bigdl-core-npu 2.6.0b20250122__cp310-cp310-win_amd64.whl → 2.6.0b20250126__cp310-cp310-win_amd64.whl
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.
- bigdl-core-npu/common.lib +0 -0
- bigdl-core-npu/ggml.dll +0 -0
- bigdl-core-npu/ggml.lib +0 -0
- bigdl-core-npu/include/llamacpp/arg.h +77 -0
- bigdl-core-npu/include/llamacpp/common.h +563 -0
- bigdl-core-npu/include/llamacpp/ggml-alloc.h +76 -0
- bigdl-core-npu/include/llamacpp/ggml-backend.h +241 -0
- bigdl-core-npu/include/llamacpp/ggml.h +2669 -0
- bigdl-core-npu/include/llamacpp/llama.h +1224 -0
- bigdl-core-npu/include/llamacpp/log.h +92 -0
- bigdl-core-npu/llama-cli-npu.exe +0 -0
- bigdl-core-npu/llama.dll +0 -0
- bigdl-core-npu/llama.lib +0 -0
- bigdl-core-npu/npu_llm.dll +0 -0
- {bigdl_core_npu-2.6.0b20250122.dist-info → bigdl_core_npu-2.6.0b20250126.dist-info}/METADATA +1 -1
- {bigdl_core_npu-2.6.0b20250122.dist-info → bigdl_core_npu-2.6.0b20250126.dist-info}/RECORD +20 -8
- intel_npu_acceleration_library/_version.py +1 -1
- intel_npu_acceleration_library/lib/Release/intel_npu_acceleration_library.dll +0 -0
- {bigdl_core_npu-2.6.0b20250122.dist-info → bigdl_core_npu-2.6.0b20250126.dist-info}/WHEEL +0 -0
- {bigdl_core_npu-2.6.0b20250122.dist-info → bigdl_core_npu-2.6.0b20250126.dist-info}/top_level.txt +0 -0
@@ -0,0 +1,92 @@
|
|
1
|
+
#pragma once
|
2
|
+
|
3
|
+
#include "ggml.h" // for ggml_log_level
|
4
|
+
|
5
|
+
#ifndef __GNUC__
|
6
|
+
# define LOG_ATTRIBUTE_FORMAT(...)
|
7
|
+
#elif defined(__MINGW32__)
|
8
|
+
# define LOG_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))
|
9
|
+
#else
|
10
|
+
# define LOG_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))
|
11
|
+
#endif
|
12
|
+
|
13
|
+
#define LOG_DEFAULT_DEBUG 1
|
14
|
+
#define LOG_DEFAULT_LLAMA 0
|
15
|
+
|
16
|
+
// needed by the LOG_TMPL macro to avoid computing log arguments if the verbosity lower
|
17
|
+
// set via gpt_log_set_verbosity()
|
18
|
+
extern int gpt_log_verbosity_thold;
|
19
|
+
|
20
|
+
void gpt_log_set_verbosity_thold(int verbosity); // not thread-safe
|
21
|
+
|
22
|
+
// the gpt_log uses an internal worker thread to print/write log messages
|
23
|
+
// when the worker thread is paused, incoming log messages are discarded
|
24
|
+
struct gpt_log;
|
25
|
+
|
26
|
+
struct gpt_log * gpt_log_init();
|
27
|
+
struct gpt_log * gpt_log_main(); // singleton, automatically destroys itself on exit
|
28
|
+
void gpt_log_pause (struct gpt_log * log); // pause the worker thread, not thread-safe
|
29
|
+
void gpt_log_resume(struct gpt_log * log); // resume the worker thread, not thread-safe
|
30
|
+
void gpt_log_free (struct gpt_log * log);
|
31
|
+
|
32
|
+
LOG_ATTRIBUTE_FORMAT(3, 4)
|
33
|
+
void gpt_log_add(struct gpt_log * log, enum ggml_log_level level, const char * fmt, ...);
|
34
|
+
|
35
|
+
// defaults: file = NULL, colors = false, prefix = false, timestamps = false
|
36
|
+
//
|
37
|
+
// regular log output:
|
38
|
+
//
|
39
|
+
// ggml_backend_metal_log_allocated_size: allocated buffer, size = 6695.84 MiB, ( 6695.91 / 21845.34)
|
40
|
+
// llm_load_tensors: ggml ctx size = 0.27 MiB
|
41
|
+
// llm_load_tensors: offloading 32 repeating layers to GPU
|
42
|
+
// llm_load_tensors: offloading non-repeating layers to GPU
|
43
|
+
//
|
44
|
+
// with prefix = true, timestamps = true, the log output will look like this:
|
45
|
+
//
|
46
|
+
// 0.00.035.060 D ggml_backend_metal_log_allocated_size: allocated buffer, size = 6695.84 MiB, ( 6695.91 / 21845.34)
|
47
|
+
// 0.00.035.064 I llm_load_tensors: ggml ctx size = 0.27 MiB
|
48
|
+
// 0.00.090.578 I llm_load_tensors: offloading 32 repeating layers to GPU
|
49
|
+
// 0.00.090.579 I llm_load_tensors: offloading non-repeating layers to GPU
|
50
|
+
//
|
51
|
+
// I - info (stdout, V = 0)
|
52
|
+
// W - warning (stderr, V = 0)
|
53
|
+
// E - error (stderr, V = 0)
|
54
|
+
// D - debug (stderr, V = LOG_DEFAULT_DEBUG)
|
55
|
+
//
|
56
|
+
|
57
|
+
void gpt_log_set_file (struct gpt_log * log, const char * file); // not thread-safe
|
58
|
+
void gpt_log_set_colors (struct gpt_log * log, bool colors); // not thread-safe
|
59
|
+
void gpt_log_set_prefix (struct gpt_log * log, bool prefix); // whether to output prefix to each log
|
60
|
+
void gpt_log_set_timestamps(struct gpt_log * log, bool timestamps); // whether to output timestamps in the prefix
|
61
|
+
|
62
|
+
// helper macros for logging
|
63
|
+
// use these to avoid computing log arguments if the verbosity of the log is higher than the threshold
|
64
|
+
//
|
65
|
+
// for example:
|
66
|
+
//
|
67
|
+
// LOG_DBG("this is a debug message: %d\n", expensive_function());
|
68
|
+
//
|
69
|
+
// this will avoid calling expensive_function() if LOG_DEFAULT_DEBUG > gpt_log_verbosity_thold
|
70
|
+
//
|
71
|
+
|
72
|
+
#define LOG_TMPL(level, verbosity, ...) \
|
73
|
+
do { \
|
74
|
+
if ((verbosity) <= gpt_log_verbosity_thold) { \
|
75
|
+
gpt_log_add(gpt_log_main(), (level), __VA_ARGS__); \
|
76
|
+
} \
|
77
|
+
} while (0)
|
78
|
+
|
79
|
+
#define LOG(...) LOG_TMPL(GGML_LOG_LEVEL_NONE, 0, __VA_ARGS__)
|
80
|
+
#define LOGV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_NONE, verbosity, __VA_ARGS__)
|
81
|
+
|
82
|
+
#define LOG_INF(...) LOG_TMPL(GGML_LOG_LEVEL_INFO, 0, __VA_ARGS__)
|
83
|
+
#define LOG_WRN(...) LOG_TMPL(GGML_LOG_LEVEL_WARN, 0, __VA_ARGS__)
|
84
|
+
#define LOG_ERR(...) LOG_TMPL(GGML_LOG_LEVEL_ERROR, 0, __VA_ARGS__)
|
85
|
+
#define LOG_DBG(...) LOG_TMPL(GGML_LOG_LEVEL_DEBUG, LOG_DEFAULT_DEBUG, __VA_ARGS__)
|
86
|
+
#define LOG_CNT(...) LOG_TMPL(GGML_LOG_LEVEL_CONT, 0, __VA_ARGS__)
|
87
|
+
|
88
|
+
#define LOG_INFV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_INFO, verbosity, __VA_ARGS__)
|
89
|
+
#define LOG_WRNV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_WARN, verbosity, __VA_ARGS__)
|
90
|
+
#define LOG_ERRV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_ERROR, verbosity, __VA_ARGS__)
|
91
|
+
#define LOG_DBGV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_DEBUG, verbosity, __VA_ARGS__)
|
92
|
+
#define LOG_CNTV(verbosity, ...) LOG_TMPL(GGML_LOG_LEVEL_CONT, verbosity, __VA_ARGS__)
|
bigdl-core-npu/llama-cli-npu.exe
CHANGED
Binary file
|
bigdl-core-npu/llama.dll
ADDED
Binary file
|
bigdl-core-npu/llama.lib
ADDED
Binary file
|
bigdl-core-npu/npu_llm.dll
CHANGED
Binary file
|
@@ -1,11 +1,23 @@
|
|
1
1
|
bigdl-core-npu/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
2
|
-
bigdl-core-npu/
|
3
|
-
bigdl-core-npu/
|
2
|
+
bigdl-core-npu/common.lib,sha256=dszuzlqhqxqlyZzDh3HBvx7Zfr1XX5AOfmpEXWmMphc,6647656
|
3
|
+
bigdl-core-npu/ggml.dll,sha256=qETfE0Uwxl_GjqqFFwhKEwC-xnzcIUDz2G3Bea0Gh0M,704000
|
4
|
+
bigdl-core-npu/ggml.lib,sha256=gnGRJzDCJ4O25sbrAp-tydMm0j50lhiigyXkpqrgW8M,139692
|
5
|
+
bigdl-core-npu/llama-cli-npu.exe,sha256=k0sX6AT1OTr5JUS0tP6_gPWf2VqkIujvaYTouaL4lIg,41984
|
6
|
+
bigdl-core-npu/llama.dll,sha256=DfKUiIdCV7VBX-3dtopFmDQeO1tY-yOAILS8w0pCJQk,1573376
|
7
|
+
bigdl-core-npu/llama.lib,sha256=Z2dQy5y6QJ1P2DubIjfo4cB-T6a2ZfAMEC7MDShrX8Q,1494020
|
8
|
+
bigdl-core-npu/npu_llm.dll,sha256=cIzJ763TQigSFQrcNrCCR4YqWB8gwmJAmxMgGoCCZf4,3924992
|
4
9
|
bigdl-core-npu/npu_llm.lib,sha256=nNYF-btjrT9Pzcd31kQV5BsLBSNdA_DSeyK5RnmOc9s,44524
|
5
10
|
bigdl-core-npu/include/common.h,sha256=P_WP5vhSSYZPmFAZxLMvhEobKRfksfvQt-O8wmGzFSQ,3059
|
6
11
|
bigdl-core-npu/include/npu_llm.h,sha256=siEFqkKnZlzR9cvq7qXiAy5Z65-X14QyWEw__wsLLpA,2605
|
12
|
+
bigdl-core-npu/include/llamacpp/arg.h,sha256=XOLwSDqJsCJJgGCkOGSWcA9yGZ0eiiY7GffHdAxFYYY,3110
|
13
|
+
bigdl-core-npu/include/llamacpp/common.h,sha256=SaGJ1cAmJWd5D-T_03udb6ADTcJ8VX-bMy0GYGDMpI4,25841
|
14
|
+
bigdl-core-npu/include/llamacpp/ggml-alloc.h,sha256=kiWITcUF9Q7kvGSBeCfouV59YyGDNy56VE_0fXXnWHE,3088
|
15
|
+
bigdl-core-npu/include/llamacpp/ggml-backend.h,sha256=cUJR-AC7hUbIxukKixR9vUSV38gmp-9eoViaBG8lyio,14347
|
16
|
+
bigdl-core-npu/include/llamacpp/ggml.h,sha256=_K-zhVjT6uHa5lXMN3k3AwSrwb9W2q1A7v5HiGQc8iQ,105629
|
17
|
+
bigdl-core-npu/include/llamacpp/llama.h,sha256=gMTNuRrEY7z0nkCewKLMZtf5urcSzkQm_HxEs3DJXHI,60967
|
18
|
+
bigdl-core-npu/include/llamacpp/log.h,sha256=D6UiundA26yZOh3ci828u5U4ZaBxWb9CwSNu_3dt6CA,4295
|
7
19
|
intel_npu_acceleration_library/__init__.py,sha256=ZKTIhGMDjF7P6pF-yX8KWcSXbeHWRk24AO_orsa18f8,536
|
8
|
-
intel_npu_acceleration_library/_version.py,sha256=
|
20
|
+
intel_npu_acceleration_library/_version.py,sha256=pfTWD8BzC-mWRwDdCEa_lX70cHhl1MEalYyS5mzHoko,112
|
9
21
|
intel_npu_acceleration_library/compiler.py,sha256=3IdgqjamSC8MLexDBJypIeZRiWIcTFnvQSU1LPXUr7Y,6225
|
10
22
|
intel_npu_acceleration_library/device.py,sha256=9bn8eVXJa5cXIqgfLsQAdkMVtVUQABb8z0-mQik5jRg,7424
|
11
23
|
intel_npu_acceleration_library/dtypes.py,sha256=gdd06Wsc9zIZFHlauUEx4xcK9WGTn1Mu6GkuYDJeA-E,4683
|
@@ -222,7 +234,7 @@ intel_npu_acceleration_library/external/openvino/torch/__init__.py,sha256=RXLzsf
|
|
222
234
|
intel_npu_acceleration_library/functional/__init__.py,sha256=WWKwKOh6Sgovv7mKctA872TbLP98Pg5m5-MREvUmlAA,204
|
223
235
|
intel_npu_acceleration_library/functional/scaled_dot_product_attention.py,sha256=yGUcg4tDQOLuUnP1g74cl-ec8TRr2SuAMcNLlN6qLvE,1620
|
224
236
|
intel_npu_acceleration_library/lib/Release/cache.json,sha256=CyrSqZUWo0Ec4_7ydOiuKIC0Gm8AybrGdozUqUuHxBw,8840377
|
225
|
-
intel_npu_acceleration_library/lib/Release/intel_npu_acceleration_library.dll,sha256=
|
237
|
+
intel_npu_acceleration_library/lib/Release/intel_npu_acceleration_library.dll,sha256=BpYRmTopXmzg3fg-MbIpKPqOGwffLMVjSLJLttnpUWI,341504
|
226
238
|
intel_npu_acceleration_library/lib/Release/openvino.dll,sha256=m7M119p3JBq2YYJJ2zzCaBDz6XivKK3nNykb8L1cvDU,13244768
|
227
239
|
intel_npu_acceleration_library/lib/Release/openvino_auto_batch_plugin.dll,sha256=2v_I9P3Qo0St1bQZMEZscnFOUVvgZQQ0HvQlG3HtTd0,203104
|
228
240
|
intel_npu_acceleration_library/lib/Release/openvino_auto_plugin.dll,sha256=e3Aj9CDRHN30dBEdPSk7OCWe52tWfhI4xeXgyFjuDHg,475488
|
@@ -252,7 +264,7 @@ intel_npu_acceleration_library/nn/functional.py,sha256=UfAKBc0u6RtyaMo14ldH2GpEn
|
|
252
264
|
intel_npu_acceleration_library/nn/linear.py,sha256=Q06SoGQeLaI86nA_ky2GnFC6H2Fw1zyMDILKnpYC2eo,5739
|
253
265
|
intel_npu_acceleration_library/nn/llm.py,sha256=P6dz36Yf6BHtzWcftaghC6QaMI_WeRfQwrCbO7fD6hk,15002
|
254
266
|
intel_npu_acceleration_library/nn/module.py,sha256=EYxoTq6I_YgBDgTF76GPDxHrT8SupOTDGMzQaomBeq8,12667
|
255
|
-
bigdl_core_npu-2.6.
|
256
|
-
bigdl_core_npu-2.6.
|
257
|
-
bigdl_core_npu-2.6.
|
258
|
-
bigdl_core_npu-2.6.
|
267
|
+
bigdl_core_npu-2.6.0b20250126.dist-info/METADATA,sha256=Phpf0TJvqJypcAsoOQmwsksNqAIbGDJOOSsEVh_XXNI,1762
|
268
|
+
bigdl_core_npu-2.6.0b20250126.dist-info/WHEEL,sha256=rzGfZgUcGeKSgIHGYMuqg4xE4VPHxnaldXH6BG0zjVk,101
|
269
|
+
bigdl_core_npu-2.6.0b20250126.dist-info/top_level.txt,sha256=iMQZlTsFPJjlD-Y0MqZEP_9ifI0LlbNCJIOTaMoGMjk,46
|
270
|
+
bigdl_core_npu-2.6.0b20250126.dist-info/RECORD,,
|
Binary file
|
File without changes
|
{bigdl_core_npu-2.6.0b20250122.dist-info → bigdl_core_npu-2.6.0b20250126.dist-info}/top_level.txt
RENAMED
File without changes
|