mostlyai-engine 1.1.7__tar.gz → 1.1.8__tar.gz
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.
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/PKG-INFO +15 -15
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/__init__.py +1 -1
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/training.py +1 -1
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/pyproject.toml +15 -15
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/.gitignore +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/LICENSE +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/README.md +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_common.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_dtypes.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/__init__.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/language/__init__.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/language/categorical.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/language/datetime.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/language/numeric.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/language/text.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/__init__.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/categorical.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/character.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/datetime.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/itt.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/lat_long.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/numeric.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/__init__.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/common.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/encoding.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/engine/__init__.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/engine/base.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/engine/hf_engine.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/engine/vllm_engine.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/formatron_utils.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/generation.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/lstm.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/tokenizer_utils.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_memory.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_tabular/__init__.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_tabular/argn.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_tabular/common.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_tabular/encoding.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_tabular/fairness.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_tabular/generation.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_tabular/training.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_training_utils.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_workspace.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/analysis.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/domain.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/encoding.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/generation.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/logging.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/splitting.py +0 -0
- {mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/training.py +0 -0
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
Metadata-Version: 2.4
|
|
2
2
|
Name: mostlyai-engine
|
|
3
|
-
Version: 1.1.
|
|
3
|
+
Version: 1.1.8
|
|
4
4
|
Summary: Synthetic Data Engine
|
|
5
5
|
Project-URL: homepage, https://github.com/mostly-ai/mostlyai-engine
|
|
6
6
|
Project-URL: repository, https://github.com/mostly-ai/mostlyai-engine
|
|
@@ -39,24 +39,24 @@ Requires-Dist: pyarrow>=16.0.0
|
|
|
39
39
|
Requires-Dist: rich<14,>=13.9.4
|
|
40
40
|
Requires-Dist: setuptools<76,>=75.0.0
|
|
41
41
|
Requires-Dist: tokenizers<0.22,>=0.21.0
|
|
42
|
-
Requires-Dist: torch==2.
|
|
43
|
-
Requires-Dist: torch>=2.
|
|
44
|
-
Requires-Dist: torchaudio==2.
|
|
45
|
-
Requires-Dist: torchvision==0.
|
|
42
|
+
Requires-Dist: torch==2.6.0; sys_platform == 'linux'
|
|
43
|
+
Requires-Dist: torch>=2.6.0; sys_platform == 'darwin'
|
|
44
|
+
Requires-Dist: torchaudio==2.6.0; sys_platform == 'linux'
|
|
45
|
+
Requires-Dist: torchvision==0.21.0; sys_platform == 'linux'
|
|
46
46
|
Requires-Dist: transformers<5,>=4.45.2
|
|
47
47
|
Provides-Extra: cpu
|
|
48
|
-
Requires-Dist: torch==2.
|
|
49
|
-
Requires-Dist: torch==2.
|
|
50
|
-
Requires-Dist: torchaudio==2.
|
|
51
|
-
Requires-Dist: torchvision==0.
|
|
48
|
+
Requires-Dist: torch==2.6.0+cpu; (sys_platform == 'linux') and extra == 'cpu'
|
|
49
|
+
Requires-Dist: torch==2.6.0; (sys_platform != 'linux') and extra == 'cpu'
|
|
50
|
+
Requires-Dist: torchaudio==2.6.0+cpu; (sys_platform == 'linux') and extra == 'cpu'
|
|
51
|
+
Requires-Dist: torchvision==0.21.0+cpu; (sys_platform == 'linux') and extra == 'cpu'
|
|
52
52
|
Provides-Extra: gpu
|
|
53
53
|
Requires-Dist: bitsandbytes<0.43,>=0.42.0; (sys_platform == 'darwin') and extra == 'gpu'
|
|
54
|
-
Requires-Dist: bitsandbytes<0.
|
|
55
|
-
Requires-Dist: torch==2.
|
|
56
|
-
Requires-Dist: torchaudio==2.
|
|
57
|
-
Requires-Dist: torchvision==0.
|
|
58
|
-
Requires-Dist: vllm<0.
|
|
59
|
-
Requires-Dist: vllm<0.
|
|
54
|
+
Requires-Dist: bitsandbytes<0.46,>=0.43.3; (sys_platform == 'linux') and extra == 'gpu'
|
|
55
|
+
Requires-Dist: torch==2.6.0; extra == 'gpu'
|
|
56
|
+
Requires-Dist: torchaudio==2.6.0; extra == 'gpu'
|
|
57
|
+
Requires-Dist: torchvision==0.21.0; extra == 'gpu'
|
|
58
|
+
Requires-Dist: vllm<0.9,>=0.7.3; (sys_platform == 'darwin') and extra == 'gpu'
|
|
59
|
+
Requires-Dist: vllm<0.9,>=0.8.0; (sys_platform == 'linux') and extra == 'gpu'
|
|
60
60
|
Description-Content-Type: text/markdown
|
|
61
61
|
|
|
62
62
|
# Synthetic Data Engine 💎
|
|
@@ -22,7 +22,7 @@ from mostlyai.engine.generation import generate
|
|
|
22
22
|
|
|
23
23
|
|
|
24
24
|
__all__ = ["split", "analyze", "encode", "train", "generate", "init_logging"]
|
|
25
|
-
__version__ = "1.1.
|
|
25
|
+
__version__ = "1.1.8"
|
|
26
26
|
|
|
27
27
|
# suppress specific warning related to os.fork() in multi-threaded processes
|
|
28
28
|
warnings.filterwarnings("ignore", category=DeprecationWarning, message=".*multi-threaded.*fork.*")
|
|
@@ -480,7 +480,7 @@ def train(
|
|
|
480
480
|
gradient_accumulation_steps = default_gradient_accumulation_steps
|
|
481
481
|
else:
|
|
482
482
|
if batch_size is None:
|
|
483
|
-
batch_size = 2**
|
|
483
|
+
batch_size = 2**10 # 1024, max 10 reductions
|
|
484
484
|
if gradient_accumulation_steps is None:
|
|
485
485
|
gradient_accumulation_steps = 1
|
|
486
486
|
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
[project]
|
|
2
2
|
name = "mostlyai-engine"
|
|
3
|
-
version = "1.1.
|
|
3
|
+
version = "1.1.8"
|
|
4
4
|
description = "Synthetic Data Engine"
|
|
5
5
|
authors = [{ name = "MOSTLY AI", email = "dev@mostly.ai" }]
|
|
6
6
|
requires-python = ">=3.10"
|
|
@@ -41,27 +41,27 @@ dependencies = [
|
|
|
41
41
|
"formatron>=0.4.11,<0.5",
|
|
42
42
|
"json-repair>=0.30.0",
|
|
43
43
|
"rich>=13.9.4,<14",
|
|
44
|
-
"torch>=2.
|
|
45
|
-
"torch==2.
|
|
46
|
-
"torchaudio==2.
|
|
47
|
-
"torchvision==0.
|
|
44
|
+
"torch>=2.6.0; sys_platform == 'darwin'",
|
|
45
|
+
"torch==2.6.0; sys_platform == 'linux'", # for vllm 0.8.0 compatibility
|
|
46
|
+
"torchaudio==2.6.0; sys_platform == 'linux'", # for vllm 0.8.0 compatibility
|
|
47
|
+
"torchvision==0.21.0; sys_platform == 'linux'" # for vllm 0.8.0 compatibility
|
|
48
48
|
]
|
|
49
49
|
|
|
50
50
|
[project.optional-dependencies]
|
|
51
51
|
gpu = [
|
|
52
|
-
"torch==2.
|
|
53
|
-
"torchaudio==2.
|
|
54
|
-
"torchvision==0.
|
|
52
|
+
"torch==2.6.0", # for vllm 0.8.0 compatibility
|
|
53
|
+
"torchaudio==2.6.0", # for vllm 0.8.0 compatibility
|
|
54
|
+
"torchvision==0.21.0", # for vllm 0.8.0 compatibility
|
|
55
55
|
"bitsandbytes>=0.42.0,<0.43; sys_platform == 'darwin'",
|
|
56
|
-
"vllm>=0.7.3,<0.
|
|
57
|
-
"bitsandbytes>=0.43.3,<0.
|
|
58
|
-
"vllm>=0.
|
|
56
|
+
"vllm>=0.7.3,<0.9; sys_platform == 'darwin'",
|
|
57
|
+
"bitsandbytes>=0.43.3,<0.46; sys_platform == 'linux'",
|
|
58
|
+
"vllm>=0.8.0,<0.9; sys_platform == 'linux'",
|
|
59
59
|
]
|
|
60
60
|
cpu = [
|
|
61
|
-
"torch==2.
|
|
62
|
-
"torchaudio==2.
|
|
63
|
-
"torchvision==0.
|
|
64
|
-
"torch==2.
|
|
61
|
+
"torch==2.6.0+cpu; sys_platform == 'linux'", # for vllm 0.8.0 compatibility
|
|
62
|
+
"torchaudio==2.6.0+cpu; sys_platform == 'linux'", # for vllm 0.8.0 compatibility
|
|
63
|
+
"torchvision==0.21.0+cpu; sys_platform == 'linux'", # for vllm 0.8.0 compatibility
|
|
64
|
+
"torch==2.6.0; sys_platform != 'linux'",
|
|
65
65
|
]
|
|
66
66
|
|
|
67
67
|
[dependency-groups]
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/language/__init__.py
RENAMED
|
File without changes
|
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/language/datetime.py
RENAMED
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/language/numeric.py
RENAMED
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/language/text.py
RENAMED
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/__init__.py
RENAMED
|
File without changes
|
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/character.py
RENAMED
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/datetime.py
RENAMED
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/itt.py
RENAMED
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/lat_long.py
RENAMED
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_encoding_types/tabular/numeric.py
RENAMED
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/engine/__init__.py
RENAMED
|
File without changes
|
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/engine/hf_engine.py
RENAMED
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/engine/vllm_engine.py
RENAMED
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/formatron_utils.py
RENAMED
|
File without changes
|
|
File without changes
|
|
File without changes
|
{mostlyai_engine-1.1.7 → mostlyai_engine-1.1.8}/mostlyai/engine/_language/tokenizer_utils.py
RENAMED
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|