llama_cpp 0.3.2 → 0.3.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f1fcd28849baae5e90c466665aff4fe5da1d848193ebcf74c3fe333c5674191c
-  data.tar.gz: fcb0c64528d24c5cfad677f17bfd6e1e817a4b8279317ca5b2113302735598b9
+  metadata.gz: 35afb5cc65c290036ae7e45459eadc9b509f34f33a3f7708244cf47f1a38829f
+  data.tar.gz: 3301158526c63d9d2004e22bda0d1cc8025b4343d8d737df96260786531b074d
 SHA512:
-  metadata.gz: c70b5f919feb7a585efbe21b3360254c2f5789504cd73fecee12fd686483c77eeb763ed91a8e7434d5852208555a78f168b358d0895f15b1ea7e774d36d6910a
-  data.tar.gz: f554ad58fc9d68c39b80995b7f424468386b32a5847dbdefbceb1cba53ff7182da35be8599523d82a6daa8fee23667d07e06faedc4c727d52e8fc594d0bc7d3f
+  metadata.gz: b0a50f9f012f44f119a70790d3de07c7fcc64151246791e270e4ff9fc479a85a01c53cf2775945eba3145a3ba89da55a8d14891c6236cfeae16aed5ae455cf0d
+  data.tar.gz: ede388584e115ae93d509b6c15b288303c348f3cfe8ea46879a1b69e6c96be31a321edbb52cfbeb309a8fb456738f3f6b7cc1d3f71ce7addbd05b3a1e73d4755

data/CHANGELOG.md

@@ -1,3 +1,40 @@
+## [[0.3.4](https://github.com/yoshoku/llama_cpp.rb/compare/v0.3.3...v0.3.4)] - 2023-07-23
+
+- Bump bundled llama.cpp from master-32c5411 to master-d924522.
+  - Add `rope_freq_base` and `rope_freq_scale` options to ContextParams.
+  - Add `max_devices` module function to LLaMACpp.
+  - Add `n_vocab`, `n_ctx`, and `n_embd` methods to Model.
+  - Add `vocab`, `tokenize`, and `token_to_str` methods to Model.
+  ```ruby
+  require 'llama_cpp'
+
+  params = LLaMACpp::ContextParams.new
+  model = LLaMACpp::Model.new(model_path: '/path/to/model.bin', params: params)
+
+  p model.tokenize(text: 'hello, world')
+  # => [12199, 29892, 3186]
+
+  p model.token_to_str(12199)
+  # => "hello"
+  ```
+
+**Breaking Changes**
+- Fix to automatically call `backend_free` method when Ruby script exits.
+- Remove `smooth_factor` argument from `sample_classifier_free_guidance methos` on Context.
+
+## [[0.3.3](https://github.com/yoshoku/llama_cpp.rb/compare/v0.3.2...v0.3.3)] - 2023-07-15
+
+- Bump bundled llama.cpp from master-481f793 to master-32c5411.
+- Add MPI config options:
+  ```
+  $ gem install llama_cpp -- --with-mpi
+  ```
+- Add `backend_free` module function to `LLaMACpp`. This method should be called once at the end of the program when the MPI option is enabled.
+- Add `sample_classifier_free_guidance` method to `Context`.
+
+**Breaking Changes**
+- Rename `init_backend` method to `backend_init`. This method is called internally at `require 'llama_cpp'`.
+
 ## [[0.3.2](https://github.com/yoshoku/llama_cpp.rb/compare/v0.3.1...v0.3.2)] - 2023-07-08
 
 - Bump bundled llama.cpp from master-b8c8dda to master-481f793.

data/ext/llama_cpp/extconf.rb

@@ -7,6 +7,7 @@ abort 'libstdc++ is not found.' unless have_library('stdc++')
 
 $srcs = %w[ggml.c llama.cpp llama_cpp.cpp]
 $srcs << 'ggml-opencl.cpp' if with_config('clblast')
+$srcs << 'ggml-mpi.c' if with_config('mpi')
 $CFLAGS << ' -w -DNDEBUG'
 $CXXFLAGS << ' -std=c++11 -DNDEBUG'
 $INCFLAGS << ' -I$(srcdir)/src'
@@ -76,6 +77,14 @@ if with_config('clblast')
   end
 end
 
+if with_config('mpi')
+  abort 'libmpi is not found.' unless have_library('mpi')
+  abort 'mpi.h is not found.' unless have_header('mpi.h')
+
+  $CFLAGS << ' -DGGML_USE_MPI -Wno-cast-qual'
+  $CXXFLAGS << ' -DGGML_USE_MPI -Wno-cast-qual'
+end
+
 UNAME_M = RbConfig::CONFIG['build_cpu'] || RbConfig::CONFIG['host_cpu'] || RbConfig::CONFIG['target_cpu']
 
 # rubocop:disable Layout/LineLength