llama_cpp 0.0.6 → 0.0.7
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +4 -4
- data/CHANGELOG.md +13 -1
- data/ext/llama_cpp/extconf.rb +9 -0
- data/ext/llama_cpp/llama_cpp.cpp +26 -0
- data/ext/llama_cpp/src/ggml-cuda.h +32 -0
- data/ext/llama_cpp/src/ggml-opencl.c +216 -0
- data/ext/llama_cpp/src/ggml-opencl.h +24 -0
- data/ext/llama_cpp/src/ggml.c +1436 -624
- data/ext/llama_cpp/src/ggml.h +654 -627
- data/ext/llama_cpp/src/llama.cpp +212 -29
- data/ext/llama_cpp/src/llama.h +17 -13
- data/ext/llama_cpp/src/llama_util.h +15 -2
- data/lib/llama_cpp/client.rb +151 -0
- data/lib/llama_cpp/version.rb +2 -2
- data/lib/llama_cpp.rb +16 -8
- data/sig/llama_cpp.rbs +16 -1
- metadata +5 -2
data/sig/llama_cpp.rbs
CHANGED
|
@@ -12,9 +12,12 @@ module LLaMACpp
|
|
|
12
12
|
LLAMA_FTYPE_MOSTLY_Q4_1_SOME_F16: Integer
|
|
13
13
|
LLAMA_FTYPE_MOSTLY_Q4_2: Integer
|
|
14
14
|
LLAMA_FTYPE_MOSTLY_Q4_3: Integer
|
|
15
|
+
LLAMA_FTYPE_MOSTLY_Q8_0: Integer
|
|
16
|
+
LLAMA_FTYPE_MOSTLY_Q5_0: Integer
|
|
17
|
+
LLAMA_FTYPE_MOSTLY_Q5_1: Integer
|
|
15
18
|
|
|
16
19
|
def self?.model_quantize: (input_path: String, output_path: String, ftype: Integer, ?n_threads: Integer) -> void
|
|
17
|
-
def self?.generate: (::LLaMACpp::Context, String, ?n_threads: Integer) -> String
|
|
20
|
+
def self?.generate: (::LLaMACpp::Context, String, ?n_predict: Integer, ?n_threads: Integer) -> String
|
|
18
21
|
def self?.print_system_info: () -> void
|
|
19
22
|
def self?.token_bos: () -> Integer
|
|
20
23
|
def self?.token_eos: () -> Integer
|
|
@@ -27,6 +30,7 @@ module LLaMACpp
|
|
|
27
30
|
def initialize: (model_path: String, params: ::LLaMACpp::ContextParams) -> void
|
|
28
31
|
| () -> void
|
|
29
32
|
def embeddings: () -> Array[Float]
|
|
33
|
+
def empty?: () -> bool
|
|
30
34
|
def eval: (tokens: Array[Integer], n_past: Integer, ?n_tokens: Integer, ?n_threads: Integer) -> void
|
|
31
35
|
def free: () -> void
|
|
32
36
|
def load: (model_path: String, params: ::LLaMACpp::ContextParams) -> void
|
|
@@ -59,9 +63,20 @@ module LLaMACpp
|
|
|
59
63
|
def seed=: (Integer) -> Integer
|
|
60
64
|
def use_mlock: () -> bool
|
|
61
65
|
def use_mlock=: (bool) -> bool
|
|
66
|
+
def use_mmap: () -> bool
|
|
67
|
+
def use_mmap=: (bool) -> bool
|
|
62
68
|
def vocab_only: () -> bool
|
|
63
69
|
def vocab_only=: (bool) -> bool
|
|
64
70
|
end
|
|
65
71
|
|
|
66
72
|
class Params = ContextParams
|
|
73
|
+
|
|
74
|
+
class Client
|
|
75
|
+
def initialize(model_path: String, ?lora_adapter_path: String, ?lora_base_path: String,
|
|
76
|
+
?n_ctx: Integer, ?n_parts: Integer, ?memory_f16: bool, ?use_mmap: bool, ?use_mlock: bool,
|
|
77
|
+
?embedding: bool, ?n_threads: Integer, ?seed: Integer) -> void
|
|
78
|
+
def completions(String, ?max_tokens: Integer, ?n_keep: Integer, ?repeat_last_n: Integer, ?n_batch: Integer,
|
|
79
|
+
?top_k: Integer, ?top_p: Float, ?temperature: Float, ?repeat_penalty: Float) -> String
|
|
80
|
+
def embeddings(String) -> Array[Float]
|
|
81
|
+
end
|
|
67
82
|
end
|
metadata
CHANGED
|
@@ -1,14 +1,14 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: llama_cpp
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.0.
|
|
4
|
+
version: 0.0.7
|
|
5
5
|
platform: ruby
|
|
6
6
|
authors:
|
|
7
7
|
- yoshoku
|
|
8
8
|
autorequire:
|
|
9
9
|
bindir: exe
|
|
10
10
|
cert_chain: []
|
|
11
|
-
date: 2023-04-
|
|
11
|
+
date: 2023-04-29 00:00:00.000000000 Z
|
|
12
12
|
dependencies: []
|
|
13
13
|
description: llama_cpp.rb provides Ruby bindings for the llama.cpp.
|
|
14
14
|
email:
|
|
@@ -27,12 +27,15 @@ files:
|
|
|
27
27
|
- ext/llama_cpp/llama_cpp.h
|
|
28
28
|
- ext/llama_cpp/src/LICENSE
|
|
29
29
|
- ext/llama_cpp/src/ggml-cuda.h
|
|
30
|
+
- ext/llama_cpp/src/ggml-opencl.c
|
|
31
|
+
- ext/llama_cpp/src/ggml-opencl.h
|
|
30
32
|
- ext/llama_cpp/src/ggml.c
|
|
31
33
|
- ext/llama_cpp/src/ggml.h
|
|
32
34
|
- ext/llama_cpp/src/llama.cpp
|
|
33
35
|
- ext/llama_cpp/src/llama.h
|
|
34
36
|
- ext/llama_cpp/src/llama_util.h
|
|
35
37
|
- lib/llama_cpp.rb
|
|
38
|
+
- lib/llama_cpp/client.rb
|
|
36
39
|
- lib/llama_cpp/version.rb
|
|
37
40
|
- sig/llama_cpp.rbs
|
|
38
41
|
homepage: https://github.com/yoshoku/llama_cpp.rb
|