ai-edge-torch-nightly 0.2.0.dev20240806__py3-none-any.whl → 0.2.0.dev20240808__py3-none-any.whl

ai_edge_torch/quantize/pt2e_quantizer_utils.py

@@ -31,7 +31,7 @@ from torch.ao.quantization.quantizer import SharedQuantizationSpec
 from torch.ao.quantization.quantizer.utils import _annotate_input_qspec_map
 from torch.ao.quantization.quantizer.utils import _annotate_output_qspec
 from torch.fx import Node
-from torch.fx.passes.utils.matcher_with_name_node_map_utils import SubgraphMatcherWithNameNodeMap  # NOQA
+from torch.fx.passes.utils.matcher_with_name_node_map_utils import SubgraphMatcherWithNameNodeMap
 from torch.fx.passes.utils.source_matcher_utils import get_source_partitions
 import torch.nn.functional as F
 
@@ -95,9 +95,10 @@ class OperatorConfig(NamedTuple):
 
 
 def _is_annotated(nodes: List[Node]):
-  """
-  Given a list of nodes (that represents an operator pattern),
-  check if any of the node is annotated, return True if any of the node
+  """Checks if a list of nodes is annotated.
+
+  Given a list of nodes (that represents an operator pattern), check if any of
+  the node is annotated, return True if any of the node
   is annotated, otherwise return False
   """
   annotated = False
@@ -418,9 +419,9 @@ def _annotate_conv_bn(
     quantization_config: Optional[QuantizationConfig],
     filter_fn: Optional[Callable[[Node], bool]] = None,
 ) -> Optional[List[List[Node]]]:
-  """
-  Find conv + batchnorm parititions
-  Note: This is only used for QAT. In PTQ, batchnorm should already be fused into the conv.
+  """Find conv + batchnorm parititions Note: This is only used for QAT.
+
+  In PTQ, batchnorm should already be fused into the conv.
   """
   return _do_annotate_conv_bn(
       gm, quantization_config, filter_fn, has_relu=False
@@ -433,9 +434,9 @@ def _annotate_conv_bn_relu(
     quantization_config: Optional[QuantizationConfig],
     filter_fn: Optional[Callable[[Node], bool]] = None,
 ) -> Optional[List[List[Node]]]:
-  """
-  Find conv + batchnorm + relu parititions
-  Note: This is only used for QAT. In PTQ, batchnorm should already be fused into the conv.
+  """Find conv + batchnorm + relu parititions Note: This is only used for QAT.
+
+  In PTQ, batchnorm should already be fused into the conv.
   """
   return _do_annotate_conv_bn(gm, quantization_config, filter_fn, has_relu=True)
 
@@ -446,8 +447,8 @@ def _do_annotate_conv_bn(
     filter_fn: Optional[Callable[[Node], bool]],
     has_relu: bool,
 ) -> List[List[Node]]:
-  """
-  Given a function that takes in a `conv_fn` and returns a conv-bn[-relu] pattern,
+  """Given a function that takes in a `conv_fn` and returns a conv-bn[-relu] pattern,
+
   return a list of annotated partitions.
 
   The output of the pattern must include a dictionary from string name to node

ai_edge_torch/quantize/quant_config.py

@@ -13,27 +13,27 @@
 # limitations under the License.
 # ==============================================================================
 
-from dataclasses import dataclass
+import dataclasses
 import enum
 from typing import Optional
 
-from ai_edge_torch.generative.quantize import quant_attrs
 from ai_edge_torch.generative.quantize import quant_recipe
 from ai_edge_torch.quantize import pt2e_quantizer as pt2eq
 
 
-@dataclass(frozen=True)
+@dataclasses.dataclass(frozen=True)
 class QuantConfig:
-  """
+  """Encapsulates a quantization configuration.
+
   Encapsulates all different quantization methods and schemes available for
   models converted with ai_edge_torch.
 
-  Args:
+  Attributes:
     pt2e_quantizer: The instance of PT2EQuantizer used to quantize the model
       with PT2E quantization. This method of quantization is not applicable to
       models created with the Edge Generative API.
-    generative_recipe: Quantization recipe to be applied on a model created
-      with the Edge Generative API.
+    generative_recipe: Quantization recipe to be applied on a model created with
+      the Edge Generative API.
   """
 
   pt2e_quantizer: pt2eq.PT2EQuantizer = None

ai_edge_torch/testing/model_coverage/model_coverage.py

@@ -13,11 +13,11 @@
 # limitations under the License.
 # ==============================================================================
 
-"""Utility Functions to test TFLite models exported from PyTorch"""
+"""Contains utility functions to test TFLite models exported from PyTorch."""
 
 from collections.abc import Callable
 
-from ai_edge_torch.model import Model
+from ai_edge_torch import model
 import numpy as np
 import torch
 from torch.utils import _pytree as pytree
@@ -26,12 +26,20 @@ from torch.utils import _pytree as pytree
 # Utility to flatten the order to make it deterministic.
 # Ordering is done in left-to-right depth-first tree traversal.
 def _flatten(data):
-  out, spec = pytree.tree_flatten(data)
+  out, _ = pytree.tree_flatten(data)
   return out
 
 
 # Convert a Torch Tensor to a numpy array
 def _torch_tensors_to_np(*argv):
+  """Converts a Torch Tensor to a numpy array.
+
+  Args:
+    *argv: A list of torch.tensor or a single torch.tensor.
+
+  Returns:
+    A list of numpy array or a single numpy array.
+  """
   if len(argv) > 1:
     data = list(argv)
   else:
@@ -57,7 +65,7 @@ def _torch_tensors_to_np(*argv):
 
 
 def compare_tflite_torch(
-    edge_model: Model,
+    edge_model: model.Model,
     torch_eval_func: Callable,
     args=None,
     kwargs=None,
@@ -68,15 +76,17 @@ def compare_tflite_torch(
     rtol: float = 1e-5
 ):
   """Compares torch models and TFLite models.
+
   Args:
     edge_model: Serialized ai_edge_torch.model.Model object.
     torch_eval_func: Callable function to evaluate torch model.
-    args: torch.tensor array or a callable to generate a torch.tensor array
-      with random data, to pass into models during inference. (default None).
+    args: torch.tensor array or a callable to generate a torch.tensor array with
+      random data, to pass into models during inference. (default None).
     kwargs: dict of str to torch.tensor, or a callable to generate such.
-    num_valid_inputs: Defines the number of times the random inputs will be generated (if a callable is provided for input_data).
-    signature_name: If provided, specifies the name for the signature of the edge_model to run.
-      Calls the default signature if not provided.
+    num_valid_inputs: Defines the number of times the random inputs will be
+      generated (if a callable is provided for input_data).
+    signature_name: If provided, specifies the name for the signature of the
+      edge_model to run. Calls the default signature if not provided.
     atol: Absolute tolerance (see `numpy.allclose`)
     rtol: Relative tolerance (see `numpy.allclose`)
   """
@@ -118,7 +128,6 @@ def compare_tflite_torch(
   for idx, np_input in enumerate(np_inputs):
     output = get_edge_output(np_input)
     golden_output = np_outputs[idx]
-
     is_output_len_eq = len(golden_output) == len(output)
 
     output = [v.astype(np.float32) for v in output]

ai_edge_torch/version.py

@@ -13,4 +13,4 @@
 # limitations under the License.
 # ==============================================================================
 
-__version__ = "0.2.0.dev20240806"
+__version__ = "0.2.0.dev20240808"

{ai_edge_torch_nightly-0.2.0.dev20240806.dist-info → ai_edge_torch_nightly-0.2.0.dev20240808.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ai-edge-torch-nightly
-Version: 0.2.0.dev20240806
+Version: 0.2.0.dev20240808
 Summary: Supporting PyTorch models with the Google AI Edge TFLite runtime.
 Home-page: https://github.com/google-ai-edge/ai-edge-torch
 Keywords: On-Device ML,AI,Google,TFLite,PyTorch,LLMs,GenAI

ai_edge_torch_nightly-0.2.0.dev20240808.dist-info/RECORD

@@ -0,0 +1,141 @@
+ai_edge_torch/__init__.py,sha256=48qP37uHT90YPs4eIUQxCiWVwqGEX3idCUs6mQKvX1U,1168
+ai_edge_torch/config.py,sha256=PCd9PVrbUNeVIUDFUCnW4goDWU4bjouK28yMYU6VOi0,877
+ai_edge_torch/conftest.py,sha256=r0GTrhMRhlmOGrrkvumHN8hkmyug6WvF60vWq8wRIBI,758
+ai_edge_torch/model.py,sha256=5DYNpFVwvI1w0JbAC1hn83NJVGS1WPX7n742419PMqs,4558
+ai_edge_torch/version.py,sha256=teikEu_awfqM3hip0zXzFvYahGo9H3DdPunvXrDqUgE,706
+ai_edge_torch/_convert/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/_convert/conversion.py,sha256=kcv_QgNgeyDmrqwdzHicGNP68w6zF7GJg7YkMEIXp4Q,3759
+ai_edge_torch/_convert/conversion_utils.py,sha256=Sr8qXVcTwc-ZnZmK7yxVrIOOp1S_vNrwzC0zUvLTI2o,2160
+ai_edge_torch/_convert/converter.py,sha256=ezmaATnQi7NWDo37LUb-hEXtZSmT7_AT6vqXC6Fcq1o,8615
+ai_edge_torch/_convert/signature.py,sha256=zSSG4_u38pOw08pHgXssZDK_EanACDx7lcI-PzBgh5A,3313
+ai_edge_torch/_convert/to_channel_last_io.py,sha256=_31phf7TYgZY2ftpNbrdlB1RhDium1lz_BXEQ6IsMFc,2893
+ai_edge_torch/_convert/fx_passes/__init__.py,sha256=hVuNftOcZIpwkUcPab44mhFmi9Z1f1REV5o3j39Sf-w,2818
+ai_edge_torch/_convert/fx_passes/_pass_base.py,sha256=WVYZuocpygHAzk9u1GNoGowAIOHTlJXyA_NklmYkRms,1672
+ai_edge_torch/_convert/fx_passes/build_aten_composite_pass.py,sha256=3aShe8t2VTtgWntmmVBCOwurFeBo0YI6olkprfLFxSY,9126
+ai_edge_torch/_convert/fx_passes/build_interpolate_composite_pass.py,sha256=izep1DsIz2jwVtqB0CBKg4FcTeEzuPfd2G5Qo4gpTk4,4274
+ai_edge_torch/_convert/fx_passes/canonicalize_pass.py,sha256=8jcKqWzG7p5r3Cu7DXNP-4o4X2bqLaoXY7N6W8QsZXo,1582
+ai_edge_torch/_convert/fx_passes/inject_mlir_debuginfo_pass.py,sha256=WKI8V9-V50agkiNVpBFWWp0BEpUfemdENuN1cEaGD-g,2370
+ai_edge_torch/_convert/fx_passes/optimize_layout_transposes_pass/__init__.py,sha256=lxnoH-WGLeiQIF8XjMGodjiZEFTxucl7g05N7MR9OPk,796
+ai_edge_torch/_convert/fx_passes/optimize_layout_transposes_pass/layout_check.py,sha256=DIfrWDZ1ufAN_uH-oW3k66jTciY7DlLDAb6UKMN14zE,7528
+ai_edge_torch/_convert/fx_passes/optimize_layout_transposes_pass/layout_mark.py,sha256=4RyGUwR22bZqkn_TnptenFJodc_Q43f4_SBG7gmTbos,1621
+ai_edge_torch/_convert/fx_passes/optimize_layout_transposes_pass/layout_rewrite.py,sha256=GcxDxj-5KKWTR5xnRKuhRsb6TDHLCXiPXjGnc_97QXs,12604
+ai_edge_torch/_convert/fx_passes/optimize_layout_transposes_pass/op_func_registry.py,sha256=bsYnudRlXp1PJlu4GF25KSogSkBGQPSaecBrUTONKaw,1031
+ai_edge_torch/_convert/fx_passes/optimize_layout_transposes_pass/pass_body.py,sha256=HXTDEP6_Z0I0s58H6I0yHz9qrkOxptIjKhxywfe8F80,10637
+ai_edge_torch/_convert/fx_passes/optimize_layout_transposes_pass/utils.py,sha256=YLMttMg5PdvXTtQ8lxpKb434UGVvYVALV1-xeuH4UGc,2131
+ai_edge_torch/_convert/fx_passes/optimize_layout_transposes_pass/layout_partitioners/__init__.py,sha256=D8VX8SbCzfoyvPgMFHK7yxD7R-bzLxp2gfdKxgrWekA,742
+ai_edge_torch/_convert/fx_passes/optimize_layout_transposes_pass/layout_partitioners/greedy.py,sha256=L_x8BrF7UDah-SYl-pG11I6CIckdU9kBTUHcmwW4cts,2420
+ai_edge_torch/_convert/fx_passes/optimize_layout_transposes_pass/layout_partitioners/min_cut.py,sha256=mzfL9cf0qBnpmxM_OlMQFvQsEZV2B_Mia9yEJV4J7rI,7135
+ai_edge_torch/_convert/test/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/_convert/test/test_convert.py,sha256=y0ZRivdglGx217rnacze8N6nd7aafk28NkbBFUSa9DQ,13121
+ai_edge_torch/_convert/test/test_convert_composites.py,sha256=CBiOqq-m7QT2ggBI1jBl9MkTIT5d0nK1tA0BUga0LGs,7994
+ai_edge_torch/_convert/test/test_convert_multisig.py,sha256=4jm5blAfzLMjvrJt0ntuG_Fgy4Ie3SoUOGBOy9tf6bg,4725
+ai_edge_torch/_convert/test/test_to_channel_last_io.py,sha256=jLAmyHw5llT2ff8qA8mem3eVN57e_o5EpBnW72ZtP2I,3026
+ai_edge_torch/debug/__init__.py,sha256=N05Mmvi41KgSuK0JhuMejERESgP8QekiGdp9_PEyuKU,742
+ai_edge_torch/debug/culprit.py,sha256=7UYVpVWpiCXbMAyThVtHt_kc_poT7sCTh5UUPvcycgk,14832
+ai_edge_torch/debug/utils.py,sha256=vOAL4t6Lj47uhKapfEsc_WHmvwew3eKO9hSJyzvPXnU,1625
+ai_edge_torch/debug/test/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/debug/test/test_culprit.py,sha256=GjQv4bpz5EVwgxQt7HmpqTzIo_BpsvRmDVWeOmr29HE,3775
+ai_edge_torch/debug/test/test_search_model.py,sha256=3rUSl7pFBfWjK47YhK5B8J1bXrvNhKKIEuNDNfFShHc,1670
+ai_edge_torch/experimental/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/experimental/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/experimental/gemma/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/experimental/gemma/convert_to_tflite.py,sha256=lpiPFSh3SJd6WwuZ0QegSva3__iSz2tUD7L7QfkAe4I,3085
+ai_edge_torch/generative/examples/experimental/gemma/gemma.py,sha256=QoFbUUCTJrW1IYZg0vfb2-K-X0q1-NJFbWNGPQGwBgk,6688
+ai_edge_torch/generative/examples/experimental/phi/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/experimental/phi/convert_to_tflite.py,sha256=DavrdGmqUgoThsGNRv3LXMW5tvJdYEvj66Hf1XRqkXU,3055
+ai_edge_torch/generative/examples/experimental/phi/phi2.py,sha256=u-VJX5mjzQKspXtAhNi53LCITtag-3nCaRTKdk5Z1sc,6231
+ai_edge_torch/generative/examples/experimental/tiny_llama/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/experimental/tiny_llama/convert_to_tflite.py,sha256=xPVvHQjLJHFiRv_-Fy2sDm0Aft7SG8SXiV6o3rF03cQ,3108
+ai_edge_torch/generative/examples/experimental/tiny_llama/tiny_llama.py,sha256=GOLLd9yCBnlNXeW7xrVy1wjOltcTbRdSpiJycbMj8TA,6372
+ai_edge_torch/generative/examples/gemma/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/gemma/convert_to_tflite.py,sha256=w589IJETATd6Z9_1XCIWbrlCV3E92X_5ac3VVCVFXG0,2522
+ai_edge_torch/generative/examples/gemma/gemma.py,sha256=5Dn9JgJiXN-hWGQj9YqCr8Iik8mh5s0dX0VfyY8KDDo,6236
+ai_edge_torch/generative/examples/phi2/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/phi2/convert_to_tflite.py,sha256=ON6zLO-nFS8eJ2yhyWzT5x2Somr-Ca-VjpjT7OGFU10,2506
+ai_edge_torch/generative/examples/phi2/phi2.py,sha256=C_kFYsPrEQ9GJCnc6h-jh8B5qQryvEpI6O6t4FBxg1I,5858
+ai_edge_torch/generative/examples/stable_diffusion/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/stable_diffusion/attention.py,sha256=kDWG6MlIGa89zC5KSRcJlw2c4ITuw8KcchtfmF55f4g,3545
+ai_edge_torch/generative/examples/stable_diffusion/clip.py,sha256=VR09iAnj1e-sr-oam2rh24Wnb_JdZZQvpJIjylfgnS8,4468
+ai_edge_torch/generative/examples/stable_diffusion/convert_to_tflite.py,sha256=7ra36nM5tQwSw-vi6QCFLx5IssZhT-6yVK4H3XsAc4w,5044
+ai_edge_torch/generative/examples/stable_diffusion/decoder.py,sha256=slieF2-QcDCwd4DRZ7snsZIphT97IXpp4plRRsRSwL8,13983
+ai_edge_torch/generative/examples/stable_diffusion/diffusion.py,sha256=7oUIJ6HO0vmlhFdkXpqGm9KTB-eM4Ob9VrHSDlIGFOg,30926
+ai_edge_torch/generative/examples/stable_diffusion/encoder.py,sha256=CAPsW84A8f00nS6fLFeh_XUjCPsDCA5UxHOUsMrLfSU,3450
+ai_edge_torch/generative/examples/stable_diffusion/pipeline.py,sha256=Wc94X_kEnbInTAXFgf-VuCvv1A0HxxWrFZ7Tsq3Z8n0,8662
+ai_edge_torch/generative/examples/stable_diffusion/tokenizer.py,sha256=xychak9hdLd6ieXBYEwrK2BkF8NRZWZSSCijIsESpBA,3420
+ai_edge_torch/generative/examples/stable_diffusion/util.py,sha256=XIXIB0vCvQKOGyIyiZeiIA5DLeSXjkudywvJS4FK7AM,2431
+ai_edge_torch/generative/examples/stable_diffusion/samplers/__init__.py,sha256=uQWKzCD_49ackNFrt50H04dkDXxfAwUCtMWWQre5SVE,830
+ai_edge_torch/generative/examples/stable_diffusion/samplers/k_euler.py,sha256=wBBNM24waZ57M1rXonwesfUkKe9DqpqO3eW6BfZkrD0,2323
+ai_edge_torch/generative/examples/stable_diffusion/samplers/k_euler_ancestral.py,sha256=c89ldwtuQ2_yspGrGa7oh7fsvTt6A86Whxa6fBK9YOQ,2526
+ai_edge_torch/generative/examples/stable_diffusion/samplers/k_lms.py,sha256=ZE6HyOoBJrmTh54KVFf7DjNBnBS0pT4cgviYaq8HGMU,2801
+ai_edge_torch/generative/examples/stable_diffusion/samplers/sampler.py,sha256=RxR5rw0wFFm_5CfAY-3-EIz83vhM9EKye8Bb5zBb0Ok,1341
+ai_edge_torch/generative/examples/t5/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/t5/convert_to_tflite.py,sha256=CZVuNEL8OHPkdsz70WOvNpTJ9LFkiDnlwgJiXfUZCVk,4548
+ai_edge_torch/generative/examples/t5/t5.py,sha256=6Rkisv7UI2w5KV8ogPPzeIiPWYwDLfFfSIncqD7Eenc,20854
+ai_edge_torch/generative/examples/t5/t5_attention.py,sha256=gp7DV8pv4FwICQhYlUYfYZ7BE5jzDIsD_V3a_4-T4Ds,8492
+ai_edge_torch/generative/examples/test_models/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/test_models/toy_model.py,sha256=DhxOrIKe-tilBjbh1q4MsmCmmKMc4c1BPUzhnaJDD6M,3955
+ai_edge_torch/generative/examples/test_models/toy_model_with_external_kv_cache.py,sha256=bW0QB-_h9cfwAQf11AxFxOBq3HrEep_UlpBjXz3JSew,5801
+ai_edge_torch/generative/examples/test_models/toy_model_with_kv_cache.py,sha256=CRja_CT0_eaH16rSDxwHKJS_CGUJMW0Fxd4r45Ii8Uo,4833
+ai_edge_torch/generative/examples/tiny_llama/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/examples/tiny_llama/convert_to_tflite.py,sha256=CLRqO7ycMbpy7J3_Czp1sLx6hcdwGD9zVq04yRba0e8,2550
+ai_edge_torch/generative/examples/tiny_llama/tiny_llama.py,sha256=nu3Il8Vxe7JwM8-AnGNXoGoZ9eVXKHMYEAqVEP-gwe8,5929
+ai_edge_torch/generative/fx_passes/__init__.py,sha256=fmNNXawJ722M4cTUuTx289rT0NHxBEsOy_k8baqCOms,1173
+ai_edge_torch/generative/fx_passes/remove_sdpa_zero_mask_pass.py,sha256=sXis0U4u-RoIp_NyrmWJNnqFqpqRuZOrhfsJIO6rMps,2028
+ai_edge_torch/generative/layers/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/layers/attention.py,sha256=xq10Gw4GudK4M2eY8-H4fi3qmpmZCfE-CziAXDZvqiQ,12177
+ai_edge_torch/generative/layers/attention_utils.py,sha256=2hzBVZvWCqqLfI-f3RJA1hi6T8cuaIJBPt8cdjQCA5s,6420
+ai_edge_torch/generative/layers/builder.py,sha256=JvPmwrG8_M4-kO2MM6sDZhpS32Wx3wVVhlVO4yPJKJ0,4161
+ai_edge_torch/generative/layers/feed_forward.py,sha256=RukSYr9h_DehcYVZWLS_rfCTY73Uj__pTRUatjxJtv8,2788
+ai_edge_torch/generative/layers/kv_cache.py,sha256=Ob8QeXWW5xt-6hcGA0uoC48eRQ8lfvKca8JbWtFx2CE,3082
+ai_edge_torch/generative/layers/model_config.py,sha256=CTvKFwsBR3Rc-Kf73NA7k0799m1WnEvaEBKCnnfNkyo,4961
+ai_edge_torch/generative/layers/normalization.py,sha256=u8lv0p-ktKcRqCDlOqZQa9WQcfDK9JM2IaUQFQdn7xs,1860
+ai_edge_torch/generative/layers/rotary_position_embedding.py,sha256=CZqOoibLcHvUgrgaIIWAlmk3XgE2inzx340MN-npLoU,1347
+ai_edge_torch/generative/layers/scaled_dot_product_attention.py,sha256=6WMe-A5KSSujQcZ34hIeSnnor3AXrw10cQ5FKy-30IU,3390
+ai_edge_torch/generative/layers/unet/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/layers/unet/blocks_2d.py,sha256=4a0wh0co8Avz1wvxS3XqsgrgL5G-X1GSARI5Rj3L-xg,26995
+ai_edge_torch/generative/layers/unet/builder.py,sha256=zAqWXdimmMrQRhmE_t9XkS68mh6PSrzwb-2NZZXrR5I,1901
+ai_edge_torch/generative/layers/unet/model_config.py,sha256=NvBJj09a7ZC-ChGE_ex-_kLnE_fjzrY6txbLSh1pMKA,9208
+ai_edge_torch/generative/quantize/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/quantize/example.py,sha256=mqi3zFUp4w198DGnRkmZCWUZdUXTkvg1_tdTdOk9IkA,1535
+ai_edge_torch/generative/quantize/quant_attrs.py,sha256=n1Fm8BFC8gJa_oiwwAOOghJyHtOXYZ4q-5ZRy4pHrIw,1957
+ai_edge_torch/generative/quantize/quant_recipe.py,sha256=tKnuJq6hPD23JPCB9nPAlE1UHAwdbChkgPShiVaz4CE,5156
+ai_edge_torch/generative/quantize/quant_recipe_utils.py,sha256=4fgmP_GgeiFUOkIaC9ZZXC12eO3DQZdrWDXRz5YXiwU,2270
+ai_edge_torch/generative/quantize/quant_recipes.py,sha256=0Kvr_o7pbMnE8VMe6Ml0FBxkHM6RJ3C14B2I1mjItjc,2030
+ai_edge_torch/generative/quantize/supported_schemes.py,sha256=FjdycEOvxRgBmQdZVufetPvkDoD7rUowIOSKV9oV5Kk,1418
+ai_edge_torch/generative/quantize/ai_edge_quantizer_glue/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/quantize/ai_edge_quantizer_glue/translate_recipe.py,sha256=sSHc_4hUEvi-3KmqbpqWbrRKBjCI1AOctM3dr2EH3vk,5263
+ai_edge_torch/generative/test/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/generative/test/test_experimental_ekv.py,sha256=T5-O2RVLJTH7v9w1_uBfp-Y7o3sdGzYq2Tj2wLRNHyI,4357
+ai_edge_torch/generative/test/test_loader.py,sha256=1ZqAq0HY5uIioumsReOVIsbGBx0WkYcl18PvttdJKrk,3381
+ai_edge_torch/generative/test/test_model_conversion.py,sha256=4RTB1oPA2eWPyuof2-ZB1BxVKzKy5Q9vCux7psmV6zc,7615
+ai_edge_torch/generative/test/test_quantize.py,sha256=3SmJm7Kq98gAneU6IGwwJrJYCVH1qwWR6oUxPfb6qiI,5346
+ai_edge_torch/generative/utilities/__init__.py,sha256=-_jxnnFnCgnTU4oTm4MnRsvL5lqhomBNdFBbqfmfHPo,720
+ai_edge_torch/generative/utilities/loader.py,sha256=XfVRvwvZyQuofctxIedLNDKQrsy9UlRr4wpScZJLWcw,11779
+ai_edge_torch/generative/utilities/stable_diffusion_loader.py,sha256=pKp3AMSbS3otCvgwJRF5M1l4JRNKk-aCKimXzIMSrds,35679
+ai_edge_torch/generative/utilities/t5_loader.py,sha256=jz2qnDtH6oyxcqaBwEVfiiKmq_93LTDeUKNJ2cWpLwg,16856
+ai_edge_torch/hlfb/__init__.py,sha256=sH4um75na-O8tzxN6chFyp6Y4xnexsE7kUQpZySv6dE,735
+ai_edge_torch/hlfb/mark_pattern/__init__.py,sha256=cjTprggj_cuktSCm7-A25e7Shop3k63ylp7sdZmtZ8o,4790
+ai_edge_torch/hlfb/mark_pattern/passes.py,sha256=pjkKcI1nHECPluAt87cFBrt1DP0f3ge7rHq1NhCkBIE,1936
+ai_edge_torch/hlfb/mark_pattern/pattern.py,sha256=uiYRfzD1T8deCEAGfdAFusRbI41m14zeTt0Lz5lNT3M,9808
+ai_edge_torch/hlfb/test/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/hlfb/test/test_mark_pattern.py,sha256=oYB0RPW-tHOwW9gQFH9GtHKO_Mmh1lkoiemXmTfySqc,4383
+ai_edge_torch/hlfb/test/test_stablehlo_composite_builder.py,sha256=3vSX5E9ZFFhTPZZX6TMiAsGa_kaXABbN851bRbTFsC0,8297
+ai_edge_torch/lowertools/__init__.py,sha256=0M9TOR80sS5y6dikOsIFYx0P9IomqAdNIuYpgkP4PcI,693
+ai_edge_torch/lowertools/_shim.py,sha256=ilL7x1ebUBj1clg7bagrX4y_nVSHiGrvDrOVfuTeenE,3039
+ai_edge_torch/lowertools/common_utils.py,sha256=emClsZ_MBlbLG_0BBtyLpkdz4dMWp6SyrNioygRBylk,2973
+ai_edge_torch/lowertools/odml_torch_utils.py,sha256=5XVI2ovptp1wsDZcyyaZDgT4oUa1McOiE-PrKhXNhFo,6316
+ai_edge_torch/lowertools/torch_xla_utils.py,sha256=-g0NldtVOTCQtX3V2XEjuCQO_I52nSNQlu0r_rIS2IE,8635
+ai_edge_torch/quantize/__init__.py,sha256=aB5dXot04bqyUhpsDFvxt9CIi15QAC4euvqOndJ0XLU,714
+ai_edge_torch/quantize/pt2e_quantizer.py,sha256=CKIEhs9jCcna64qj1jFH9zEbMbRdyeGV_TmSqEBPjes,15741
+ai_edge_torch/quantize/pt2e_quantizer_utils.py,sha256=eARD1LxLi5m7Z0n_psAkeX_AtUp4fNkE--oECBfivv4,36208
+ai_edge_torch/quantize/quant_config.py,sha256=U0KisSW-uZkoMJcy-ZP9W57p3tsa594fr9PphCRdO8o,3172
+ai_edge_torch/testing/__init__.py,sha256=hHLluseD2R0Hh4W6XZRIXY_dRQeYudjsrKGf6LZz65g,671
+ai_edge_torch/testing/model_coverage/__init__.py,sha256=5P8J6Zk5YYtDvTBucFvB9NGSRI7Gw_24WnrbhXgycEE,765
+ai_edge_torch/testing/model_coverage/model_coverage.py,sha256=UPB448aMDUyC0HNYVqio2rcJPnDN0tBQMP08J6vPYew,4718
+ai_edge_torch_nightly-0.2.0.dev20240808.dist-info/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
+ai_edge_torch_nightly-0.2.0.dev20240808.dist-info/METADATA,sha256=bzMN4crnlIS6AQjGGcYzSIPuOK-u10kqqBxlVsxCuP4,1885
+ai_edge_torch_nightly-0.2.0.dev20240808.dist-info/WHEEL,sha256=eOLhNAGa2EW3wWl_TU484h7q1UNgy0JXjjoqKoxAAQc,92
+ai_edge_torch_nightly-0.2.0.dev20240808.dist-info/top_level.txt,sha256=5KXRaF2hwkApYxf7Y8y_tVb9aulGTlbOoNdbx1aKRkE,14
+ai_edge_torch_nightly-0.2.0.dev20240808.dist-info/RECORD,,