CUDA_V2].call(session_options, cuda_options)
         when "CoreMLExecutionProvider"
           unless FFI.respond_to?(:OrtSessionOptionsAppendExecutionProvider_CoreML)
             raise ArgumentError, "Provider not available: #{provider}"
           end
 
           coreml_flags = 0
-          check_status FFI.OrtSessionOptionsAppendExecutionProvider_CoreML(session_options.read_pointer, coreml_flags)
+          check_status FFI.OrtSessionOptionsAppendExecutionProvider_CoreML(session_options, coreml_flags)
         when "CPUExecutionProvider"
           break
         else
@@ -81,13 +83,9 @@ module OnnxRuntime
       end
 
       @session = load_session(path_or_bytes, session_options)
-      ObjectSpace.define_finalizer(@session, self.class.finalize(read_pointer.to_i))
-
       @allocator = Utils.allocator
       @inputs = load_inputs
       @outputs = load_outputs
-    ensure
-      release :SessionOptions, session_options
     end
 
     def run(output_names, input_feed, log_severity_level: nil, log_verbosity_level: nil, logid: nil, terminate: nil, output_type: :ruby)
@@ -119,51 +117,57 @@ module OnnxRuntime
       # run options
       run_options = ::FFI::MemoryPointer.new(:pointer)
       check_status api[:CreateRunOptions].call(run_options)
-      check_status api[:RunOptionsSetRunLogSeverityLevel].call(run_options.read_pointer, log_severity_level) if log_severity_level
-      check_status api[:RunOptionsSetRunLogVerbosityLevel].call(run_options.read_pointer, log_verbosity_level) if log_verbosity_level
-      check_status api[:RunOptionsSetRunTag].call(run_options.read_pointer, logid) if logid
-      check_status api[:RunOptionsSetTerminate].call(run_options.read_pointer) if terminate
+      run_options = ::FFI::AutoPointer.new(run_options.read_pointer, api[:ReleaseRunOptions])
 
-      check_status api[:Run].call(read_pointer, run_options.read_pointer, input_node_names, input_tensor, input_feed.size, output_node_names, output_names.size, output_tensor)
+      check_status api[:RunOptionsSetRunLogSeverityLevel].call(run_options, log_severity_level) if log_severity_level
+      check_status api[:RunOptionsSetRunLogVerbosityLevel].call(run_options, log_verbosity_level) if log_verbosity_level
+      check_status api[:RunOptionsSetRunTag].call(run_options, logid) if logid
+      check_status api[:RunOptionsSetTerminate].call(run_options) if terminate
 
-      output_names.size.times.map { |i| OrtValue.new(output_tensor[i]) }
-    ensure
-      release :RunOptions, run_options
+      check_status api[:Run].call(@session, run_options, input_node_names, input_tensor, input_feed.size, output_node_names, output_names.size, output_tensor)
+
+      output_names.size.times.map { |i| OrtValue.new(output_tensor[i].read_pointer) }
     end
 
     def modelmeta
+      metadata = ::FFI::MemoryPointer.new(:pointer)
+      check_status api[:SessionGetModelMetadata].call(@session, metadata)
+      metadata = ::FFI::AutoPointer.new(metadata.read_pointer, api[:ReleaseModelMetadata])
+
       keys = ::FFI::MemoryPointer.new(:pointer)
       num_keys = ::FFI::MemoryPointer.new(:int64_t)
-      description = ::FFI::MemoryPointer.new(:string)
-      domain = ::FFI::MemoryPointer.new(:string)
-      graph_name = ::FFI::MemoryPointer.new(:string)
-      graph_description = ::FFI::MemoryPointer.new(:string)
-      producer_name = ::FFI::MemoryPointer.new(:string)
-      version = ::FFI::MemoryPointer.new(:int64_t)
-
-      metadata = ::FFI::MemoryPointer.new(:pointer)
-      check_status api[:SessionGetModelMetadata].call(read_pointer, metadata)
+      check_status api[:ModelMetadataGetCustomMetadataMapKeys].call(metadata, @allocator, keys, num_keys)
+      keys = keys.read_pointer
 
       custom_metadata_map = {}
-      check_status api[:ModelMetadataGetCustomMetadataMapKeys].call(metadata.read_pointer, @allocator.read_pointer, keys, num_keys)
       num_keys.read(:int64_t).times do |i|
-        key_ptr = keys.read_pointer[i * ::FFI::Pointer.size]
-        key = key_ptr.read_pointer.read_string
-        value = ::FFI::MemoryPointer.new(:string)
-        check_status api[:ModelMetadataLookupCustomMetadataMap].call(metadata.read_pointer, @allocator.read_pointer, key, value)
+        key_ptr = keys.get_pointer(i * ::FFI::Pointer.size)
+        key = key_ptr.read_string
+        value = ::FFI::MemoryPointer.new(:pointer)
+        check_status api[:ModelMetadataLookupCustomMetadataMap].call(metadata, @allocator, key, value)
         custom_metadata_map[key] = value.read_pointer.read_string
 
         allocator_free key_ptr
-        allocator_free value
+        allocator_free value.read_pointer
       end
-      allocator_free keys
 
-      check_status api[:ModelMetadataGetDescription].call(metadata.read_pointer, @allocator.read_pointer, description)
-      check_status api[:ModelMetadataGetDomain].call(metadata.read_pointer, @allocator.read_pointer, domain)
-      check_status api[:ModelMetadataGetGraphName].call(metadata.read_pointer, @allocator.read_pointer, graph_name)
-      check_status api[:ModelMetadataGetGraphDescription].call(metadata.read_pointer, @allocator.read_pointer, graph_description)
-      check_status api[:ModelMetadataGetProducerName].call(metadata.read_pointer, @allocator.read_pointer, producer_name)
-      check_status api[:ModelMetadataGetVersion].call(metadata.read_pointer, version)
+      description = ::FFI::MemoryPointer.new(:pointer)
+      check_status api[:ModelMetadataGetDescription].call(metadata, @allocator, description)
+
+      domain = ::FFI::MemoryPointer.new(:pointer)
+      check_status api[:ModelMetadataGetDomain].call(metadata, @allocator, domain)
+
+      graph_name = ::FFI::MemoryPointer.new(:pointer)
+      check_status api[:ModelMetadataGetGraphName].call(metadata, @allocator, graph_name)
+
+      graph_description = ::FFI::MemoryPointer.new(:pointer)
+      check_status api[:ModelMetadataGetGraphDescription].call(metadata, @allocator, graph_description)
+
+      producer_name = ::FFI::MemoryPointer.new(:pointer)
+      check_status api[:ModelMetadataGetProducerName].call(metadata, @allocator, producer_name)
+
+      version = ::FFI::MemoryPointer.new(:int64_t)
+      check_status api[:ModelMetadataGetVersion].call(metadata, version)
 
       {
         custom_metadata_map: custom_metadata_map,
@@ -175,19 +179,23 @@ module OnnxRuntime
         version: version.read(:int64_t)
       }
     ensure
-      release :ModelMetadata, metadata
-      allocator_free description
-      allocator_free domain
-      allocator_free graph_name
-      allocator_free graph_description
-      allocator_free producer_name
+      allocator_free keys
+      allocator_free description.read_pointer
+      allocator_free domain.read_pointer
+      allocator_free graph_name.read_pointer
+      allocator_free graph_description.read_pointer
+      allocator_free producer_name.read_pointer
     end
 
     # return value has double underscore like Python
     def end_profiling
-      out = ::FFI::MemoryPointer.new(:string)
-      check_status api[:SessionEndProfiling].call(read_pointer, @allocator.read_pointer, out)
-      out.read_pointer.read_string
+      out = ::FFI::MemoryPointer.new(:pointer)
+      check_status api[:SessionEndProfiling].call(@session, @allocator, out)
+      begin
+        out.read_pointer.read_string
+      ensure
+        allocator_free out.read_pointer
+      end
     end
 
     # no way to set providers with C API yet
@@ -197,10 +205,7 @@ module OnnxRuntime
       length_ptr = ::FFI::MemoryPointer.new(:int)
       check_status api[:GetAvailableProviders].call(out_ptr, length_ptr)
       length = length_ptr.read_int
-      providers = []
-      length.times do |i|
-        providers << out_ptr.read_pointer[i * ::FFI::Pointer.size].read_pointer.read_string
-      end
+      providers = out_ptr.read_pointer.read_array_of_pointer(length).map(&:read_string)
       api[:ReleaseAvailableProviders].call(out_ptr.read_pointer, length)
       providers
     end
@@ -208,7 +213,6 @@ module OnnxRuntime
     private
 
     def load_session(path_or_bytes, session_options)
-      session = ::FFI::MemoryPointer.new(:pointer)
       from_memory =
         if path_or_bytes.respond_to?(:read)
           path_or_bytes = path_or_bytes.read
@@ -218,44 +222,49 @@ module OnnxRuntime
           false
         end
 
+      session = ::FFI::MemoryPointer.new(:pointer)
       if from_memory
-        check_status api[:CreateSessionFromArray].call(env.read_pointer, path_or_bytes, path_or_bytes.bytesize, session_options.read_pointer, session)
+        check_status api[:CreateSessionFromArray].call(env, path_or_bytes, path_or_bytes.bytesize, session_options, session)
       else
-        check_status api[:CreateSession].call(env.read_pointer, ort_string(path_or_bytes), session_options.read_pointer, session)
+        check_status api[:CreateSession].call(env, ort_string(path_or_bytes), session_options, session)
       end
-      session
+      ::FFI::AutoPointer.new(session.read_pointer, api[:ReleaseSession])
     end
 
     def load_inputs
-      inputs = []
       num_input_nodes = ::FFI::MemoryPointer.new(:size_t)
-      check_status api[:SessionGetInputCount].call(read_pointer, num_input_nodes)
-      num_input_nodes.read(:size_t).times do |i|
-        name_ptr = ::FFI::MemoryPointer.new(:string)
-        check_status api[:SessionGetInputName].call(read_pointer, i, @allocator.read_pointer, name_ptr)
-        # freed in node_info
+      check_status api[:SessionGetInputCount].call(@session, num_input_nodes)
+
+      num_input_nodes.read(:size_t).times.map do |i|
+        name_ptr = ::FFI::MemoryPointer.new(:pointer)
+        check_status api[:SessionGetInputName].call(@session, i, @allocator, name_ptr)
+        name_str = name_ptr.read_pointer.read_string
+        allocator_free name_ptr.read_pointer
+
         typeinfo = ::FFI::MemoryPointer.new(:pointer)
-        check_status api[:SessionGetInputTypeInfo].call(read_pointer, i, typeinfo)
-        inputs << {name: name_ptr.read_pointer.read_string}.merge(Utils.node_info(typeinfo))
-        allocator_free name_ptr
+        check_status api[:SessionGetInputTypeInfo].call(@session, i, typeinfo)
+        typeinfo = ::FFI::AutoPointer.new(typeinfo.read_pointer, api[:ReleaseTypeInfo])
+
+        {name: name_str}.merge(Utils.node_info(typeinfo))
       end
-      inputs
     end
 
     def load_outputs
-      outputs = []
       num_output_nodes = ::FFI::MemoryPointer.new(:size_t)
-      check_status api[:SessionGetOutputCount].call(read_pointer, num_output_nodes)
-      num_output_nodes.read(:size_t).times do |i|
-        name_ptr = ::FFI::MemoryPointer.new(:string)
-        check_status api[:SessionGetOutputName].call(read_pointer, i, @allocator.read_pointer, name_ptr)
-        # freed in node_info
+      check_status api[:SessionGetOutputCount].call(@session, num_output_nodes)
+
+      num_output_nodes.read(:size_t).times.map do |i|
+        name_ptr = ::FFI::MemoryPointer.new(:pointer)
+        check_status api[:SessionGetOutputName].call(@session, i, @allocator, name_ptr)
+        name_str = name_ptr.read_pointer.read_string
+        allocator_free name_ptr.read_pointer
+
         typeinfo = ::FFI::MemoryPointer.new(:pointer)
-        check_status api[:SessionGetOutputTypeInfo].call(read_pointer, i, typeinfo)
-        outputs << {name: name_ptr.read_pointer.read_string}.merge(Utils.node_info(typeinfo))
-        allocator_free name_ptr
+        check_status api[:SessionGetOutputTypeInfo].call(@session, i, typeinfo)
+        typeinfo = ::FFI::AutoPointer.new(typeinfo.read_pointer, api[:ReleaseTypeInfo])
+
+        {name: name_str}.merge(Utils.node_info(typeinfo))
       end
-      outputs
     end
 
     def create_input_tensor(input_feed)
@@ -285,10 +294,6 @@ module OnnxRuntime
       ptr
     end
 
-    def read_pointer
-      @session.read_pointer
-    end
-
     def check_status(status)
       Utils.check_status(status)
     end
@@ -301,27 +306,14 @@ module OnnxRuntime
       self.class.api
     end
 
-    def release(*args)
-      self.class.release(*args)
-    end
-
     def allocator_free(ptr)
-      api[:AllocatorFree].call(@allocator.read_pointer, ptr.read_pointer)
+      api[:AllocatorFree].call(@allocator, ptr)
     end
 
     def self.api
       FFI.api
     end
 
-    def self.release(type, pointer)
-      Utils.release(type, pointer)
-    end
-
-    def self.finalize(addr)
-      # must use proc instead of stabby lambda
-      proc { api[:ReleaseSession].call(::FFI::Pointer.new(:pointer, addr)) }
-    end
-
     # wide string on Windows
     # char string on Linux
     # see ORTCHAR_T in onnxruntime_c_api.h
@@ -333,7 +325,8 @@ module OnnxRuntime
         raise Error, "Expected mbstowcs to return #{str.size}, got #{ret}" if ret != str.size
         dest
       else
-        str
+        # prevent frozen string literal warnings
+        +str
       end
     end
 
@@ -343,7 +336,7 @@ module OnnxRuntime
         @@env ||= begin
           env = ::FFI::MemoryPointer.new(:pointer)
           check_status api[:CreateEnv].call(3, "Default", env)
-          at_exit { release :Env, env }
+          env = ::FFI::AutoPointer.new(env.read_pointer, api[:ReleaseEnv])
           # disable telemetry
           # https://github.com/microsoft/onnxruntime/blob/master/docs/Privacy.md
           check_status api[:DisableTelemetryEvents].call(env)

data/lib/onnxruntime/ort_value.rb

@@ -1,9 +1,8 @@
 module OnnxRuntime
   class OrtValue
     def initialize(ptr, ref = nil)
-      @ptr = ptr.read_pointer
+      @ptr = ::FFI::AutoPointer.new(ptr, FFI.api[:ReleaseValue])
       @ref = ref # keep reference to data
-      ObjectSpace.define_finalizer(@ptr, self.class.finalize(@ptr.to_i))
     end
 
     def self.from_numo(numo_obj)
@@ -27,14 +26,14 @@ module OnnxRuntime
       if element_type == :string
         # keep reference to _str_ptrs until FillStringTensor call
         input_tensor_values, _str_ptrs = create_input_strings(input)
-        Utils.check_status FFI.api[:CreateTensorAsOrtValue].call(Utils.allocator.read_pointer, input_node_dims, shape.size, type_enum, ptr)
+        Utils.check_status FFI.api[:CreateTensorAsOrtValue].call(Utils.allocator, input_node_dims, shape.size, type_enum, ptr)
         Utils.check_status FFI.api[:FillStringTensor].call(ptr.read_pointer, input_tensor_values, input_tensor_values.size / input_tensor_values.type_size)
       else
         input_tensor_values = create_input_data(input, element_type)
-        Utils.check_status FFI.api[:CreateTensorWithDataAsOrtValue].call(allocator_info.read_pointer, input_tensor_values, input_tensor_values.size, input_node_dims, shape.size, type_enum, ptr)
+        Utils.check_status FFI.api[:CreateTensorWithDataAsOrtValue].call(allocator_info, input_tensor_values, input_tensor_values.size, input_node_dims, shape.size, type_enum, ptr)
       end
 
-      new(ptr, input_tensor_values)
+      new(ptr.read_pointer, input_tensor_values)
     end
 
     def self.from_shape_and_type(shape, element_type)
@@ -45,9 +44,9 @@ module OnnxRuntime
       input_node_dims.write_array_of_int64(shape)
 
       ptr = ::FFI::MemoryPointer.new(:pointer)
-      Utils.check_status FFI.api[:CreateTensorAsOrtValue].call(Utils.allocator.read_pointer, input_node_dims, shape.size, type_enum, ptr)
+      Utils.check_status FFI.api[:CreateTensorAsOrtValue].call(Utils.allocator, input_node_dims, shape.size, type_enum, ptr)
 
-      new(ptr)
+      new(ptr.read_pointer)
     end
 
     def self.create_input_data(input, tensor_type)
@@ -88,6 +87,7 @@ module OnnxRuntime
       @data_type ||= begin
         typeinfo = ::FFI::MemoryPointer.new(:pointer)
         Utils.check_status FFI.api[:GetTypeInfo].call(@ptr, typeinfo)
+        typeinfo = ::FFI::AutoPointer.new(typeinfo.read_pointer, FFI.api[:ReleaseTypeInfo])
         Utils.node_info(typeinfo)[:type]
       end
     end
@@ -134,13 +134,10 @@ module OnnxRuntime
 
     def type_and_shape_info
       @type_and_shape_info ||= begin
-        begin
-          typeinfo = ::FFI::MemoryPointer.new(:pointer)
-          Utils.check_status FFI.api[:GetTensorTypeAndShape].call(@ptr, typeinfo)
-          Utils.tensor_type_and_shape(typeinfo)
-        ensure
-          Utils.release :TensorTypeAndShapeInfo, typeinfo
-        end
+        typeinfo = ::FFI::MemoryPointer.new(:pointer)
+        Utils.check_status FFI.api[:GetTensorTypeAndShape].call(@ptr, typeinfo)
+        typeinfo = ::FFI::AutoPointer.new(typeinfo.read_pointer, FFI.api[:ReleaseTensorTypeAndShapeInfo])
+        Utils.tensor_type_and_shape(typeinfo)
       end
     end
 
@@ -153,6 +150,7 @@ module OnnxRuntime
       when :tensor
         typeinfo = ::FFI::MemoryPointer.new(:pointer)
         Utils.check_status FFI.api[:GetTensorTypeAndShape].call(out_ptr, typeinfo)
+        typeinfo = ::FFI::AutoPointer.new(typeinfo.read_pointer, FFI.api[:ReleaseTensorTypeAndShapeInfo])
 
         type, shape = Utils.tensor_type_and_shape(typeinfo)
 
@@ -160,11 +158,9 @@ module OnnxRuntime
         Utils.check_status FFI.api[:GetTensorMutableData].call(out_ptr, tensor_data)
 
         out_size = ::FFI::MemoryPointer.new(:size_t)
-        Utils.check_status FFI.api[:GetTensorShapeElementCount].call(typeinfo.read_pointer, out_size)
+        Utils.check_status FFI.api[:GetTensorShapeElementCount].call(typeinfo, out_size)
         output_tensor_size = out_size.read(:size_t)
 
-        Utils.release :TensorTypeAndShapeInfo, typeinfo
-
         # TODO support more types
         type = FFI::TensorElementDataType[type]
 
@@ -203,20 +199,22 @@ module OnnxRuntime
 
         out.read(:size_t).times.map do |i|
           seq = ::FFI::MemoryPointer.new(:pointer)
-          Utils.check_status FFI.api[:GetValue].call(out_ptr, i, Utils.allocator.read_pointer, seq)
+          Utils.check_status FFI.api[:GetValue].call(out_ptr, i, Utils.allocator, seq)
           create_from_onnx_value(seq.read_pointer, output_type)
         end
       when :map
-        type_shape = ::FFI::MemoryPointer.new(:pointer)
         map_keys = ::FFI::MemoryPointer.new(:pointer)
+        Utils.check_status FFI.api[:GetValue].call(out_ptr, 0, Utils.allocator, map_keys)
+
         map_values = ::FFI::MemoryPointer.new(:pointer)
-        elem_type = ::FFI::MemoryPointer.new(:int)
+        Utils.check_status FFI.api[:GetValue].call(out_ptr, 1, Utils.allocator, map_values)
 
-        Utils.check_status FFI.api[:GetValue].call(out_ptr, 0, Utils.allocator.read_pointer, map_keys)
-        Utils.check_status FFI.api[:GetValue].call(out_ptr, 1, Utils.allocator.read_pointer, map_values)
+        type_shape = ::FFI::MemoryPointer.new(:pointer)
         Utils.check_status FFI.api[:GetTensorTypeAndShape].call(map_keys.read_pointer, type_shape)
-        Utils.check_status FFI.api[:GetTensorElementType].call(type_shape.read_pointer, elem_type)
-        Utils.release :TensorTypeAndShapeInfo, type_shape
+        type_shape = ::FFI::AutoPointer.new(type_shape.read_pointer, FFI.api[:ReleaseTensorTypeAndShapeInfo])
+
+        elem_type = ::FFI::MemoryPointer.new(:int)
+        Utils.check_status FFI.api[:GetTensorElementType].call(type_shape, elem_type)
 
         # TODO support more types
         elem_type = FFI::TensorElementDataType[elem_type.read_int]
@@ -255,23 +253,17 @@ module OnnxRuntime
     end
 
     def reshape(arr, dims)
-      arr = arr.flatten
       dims[1..-1].reverse_each do |dim|
         arr = arr.each_slice(dim)
       end
       arr.to_a
     end
 
-    def self.finalize(addr)
-      # must use proc instead of stabby lambda
-      proc { FFI.api[:ReleaseValue].call(::FFI::Pointer.new(:pointer, addr)) }
-    end
-
     def self.allocator_info
       @allocator_info ||= begin
         allocator_info = ::FFI::MemoryPointer.new(:pointer)
         Utils.check_status FFI.api[:CreateCpuMemoryInfo].call(1, 0, allocator_info)
-        allocator_info
+        ::FFI::AutoPointer.new(allocator_info.read_pointer, FFI.api[:ReleaseMemoryInfo])
       end
     end
   end

data/lib/onnxruntime/utils.rb

@@ -17,28 +17,24 @@ module OnnxRuntime
       FFI.api
     end
 
-    def self.release(type, pointer)
-      FFI.api[:"Release#{type}"].call(pointer.read_pointer) if pointer && !pointer.null?
-    end
-
     def self.unsupported_type(name, type)
       raise Error, "Unsupported #{name} type: #{type}"
     end
 
     def self.tensor_type_and_shape(tensor_info)
       type = ::FFI::MemoryPointer.new(:int)
-      check_status api[:GetTensorElementType].call(tensor_info.read_pointer, type)
+      check_status api[:GetTensorElementType].call(tensor_info, type)
 
       num_dims_ptr = ::FFI::MemoryPointer.new(:size_t)
-      check_status api[:GetDimensionsCount].call(tensor_info.read_pointer, num_dims_ptr)
+      check_status api[:GetDimensionsCount].call(tensor_info, num_dims_ptr)
       num_dims = num_dims_ptr.read(:size_t)
 
       node_dims = ::FFI::MemoryPointer.new(:int64, num_dims)
-      check_status api[:GetDimensions].call(tensor_info.read_pointer, node_dims, num_dims)
+      check_status api[:GetDimensions].call(tensor_info, node_dims, num_dims)
       dims = node_dims.read_array_of_int64(num_dims)
 
       symbolic_dims = ::FFI::MemoryPointer.new(:pointer, num_dims)
-      check_status api[:GetSymbolicDimensions].call(tensor_info.read_pointer, symbolic_dims, num_dims)
+      check_status api[:GetSymbolicDimensions].call(tensor_info, symbolic_dims, num_dims)
       named_dims = num_dims.times.map { |i| symbolic_dims[i].read_pointer.read_string }
       dims = named_dims.zip(dims).map { |n, d| n.empty? ? d : n }
 
@@ -47,25 +43,28 @@ module OnnxRuntime
 
     def self.node_info(typeinfo)
       onnx_type = ::FFI::MemoryPointer.new(:int)
-      check_status api[:GetOnnxTypeFromTypeInfo].call(typeinfo.read_pointer, onnx_type)
+      check_status api[:GetOnnxTypeFromTypeInfo].call(typeinfo, onnx_type)
 
       type = FFI::OnnxType[onnx_type.read_int]
       case type
       when :tensor
-        tensor_info = ::FFI::MemoryPointer.new(:pointer)
         # don't free tensor_info
-        check_status api[:CastTypeInfoToTensorInfo].call(typeinfo.read_pointer, tensor_info)
+        tensor_info = ::FFI::MemoryPointer.new(:pointer)
+        check_status api[:CastTypeInfoToTensorInfo].call(typeinfo, tensor_info)
+        type, shape = Utils.tensor_type_and_shape(tensor_info.read_pointer)
 
-        type, shape = Utils.tensor_type_and_shape(tensor_info)
         {
           type: "tensor(#{FFI::TensorElementDataType[type]})",
           shape: shape
         }
       when :sequence
+        # don't free sequence_info
         sequence_type_info = ::FFI::MemoryPointer.new(:pointer)
-        check_status api[:CastTypeInfoToSequenceTypeInfo].call(typeinfo.read_pointer, sequence_type_info)
+        check_status api[:CastTypeInfoToSequenceTypeInfo].call(typeinfo, sequence_type_info)
+
         nested_type_info = ::FFI::MemoryPointer.new(:pointer)
         check_status api[:GetSequenceElementType].call(sequence_type_info.read_pointer, nested_type_info)
+        nested_type_info = ::FFI::AutoPointer.new(nested_type_info.read_pointer, api[:ReleaseTypeInfo])
         v = node_info(nested_type_info)[:type]
 
         {
@@ -73,8 +72,9 @@ module OnnxRuntime
           shape: []
         }
       when :map
+        # don't free map_type_info
         map_type_info = ::FFI::MemoryPointer.new(:pointer)
-        check_status api[:CastTypeInfoToMapTypeInfo].call(typeinfo.read_pointer, map_type_info)
+        check_status api[:CastTypeInfoToMapTypeInfo].call(typeinfo, map_type_info)
 
         # key
         key_type = ::FFI::MemoryPointer.new(:int)
@@ -84,6 +84,7 @@ module OnnxRuntime
         # value
         value_type_info = ::FFI::MemoryPointer.new(:pointer)
         check_status api[:GetMapValueType].call(map_type_info.read_pointer, value_type_info)
+        value_type_info = ::FFI::AutoPointer.new(value_type_info.read_pointer, api[:ReleaseTypeInfo])
         v = node_info(value_type_info)[:type]
 
         {
@@ -93,8 +94,6 @@ module OnnxRuntime
       else
         Utils.unsupported_type("ONNX", type)
       end
-    ensure
-      release :TypeInfo, typeinfo
     end
 
     def self.numo_array?(obj)
@@ -135,7 +134,7 @@ module OnnxRuntime
       @allocator ||= begin
         allocator = ::FFI::MemoryPointer.new(:pointer)
         check_status api[:GetAllocatorWithDefaultOptions].call(allocator)
-        allocator
+        allocator.read_pointer # do not free default allocator
       end
     end
   end

data/lib/onnxruntime/version.rb

@@ -1,3 +1,3 @@
 module OnnxRuntime
-  VERSION = "0.9.4"
+  VERSION = "0.10.0"
 end

data/vendor/ThirdPartyNotices.txt

@@ -6080,3 +6080,77 @@ https://dawn.googlesource.com/dawn
    CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
    OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
    OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+_____
+
+KleidiAI
+
+https://gitlab.arm.com/kleidi/kleidiai
+
+Apache License
+Version 2.0, January 2004
+http://www.apache.org/licenses/
+
+TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+1. Definitions.
+
+"License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document.
+
+"Licensor" shall mean the copyright owner or entity authorized by the copyright owner that is granting the License.
+
+"Legal Entity" shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity.
+
+"You" (or "Your") shall mean an individual or Legal Entity exercising permissions granted by this License.
+
+"Source" form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files.
+
+"Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types.
+
+"Work" shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below).
+
+"Derivative Works" shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof.
+
+"Contribution" shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as "Not a Contribution."
+
+"Contributor" shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by Licensor and subsequently incorporated within the Work.
+
+2. Grant of Copyright License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form.
+
+3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed.
+
+4. Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions:
+
+     (a) You must give any other recipients of the Work or Derivative Works a copy of this License; and
+
+     (b) You must cause any modified files to carry prominent notices stating that You changed the files; and
+
+     (c) You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and
+
+     (d) If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding those notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License.
+
+     You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License.
+
+5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions.
+
+6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file.
+
+7. Disclaimer of Warranty. Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License.
+
+8. Limitation of Liability. In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages.
+
+9. Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability.
+
+END OF TERMS AND CONDITIONS
+
+Copyright 2024-2025 Arm Limited and/or its affiliates <open-source-office@arm.com>
+
+Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
+
+1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
+
+2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
+
+3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: onnxruntime
 version: !ruby/object:Gem::Version
-  version: 0.9.4
+  version: 0.10.0
 platform: aarch64-linux
 authors:
 - Andrew Kane
 bindir: bin
 cert_chain: []
-date: 2025-03-08 00:00:00.000000000 Z
+date: 1980-01-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ffi
@@ -53,14 +53,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: '3'
+      version: '3.2'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.6.2
+rubygems_version: 3.6.7
 specification_version: 4
 summary: High performance scoring engine for ML models
 test_files: []