compressed-tensors 0.3.0__tar.gz → 0.3.2__tar.gz

compressed-tensors-0.3.2/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

compressed-tensors-0.3.2/PKG-INFO

@@ -0,0 +1,100 @@
+Metadata-Version: 2.1
+Name: compressed-tensors
+Version: 0.3.2
+Summary: Library for utilization of compressed safetensors of neural network models
+Home-page: https://github.com/neuralmagic/compressed-tensors
+Author: Neuralmagic, Inc.
+Author-email: support@neuralmagic.com
+License: Apache 2.0
+Platform: UNKNOWN
+Description-Content-Type: text/markdown
+Provides-Extra: dev
+License-File: LICENSE
+
+# compressed-tensors
+
+This repository extends a [safetensors](https://github.com/huggingface/safetensors) format to efficiently store sparse and/or quantized tensors on disk. `compressed-tensors` format supports multiple compression types to minimize the disk space and facilitate the tensor manipulation.
+
+## Motivation
+
+### Reduce disk space by saving sparse tensors in a compressed format
+
+The compressed format stores the data much more efficiently by taking advantage of two properties of tensors:
+
+- Sparse tensors -> due to a large number of entries that are equal to zero.
+- Quantized -> due to their low precision representation.
+
+### Introduce an elegant interface to save/load compressed tensors
+
+The library provides the user with the ability to compress/decompress tensors. The properties of tensors are defined by human-readable configs, allowing the users to understand the compression format at a quick glance.
+
+## Installation
+
+### Pip
+
+```bash
+pip install compressed-tensors
+```
+
+### From source
+
+```bash
+git clone https://github.com/neuralmagic/compressed-tensors
+cd compressed-tensors
+pip install -e .
+```
+
+## Getting started
+
+### Saving/Loading Compressed Tensors (Bitmask Compression)
+
+The function `save_compressed` uses the `compression_format` argument to apply compression to tensors.
+The function `load_compressed` reverses the process: converts the compressed weights on disk to decompressed weights in device memory.
+
+```python
+from compressed_tensors import save_compressed, load_compressed, BitmaskConfig
+from torch import Tensor
+from typing import Dict
+
+# the example BitmaskConfig method efficiently compresses 
+# tensors with large number of zero entries 
+compression_config = BitmaskConfig()
+
+tensors: Dict[str, Tensor] = {"tensor_1": Tensor(
+    [[0.0, 0.0, 0.0], 
+     [1.0, 1.0, 1.0]]
+)}
+# compress tensors using BitmaskConfig compression format (save them efficiently on disk)
+save_compressed(tensors, "model.safetensors", compression_format=compression_config.format)
+
+# decompress tensors (load_compressed returns a generator for memory efficiency)
+decompressed_tensors = {}
+for tensor_name, tensor in load_compressed("model.safetensors", compression_config = compression_config):
+    decompressed_tensors[tensor_name] = tensor
+```
+
+## Saving/Loading Compressed Models (Bitmask Compression)
+
+We can apply bitmask compression to a whole model. For more detailed example see `example` directory.
+```python
+from compressed_tensors import save_compressed_model, load_compressed, BitmaskConfig
+from transformers import AutoModelForCausalLM
+
+model_name = "neuralmagic/llama2.c-stories110M-pruned50"
+model = AutoModelForCausalLM.from_pretrained(model_name)
+
+original_state_dict = model.state_dict()
+
+compression_config = BitmaskConfig()
+
+# save compressed model weights
+save_compressed_model(model, "compressed_model.safetensors", compression_format=compression_config.format)
+
+# load compressed model weights (`dict` turns generator into a dictionary)
+state_dict = dict(load_compressed("compressed_model.safetensors", compression_config))
+```
+
+For more in-depth tutorial on bitmask compression, refer to the [notebook](https://github.com/neuralmagic/compressed-tensors/blob/d707c5b84bc3fef164aebdcd97cb6eaa571982f8/examples/bitmask_compression.ipynb).
+
+
+

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/README.md

@@ -80,3 +80,6 @@ save_compressed_model(model, "compressed_model.safetensors", compression_format=
 # load compressed model weights (`dict` turns generator into a dictionary)
 state_dict = dict(load_compressed("compressed_model.safetensors", compression_config))
 ```
+
+For more in-depth tutorial on bitmask compression, refer to the [notebook](https://github.com/neuralmagic/compressed-tensors/blob/d707c5b84bc3fef164aebdcd97cb6eaa571982f8/examples/bitmask_compression.ipynb).
+

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/setup.py

@@ -14,7 +14,10 @@
 
 
 from setuptools import setup, find_packages
-from typing import List, Dict
+from typing import List, Dict, Tuple
+
+def _setup_long_description() -> Tuple[str, str]:
+    return open("README.md", "r", encoding="utf-8").read(), "text/markdown"
 
 def _setup_packages() -> List:
     return find_packages(
@@ -22,17 +25,21 @@ def _setup_packages() -> List:
     )
     
 def _setup_install_requires() -> List:
-    return ["torch>=1.7.0", "transformers<=4.40", "pydantic<2.7"]
+    return ["torch>=1.7.0", "transformers<4.41", "pydantic<2.7"]
 
 def _setup_extras() -> Dict:
     return {"dev": ["black==22.12.0", "isort==5.8.0", "wheel>=0.36.2", "flake8>=3.8.3", "pytest>=6.0.0", "nbconvert>=7.16.3"]}
 
 setup(
     name="compressed-tensors",
-    version="0.3.0",
+    version="0.3.2",
     author="Neuralmagic, Inc.",
     author_email="support@neuralmagic.com",
+    license="Apache 2.0",
     description="Library for utilization of compressed safetensors of neural network models",
+    long_description=_setup_long_description()[0],
+    long_description_content_type=_setup_long_description()[1],
+    url="https://github.com/neuralmagic/compressed-tensors",
     extras_require=_setup_extras(),
     install_requires=_setup_install_requires(),
     package_dir={"": "src"},

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors/compressors/__init__.py

@@ -16,10 +16,5 @@
 
 from .base import ModelCompressor
 from .dense import DenseCompressor
-from .helpers import (
-    infer_compressor_from_model_config,
-    load_compressed,
-    save_compressed,
-    save_compressed_model,
-)
+from .helpers import load_compressed, save_compressed, save_compressed_model
 from .sparse_bitmask import BitmaskCompressor, BitmaskTensor

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors/compressors/base.py

@@ -22,6 +22,7 @@ from compressed_tensors.utils import get_safetensors_folder
 from torch import Tensor
 from torch.nn import Module, Parameter
 from tqdm import tqdm
+from transformers import AutoConfig
 
 
 __all__ = ["ModelCompressor"]
@@ -34,6 +35,29 @@ class ModelCompressor(RegistryMixin):
     :param config: config specifying compression parameters
     """
 
+    @classmethod
+    def from_pretrained(
+        cls, pretrained_model_name_or_path: str
+    ) -> Optional["ModelCompressor"]:
+        """
+        Given a path to a model config, extract a sparsity config if it exists and
+        return the associated ModelCompressor
+
+        :param pretrained_model_name_or_path: path to model config on disk or HF hub
+        :return: matching compressor if config contains a sparsity config
+        """
+        config = AutoConfig.from_pretrained(pretrained_model_name_or_path)
+        sparsity_config = getattr(config, SPARSITY_CONFIG_NAME, None)
+        if sparsity_config is None:
+            return None
+
+        format = sparsity_config.get("format")
+        sparsity_config = CompressionConfig.load_from_registry(
+            format, **sparsity_config
+        )
+        compressor = cls.load_from_registry(format, config=sparsity_config)
+        return compressor
+
     def __init__(self, config: Optional[CompressionConfig] = None):
         self.config = config
 
@@ -47,7 +71,7 @@ class ModelCompressor(RegistryMixin):
         raise NotImplementedError()
 
     def decompress(
-        self, path_to_model_or_tensors: str
+        self, path_to_model_or_tensors: str, device: str = "cpu"
     ) -> Generator[Tuple[str, Tensor], None, None]:
         """
         Reads a compressed state dict located at path_to_model_or_tensors

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors/compressors/dense.py

@@ -29,6 +29,6 @@ class DenseCompressor(ModelCompressor):
         return model_state
 
     def decompress(
-        self, path_to_model_or_tensors: str, device: str
+        self, path_to_model_or_tensors: str, device: str = "cpu"
     ) -> Generator[Tuple[str, Tensor], None, None]:
         return iter([])

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors/compressors/helpers.py

@@ -16,45 +16,21 @@ from pathlib import Path
 from typing import Dict, Generator, Optional, Tuple, Union
 
 import torch
-from compressed_tensors.base import SPARSITY_CONFIG_NAME
 from compressed_tensors.compressors import ModelCompressor
 from compressed_tensors.config import CompressionConfig, CompressionFormat
 from compressed_tensors.utils.safetensors_load import get_weight_mappings
 from safetensors import safe_open
 from safetensors.torch import save_file
 from torch import Tensor
-from transformers import AutoConfig
 
 
 __all__ = [
-    "infer_compressor_from_model_config",
     "load_compressed",
     "save_compressed",
     "save_compressed_model",
 ]
 
 
-def infer_compressor_from_model_config(
-    pretrained_model_name_or_path: str,
-) -> Optional[ModelCompressor]:
-    """
-    Given a path to a model config, extract a sparsity config if it exists and return
-    the associated ModelCompressor
-
-    :param pretrained_model_name_or_path: path to model config on disk or HF hub
-    :return: matching compressor if config contains a sparsity config
-    """
-    config = AutoConfig.from_pretrained(pretrained_model_name_or_path)
-    sparsity_config = getattr(config, SPARSITY_CONFIG_NAME, None)
-    if sparsity_config is None:
-        return None
-
-    format = sparsity_config.get("format")
-    sparsity_config = CompressionConfig.load_from_registry(format, **sparsity_config)
-    compressor = ModelCompressor.load_from_registry(format, config=sparsity_config)
-    return compressor
-
-
 def save_compressed(
     tensors: Dict[str, Tensor],
     save_path: Union[str, Path],

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors/compressors/sparse_bitmask.py

@@ -75,8 +75,9 @@ class BitmaskCompressor(ModelCompressor):
         self, path_to_model_or_tensors: str, device: str = "cpu"
     ) -> Generator[Tuple[str, Tensor], None, None]:
         """
-        Reads a bitmask compressed state dict located at path_to_model_or_tensors
-        and returns a generator for sequentially decompressing back to a dense state dict
+        Reads a bitmask compressed state dict located
+        at path_to_model_or_tensors and returns a generator
+        for sequentially decompressing back to a dense state dict
 
         :param model_path: path to compressed safetensors model (directory with
             one or more safetensors files) or compressed tensors file

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors/quantization/lifecycle/forward.py

@@ -111,7 +111,7 @@ def wrap_module_forward_quantized(module: Module, scheme: QuantizationScheme):
 
 
 def _maybe_calibrate_or_quantize(
-    module: Module, value: Module, base_name: str, args: "QuantizationArgs"
+    module: Module, value: torch.Tensor, base_name: str, args: "QuantizationArgs"
 ) -> torch.Tensor:
     # only run quantized for the included stages
     if module.quantization_status not in {
@@ -120,17 +120,23 @@ def _maybe_calibrate_or_quantize(
     }:
         return value
 
-    device = next(module.parameters()).device
-    scale = getattr(module, f"{base_name}_scale")
-    zero_point = getattr(module, f"{base_name}_zero_point")
-
-    if module.quantization_status == QuantizationStatus.CALIBRATION:
-        # get observer and get new quant params from observation
+    if args.dynamic:
+        # dynamic quantization - get scale and zero point directly from observer
         observer = getattr(module, f"{base_name}_observer")
-        updated_scale, updated_zero_point = observer(value)
-
-        # update scale and zero point
-        scale.data = updated_scale.to(device)
-        zero_point.data = updated_zero_point.to(device)
+        scale, zero_point = observer(value)
+    else:
+        # static quantization - get previous scale and zero point from layer
+        scale = getattr(module, f"{base_name}_scale")
+        zero_point = getattr(module, f"{base_name}_zero_point")
+
+        if module.quantization_status == QuantizationStatus.CALIBRATION:
+            # calibration mode - get new quant params from observer
+            observer = getattr(module, f"{base_name}_observer")
+            updated_scale, updated_zero_point = observer(value)
+
+            # update scale and zero point
+            device = next(module.parameters()).device
+            scale.data = updated_scale.to(device)
+            zero_point.data = updated_zero_point.to(device)
 
     return fake_quantize(value, scale, zero_point, args)

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors/quantization/lifecycle/frozen.py

@@ -30,17 +30,17 @@ def freeze_module_quantization(module: Module):
 
     :param module: module to freeze quantization for
     """
-    if not getattr(module, "quantization_scheme", None):
+    scheme = getattr(module, "quantization_scheme", None)
+    if not scheme:
         # no quantization scheme nothing to do
         return
 
-    # delete observers from module
-    observer_names = []
-    for submodule_name, _ in module.named_modules():
-        if "." not in submodule_name and submodule_name.endswith("_observer"):
-            # delete any observers that belong directly to this module
-            observer_names.append(submodule_name)
-    for observer_name in observer_names:
-        delattr(module, observer_name)
+    # delete observers from module if not dynamic
+    if scheme.input_activations and not scheme.input_activations.dynamic:
+        delattr(module, "input_observer")
+    if scheme.weights and not scheme.weights.dynamic:
+        delattr(module, "weight_observer")
+    if scheme.output_activations and not scheme.output_activations.dynamic:
+        delattr(module, "output_observer")
 
     module.quantization_status = QuantizationStatus.FROZEN

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors/quantization/lifecycle/initialize.py

@@ -80,6 +80,13 @@ def initialize_module_for_quantization(
 def _initialize_scale_zero_point_observer(
     module: Module, base_name: str, quantization_args: QuantizationArgs
 ):
+    # initialize observer module and attach as submodule
+    observer = quantization_args.get_observer()
+    module.register_module(f"{base_name}_observer", observer)
+
+    if quantization_args.dynamic:
+        return  # no need to register a scale and zero point for a dynamic observer
+
     device = next(module.parameters()).device
 
     # initializes empty scale and zero point parameters for the module
@@ -90,7 +97,3 @@ def _initialize_scale_zero_point_observer(
         torch.empty(0, device=device, dtype=int), requires_grad=False
     )
     module.register_parameter(f"{base_name}_zero_point", init_zero_point)
-
-    # initialize observer module and attach as submodule
-    observer = quantization_args.get_observer()
-    module.register_module(f"{base_name}_observer", observer)

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors/quantization/observers/memoryless.py

@@ -23,10 +23,10 @@ from torch import FloatTensor, IntTensor, Tensor
 __all__ = ["MemorylessObserver"]
 
 
-@Observer.register("memoryless")
+@Observer.register("memoryless", alias=["dynamic"])
 class MemorylessObserver(Observer):
     """
-    Implements a dynamic quantization observer that sets the scale and
+    Implements a quantization observer that sets the scale and
     zero point based on the latest observed value without tracking state
     """
 

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors/quantization/quant_args.py

@@ -53,6 +53,11 @@ class QuantizationArgs(BaseModel):
     :param group_size: group length to use for the group strategy
     :param block_structure: 2d block structure to use for the block strategy, must be
     of the format "2x4", "8x16", etc.
+    :param dynamic: set True to perform dynamic quantization - values will not be
+        calibrated during calibration phase, instead during inference new quantization
+        ranges will be observed with every sample. Defaults to False for static
+        quantization. Note that enabling dynamic quantization will change the default
+        observer to a memoryless one
     """
 
     num_bits: int = 8
@@ -61,6 +66,7 @@ class QuantizationArgs(BaseModel):
     strategy: QuantizationStrategy = QuantizationStrategy.TENSOR
     group_size: Optional[int] = None
     block_structure: Optional[str] = None
+    dynamic: bool = False
     observer: str = Field(
         default="minmax",
         description=(
@@ -82,4 +88,9 @@ class QuantizationArgs(BaseModel):
         """
         from compressed_tensors.quantization.observers.base import Observer
 
+        if self.observer == "minmax" and self.dynamic:
+            # override defualt observer for dynamic, you never want minmax which
+            # keeps state across samples for dynamic
+            self.observer = "memoryless"
+
         return Observer.load_from_registry(self.observer, quantization_args=self)

compressed-tensors-0.3.2/src/compressed_tensors.egg-info/PKG-INFO

@@ -0,0 +1,100 @@
+Metadata-Version: 2.1
+Name: compressed-tensors
+Version: 0.3.2
+Summary: Library for utilization of compressed safetensors of neural network models
+Home-page: https://github.com/neuralmagic/compressed-tensors
+Author: Neuralmagic, Inc.
+Author-email: support@neuralmagic.com
+License: Apache 2.0
+Platform: UNKNOWN
+Description-Content-Type: text/markdown
+Provides-Extra: dev
+License-File: LICENSE
+
+# compressed-tensors
+
+This repository extends a [safetensors](https://github.com/huggingface/safetensors) format to efficiently store sparse and/or quantized tensors on disk. `compressed-tensors` format supports multiple compression types to minimize the disk space and facilitate the tensor manipulation.
+
+## Motivation
+
+### Reduce disk space by saving sparse tensors in a compressed format
+
+The compressed format stores the data much more efficiently by taking advantage of two properties of tensors:
+
+- Sparse tensors -> due to a large number of entries that are equal to zero.
+- Quantized -> due to their low precision representation.
+
+### Introduce an elegant interface to save/load compressed tensors
+
+The library provides the user with the ability to compress/decompress tensors. The properties of tensors are defined by human-readable configs, allowing the users to understand the compression format at a quick glance.
+
+## Installation
+
+### Pip
+
+```bash
+pip install compressed-tensors
+```
+
+### From source
+
+```bash
+git clone https://github.com/neuralmagic/compressed-tensors
+cd compressed-tensors
+pip install -e .
+```
+
+## Getting started
+
+### Saving/Loading Compressed Tensors (Bitmask Compression)
+
+The function `save_compressed` uses the `compression_format` argument to apply compression to tensors.
+The function `load_compressed` reverses the process: converts the compressed weights on disk to decompressed weights in device memory.
+
+```python
+from compressed_tensors import save_compressed, load_compressed, BitmaskConfig
+from torch import Tensor
+from typing import Dict
+
+# the example BitmaskConfig method efficiently compresses 
+# tensors with large number of zero entries 
+compression_config = BitmaskConfig()
+
+tensors: Dict[str, Tensor] = {"tensor_1": Tensor(
+    [[0.0, 0.0, 0.0], 
+     [1.0, 1.0, 1.0]]
+)}
+# compress tensors using BitmaskConfig compression format (save them efficiently on disk)
+save_compressed(tensors, "model.safetensors", compression_format=compression_config.format)
+
+# decompress tensors (load_compressed returns a generator for memory efficiency)
+decompressed_tensors = {}
+for tensor_name, tensor in load_compressed("model.safetensors", compression_config = compression_config):
+    decompressed_tensors[tensor_name] = tensor
+```
+
+## Saving/Loading Compressed Models (Bitmask Compression)
+
+We can apply bitmask compression to a whole model. For more detailed example see `example` directory.
+```python
+from compressed_tensors import save_compressed_model, load_compressed, BitmaskConfig
+from transformers import AutoModelForCausalLM
+
+model_name = "neuralmagic/llama2.c-stories110M-pruned50"
+model = AutoModelForCausalLM.from_pretrained(model_name)
+
+original_state_dict = model.state_dict()
+
+compression_config = BitmaskConfig()
+
+# save compressed model weights
+save_compressed_model(model, "compressed_model.safetensors", compression_format=compression_config.format)
+
+# load compressed model weights (`dict` turns generator into a dictionary)
+state_dict = dict(load_compressed("compressed_model.safetensors", compression_config))
+```
+
+For more in-depth tutorial on bitmask compression, refer to the [notebook](https://github.com/neuralmagic/compressed-tensors/blob/d707c5b84bc3fef164aebdcd97cb6eaa571982f8/examples/bitmask_compression.ipynb).
+
+
+

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors.egg-info/SOURCES.txt

@@ -1,3 +1,4 @@
+LICENSE
 README.md
 pyproject.toml
 setup.cfg

{compressed-tensors-0.3.0 → compressed-tensors-0.3.2}/src/compressed_tensors.egg-info/requires.txt

@@ -1,6 +1,6 @@
 pydantic<2.7
 torch>=1.7.0
-transformers<=4.40
+transformers<4.41
 
 [dev]
 black==22.12.0

compressed-tensors-0.3.0/PKG-INFO

@@ -1,13 +0,0 @@
-Metadata-Version: 2.1
-Name: compressed-tensors
-Version: 0.3.0
-Summary: Library for utilization of compressed safetensors of neural network models
-Home-page: UNKNOWN
-Author: Neuralmagic, Inc.
-Author-email: support@neuralmagic.com
-License: UNKNOWN
-Platform: UNKNOWN
-Provides-Extra: dev
-
-UNKNOWN
-

compressed-tensors-0.3.0/src/compressed_tensors.egg-info/PKG-INFO

@@ -1,13 +0,0 @@
-Metadata-Version: 2.1
-Name: compressed-tensors
-Version: 0.3.0
-Summary: Library for utilization of compressed safetensors of neural network models
-Home-page: UNKNOWN
-Author: Neuralmagic, Inc.
-Author-email: support@neuralmagic.com
-License: UNKNOWN
-Platform: UNKNOWN
-Provides-Extra: dev
-
-UNKNOWN
-