fine 0.1.0

data/lib/fine/transforms/normalize.rb

@@ -0,0 +1,33 @@
+# frozen_string_literal: true
+
+module Fine
+  module Transforms
+    # Normalize a tensor with mean and standard deviation
+    class Normalize
+      # ImageNet normalization values (commonly used)
+      IMAGENET_MEAN = [0.485, 0.456, 0.406].freeze
+      IMAGENET_STD = [0.229, 0.224, 0.225].freeze
+
+      attr_reader :mean, :std
+
+      # @param mean [Array<Float>] Mean values for each channel
+      # @param std [Array<Float>] Standard deviation for each channel
+      def initialize(mean: IMAGENET_MEAN, std: IMAGENET_STD)
+        @mean = mean
+        @std = std
+      end
+
+      def call(tensor)
+        # Expect tensor shape: (C, H, W)
+        raise ArgumentError, "Expected tensor, got #{tensor.class}" unless tensor.is_a?(Torch::Tensor)
+
+        # Convert mean and std to tensors with shape (C, 1, 1)
+        mean_tensor = Torch.tensor(@mean).view(-1, 1, 1)
+        std_tensor = Torch.tensor(@std).view(-1, 1, 1)
+
+        # Normalize: (x - mean) / std
+        (tensor - mean_tensor) / std_tensor
+      end
+    end
+  end
+end

data/lib/fine/transforms/resize.rb

@@ -0,0 +1,35 @@
+# frozen_string_literal: true
+
+module Fine
+  module Transforms
+    # Resize image to specified dimensions
+    class Resize
+      attr_reader :width, :height, :interpolation
+
+      # @param width [Integer] Target width
+      # @param height [Integer, nil] Target height (defaults to width for square)
+      # @param interpolation [Symbol] Interpolation method (:bilinear, :nearest, :bicubic)
+      def initialize(width, height = nil, interpolation: :bilinear)
+        @width = width
+        @height = height || width
+        @interpolation = interpolation
+      end
+
+      def call(image)
+        # Map interpolation to vips kernel names
+        vips_kernel = case @interpolation
+                      when :nearest then :nearest
+                      when :bilinear then :linear
+                      when :bicubic then :cubic
+                      else :linear
+                      end
+
+        # Calculate scale factors
+        h_scale = @width.to_f / image.width
+        v_scale = @height.to_f / image.height
+
+        image.resize(h_scale, vscale: v_scale, kernel: vips_kernel)
+      end
+    end
+  end
+end

data/lib/fine/transforms/to_tensor.rb

@@ -0,0 +1,53 @@
+# frozen_string_literal: true
+
+module Fine
+  module Transforms
+    # Convert a Vips::Image to a Torch::Tensor
+    class ToTensor
+      # @param scale [Boolean] If true, scale pixel values to [0, 1]
+      def initialize(scale: true)
+        @scale = scale
+      end
+
+      def call(image)
+        # Get image as array of bytes
+        # Vips images are (H, W, C) format
+
+        # Ensure image is in RGB format
+        image = ensure_rgb(image)
+
+        # Get raw pixel data as a flat array
+        width = image.width
+        height = image.height
+        bands = image.bands
+
+        # Convert to array of floats
+        data = image.write_to_memory.unpack("C*")
+
+        # Create tensor with shape (H, W, C)
+        tensor = Torch.tensor(data, dtype: :float32).reshape([height, width, bands])
+
+        # Scale to [0, 1] if requested
+        tensor = tensor / 255.0 if @scale
+
+        # Permute to (C, H, W) format expected by PyTorch
+        tensor.permute([2, 0, 1])
+      end
+
+      private
+
+      def ensure_rgb(image)
+        case image.bands
+        when 1
+          # Grayscale to RGB
+          image.bandjoin([image, image])
+        when 4
+          # RGBA to RGB (drop alpha)
+          image.extract_band(0, n: 3)
+        else
+          image
+        end
+      end
+    end
+  end
+end

data/lib/fine/version.rb

@@ -0,0 +1,3 @@
+module Fine
+  VERSION = "0.1.0"
+end

data/lib/fine.rb

@@ -0,0 +1,112 @@
+# frozen_string_literal: true
+
+require "torch"
+require "safetensors"
+require "vips"
+require "tokenizers"
+require "tty-progressbar"
+require "down"
+require "json"
+require "fileutils"
+
+require_relative "fine/version"
+require_relative "fine/error"
+require_relative "fine/configuration"
+
+# Hub
+require_relative "fine/hub/config_loader"
+require_relative "fine/hub/model_downloader"
+require_relative "fine/hub/safetensors_loader"
+
+# Tokenizers
+require_relative "fine/tokenizers/auto_tokenizer"
+
+# Transforms (Image)
+require_relative "fine/transforms/compose"
+require_relative "fine/transforms/resize"
+require_relative "fine/transforms/normalize"
+require_relative "fine/transforms/to_tensor"
+
+# Datasets
+require_relative "fine/datasets/image_dataset"
+require_relative "fine/datasets/data_loader"
+require_relative "fine/datasets/text_dataset"
+require_relative "fine/datasets/text_data_loader"
+require_relative "fine/datasets/instruction_dataset"
+
+# Models - Vision
+require_relative "fine/models/base"
+require_relative "fine/models/siglip2_vision_encoder"
+require_relative "fine/models/classification_head"
+require_relative "fine/models/siglip2_for_image_classification"
+
+# Models - Text
+require_relative "fine/models/bert_encoder"
+require_relative "fine/models/bert_for_sequence_classification"
+require_relative "fine/models/sentence_transformer"
+
+# Models - LLM
+require_relative "fine/models/llama_decoder"
+require_relative "fine/models/gemma3_decoder"
+require_relative "fine/models/causal_lm"
+
+# Training
+require_relative "fine/training/trainer"
+require_relative "fine/training/text_trainer"
+require_relative "fine/training/llm_trainer"
+
+# Callbacks
+require_relative "fine/callbacks/base"
+require_relative "fine/callbacks/progress_bar"
+
+# High-level API
+require_relative "fine/image_classifier"
+require_relative "fine/text_classifier"
+require_relative "fine/text_embedder"
+require_relative "fine/llm"
+
+# Export
+require_relative "fine/export"
+
+module Fine
+  class << self
+    attr_accessor :configuration
+
+    def configure
+      self.configuration ||= GlobalConfiguration.new
+      yield(configuration) if block_given?
+      configuration
+    end
+
+    def cache_dir
+      configuration&.cache_dir || File.expand_path("~/.cache/fine")
+    end
+
+    def device
+      configuration&.device || detect_device
+    end
+
+    private
+
+    def detect_device
+      if Torch::CUDA.available?
+        "cuda"
+      elsif defined?(Torch::Backends::MPS) && Torch::Backends::MPS.available?
+        "mps"
+      else
+        "cpu"
+      end
+    end
+  end
+
+  class GlobalConfiguration
+    attr_accessor :cache_dir, :device, :log_level, :progress_bar
+
+    def initialize
+      @cache_dir = File.expand_path("~/.cache/fine")
+      @device = nil # auto-detect
+      @log_level = :info
+      @progress_bar = true
+    end
+  end
+end

data/mise.toml

@@ -0,0 +1,2 @@
+[tools]
+ruby = "3.3"

metadata

@@ -0,0 +1,240 @@
+--- !ruby/object:Gem::Specification
+name: fine
+version: !ruby/object:Gem::Version
+  version: 0.1.0
+platform: ruby
+authors:
+- Chris Hasinski
+autorequire:
+bindir: exe
+cert_chain: []
+date: 2026-01-26 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: torch-rb
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0.17'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0.17'
+- !ruby/object:Gem::Dependency
+  name: safetensors
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0.1'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0.1'
+- !ruby/object:Gem::Dependency
+  name: tokenizers
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0.4'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0.4'
+- !ruby/object:Gem::Dependency
+  name: ruby-vips
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '2.1'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '2.1'
+- !ruby/object:Gem::Dependency
+  name: tty-progressbar
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0.18'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0.18'
+- !ruby/object:Gem::Dependency
+  name: down
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '5.0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '5.0'
+- !ruby/object:Gem::Dependency
+  name: rake
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '13.0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '13.0'
+- !ruby/object:Gem::Dependency
+  name: rspec
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '3.12'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '3.12'
+- !ruby/object:Gem::Dependency
+  name: rubocop
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.50'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.50'
+description: A Ruby-native interface for fine-tuning machine learning models, starting
+  with image classification using SigLIP2
+email:
+- krzysztof.hasinski@gmail.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- ".rspec"
+- CHANGELOG.md
+- Gemfile
+- Gemfile.lock
+- LICENSE
+- README.md
+- Rakefile
+- docs/installation.md
+- docs/tutorials/llm-fine-tuning.md
+- docs/tutorials/model-export.md
+- docs/tutorials/siglip2-image-classification.md
+- docs/tutorials/siglip2-object-recognition.md
+- docs/tutorials/siglip2-similarity-search.md
+- docs/tutorials/text-classification.md
+- docs/tutorials/text-embeddings.md
+- examples/basic_classification.rb
+- examples/data/tool_calls.jsonl
+- examples/demo_training.rb
+- examples/finetune_gemma3_tools.rb
+- examples/real_llm_test.rb
+- examples/real_text_classification_test.rb
+- examples/real_text_embedder_test.rb
+- examples/real_training_test.rb
+- examples/test_export.rb
+- examples/test_image_classifier.rb
+- examples/test_llm.rb
+- examples/test_text_classifier.rb
+- lib/fine.rb
+- lib/fine/callbacks/base.rb
+- lib/fine/callbacks/progress_bar.rb
+- lib/fine/configuration.rb
+- lib/fine/datasets/data_loader.rb
+- lib/fine/datasets/image_dataset.rb
+- lib/fine/datasets/instruction_dataset.rb
+- lib/fine/datasets/text_data_loader.rb
+- lib/fine/datasets/text_dataset.rb
+- lib/fine/error.rb
+- lib/fine/export.rb
+- lib/fine/export/gguf_exporter.rb
+- lib/fine/export/onnx_exporter.rb
+- lib/fine/hub/config_loader.rb
+- lib/fine/hub/model_downloader.rb
+- lib/fine/hub/safetensors_loader.rb
+- lib/fine/image_classifier.rb
+- lib/fine/llm.rb
+- lib/fine/models/base.rb
+- lib/fine/models/bert_encoder.rb
+- lib/fine/models/bert_for_sequence_classification.rb
+- lib/fine/models/causal_lm.rb
+- lib/fine/models/classification_head.rb
+- lib/fine/models/gemma3_decoder.rb
+- lib/fine/models/llama_decoder.rb
+- lib/fine/models/sentence_transformer.rb
+- lib/fine/models/siglip2_for_image_classification.rb
+- lib/fine/models/siglip2_vision_encoder.rb
+- lib/fine/text_classifier.rb
+- lib/fine/text_embedder.rb
+- lib/fine/tokenizers/auto_tokenizer.rb
+- lib/fine/training/llm_trainer.rb
+- lib/fine/training/text_trainer.rb
+- lib/fine/training/trainer.rb
+- lib/fine/transforms/compose.rb
+- lib/fine/transforms/normalize.rb
+- lib/fine/transforms/resize.rb
+- lib/fine/transforms/to_tensor.rb
+- lib/fine/version.rb
+- mise.toml
+homepage: https://github.com/khasinski/fine
+licenses:
+- MIT
+metadata:
+  homepage_uri: https://github.com/khasinski/fine
+  source_code_uri: https://github.com/khasinski/fine
+  changelog_uri: https://github.com/khasinski/fine/blob/main/CHANGELOG.md
+post_install_message:
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '3.1'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubygems_version: 3.5.22
+signing_key:
+specification_version: 4
+summary: Fine-tune ML models with Ruby
+test_files: []