add basic ruff fixes

Signed-off-by: Keval Morabia <28916987+kevalmorabia97@users.noreply.github.com>

Author: kevalmorabia97

.pre-commit-config.yaml

@@ -79,6 +79,7 @@ repos:
               modelopt/onnx/quantization/ort_patching.py|
               modelopt/torch/_deploy/utils/onnx_utils.py|
               modelopt/torch/export/transformer_engine.py|
+              modelopt/torch/puzzletron/anymodel/models/gpt_oss/gpt_oss_pruned_to_mxfp4.py|
               modelopt/torch/quantization/export_onnx.py|
               modelopt/torch/quantization/plugins/attention.py|
               modelopt/torch/speculative/eagle/utils.py|
@@ -100,8 +101,6 @@ repos:
               examples/speculative_decoding/main.py|
               examples/speculative_decoding/medusa_utils.py|
               examples/speculative_decoding/server_generate.py|
-              examples/puzzletron/evaluation/lm_eval_anymodel.py|
-              modelopt/torch/puzzletron/anymodel/models/gpt_oss/gpt_oss_pruned_to_mxfp4.py|
               experimental/dms/models/qwen3/configuration_qwen3_dms.py|
               experimental/dms/models/qwen3/modeling_qwen3_dms.py|
           )$

modelopt/torch/puzzletron/anymodel/models/nemotron_h/nemotron_h_model_descriptor.py

@@ -151,7 +151,6 @@ def init_rotary_embedding(model, runtime):
         """
         NemotronH has no positional embeddings
         """
-        pass
 
     @staticmethod
     def input_embedding_name():

modelopt/torch/puzzletron/anymodel/models/nemotron_h_v2/nemotron_h_v2_model_descriptor.py

@@ -131,7 +131,6 @@ def init_rotary_embedding(model, runtime):
         """
         NemotronH has no positional embeddings
         """
-        pass
 
     @staticmethod
     def input_embedding_name():

modelopt/torch/puzzletron/anymodel/models/qwen2/qwen2_model_descriptor.py

@@ -144,5 +144,3 @@ class Qwen2FFNIntermediateLayerDescriptor(LlamaFFNIntermediateLayerDescriptor):
 
     Qwen2 uses the same FFN structure as Llama (gate_proj, up_proj, down_proj).
     """
-
-    pass

modelopt/torch/puzzletron/anymodel/models/qwen3_vl/qwen3_vl_model_descriptor.py

@@ -202,7 +202,7 @@ class Qwen3VLExpertRemovalLayerDescriptor(ExpertRemovalLayerDescriptor):
     moe_prefix_name: str = "model.language_model.layers.{layer_idx}.mlp"
     # Router: Qwen3VLMoeTextTopKRouter has self.weight, no bias
     router_weights: List[str] = field(default_factory=lambda: ["gate.weight"])
-    router_biases: List[str] = field(default_factory=lambda: [])
+    router_biases: List[str] = field(default_factory=list)
     # Fused expert format: Qwen3VLMoeTextExperts stores all experts in single tensors
     # with shape [num_experts, ...] instead of separate tensors per expert.
     is_fused_experts: bool = True

modelopt/torch/puzzletron/mip/run_puzzle.py

@@ -20,6 +20,7 @@
 import dataclasses
 import enum
 import json
+import sys
 from collections.abc import Hashable, Iterable
 from copy import deepcopy
 from pathlib import Path
@@ -401,14 +402,14 @@ def _assert_valid_config(args, puzzle_profile):
     missing_args = [arg for arg in required_args if arg not in args or getattr(args, arg) is None]
     if missing_args:
         mprint(f"error: The following arguments are required: {', '.join(missing_args)}")
-        exit(1)
+        sys.exit(1)
 
     # Make sure we have specified subblock_stats_args
     if "subblock_stats_args" not in args and "subblock_stats_args" not in puzzle_profile:
         mprint(
             "error: Must specify `subblock_stats_arrs` in either puzzle_profile or as a commandline arg."
         )
-        exit(1)
+        sys.exit(1)
 
     # Make sure we have specified constraints
     if (
@@ -420,7 +421,7 @@ def _assert_valid_config(args, puzzle_profile):
         mprint(
             "error: Must specify either `mip_constraints` or `human_constraints` in one of puzzle_profile or as a commandline argument."
         )
-        exit(1)
+        sys.exit(1)
 
 
 def _get_minimal_unique_names(dicts: list[dict]) -> list[str]:

modelopt/torch/puzzletron/mip/utils.py

@@ -21,8 +21,6 @@
 class InfeasibleError(Exception):
     """Exception raised when optimization problem is infeasible."""
 
-    pass
-
 
 def sort_replacements(layer_replacements: list[dict]) -> list[dict]:
     """Sort layer replacements by parent layer indices.

modelopt/torch/puzzletron/sewing_kit/core.py

@@ -676,8 +676,6 @@ def forward(
                                 if work is not None:
                                     work.wait()
 
-                        pass
-
                     if len(node.stitches_from) > 0:
                         assert len(peers) == 1, (
                             f"Cannot use multiple peers when using RemoteTarget as a source ({peers=})"

modelopt/torch/puzzletron/sewing_kit/utils.py

@@ -16,9 +16,9 @@
 from __future__ import annotations
 
 import inspect
-from collections.abc import Sequence
 from contextlib import contextmanager
 from typing import (
+    TYPE_CHECKING,
     Any,
     Callable,
     ContextManager,
@@ -43,6 +43,9 @@
 from torch._subclasses import FakeTensor, FakeTensorMode
 from typing_extensions import override
 
+if TYPE_CHECKING:
+    from collections.abc import Sequence
+
 Fn = TypeVar("Fn", bound=Callable)
 
 
@@ -61,11 +64,11 @@ def __call__(self, fn: Fn, disable: bool = False) -> Fn: ...
 
 
 try:
-    dynamo_skip: DynamoSkip = cast(Any, torch._dynamo.decorators).skip
-    dynamo_disable: DynamoDisable = cast(Any, torch._dynamo.decorators).disable
+    dynamo_skip: DynamoSkip = cast("Any", torch._dynamo.decorators).skip
+    dynamo_disable: DynamoDisable = cast("Any", torch._dynamo.decorators).disable
 except:
-    dynamo_skip: DynamoSkip = cast(Any, torch._dynamo.eval_frame).skip
-    dynamo_disable: DynamoDisable = cast(Any, torch._dynamo.eval_frame).disable
+    dynamo_skip: DynamoSkip = cast("Any", torch._dynamo.eval_frame).skip
+    dynamo_disable: DynamoDisable = cast("Any", torch._dynamo.eval_frame).disable
 
 
 TModule = TypeVar("TModule", bound=nn.Module)
@@ -264,7 +267,7 @@ def __new__(cls, elem, device) -> MyFakeTensor:
             dispatch_device=True,
             device_for_backend_keys=device,
         )
-        return cast(MyFakeTensor, self)
+        return cast("MyFakeTensor", self)
 
     @classmethod
     @dynamo_disable

modelopt/torch/puzzletron/subblock_stats/calc_subblock_stats.py

@@ -203,7 +203,6 @@ def calculate_subblock_stats(
         )
 
     if is_calc_runtime:
-        pass
         # TODO: fix
         # from puzzle_tools.calc_subblock_runtime import measure_non_block_runtime_ms
         # non_block_runtime_ms, embedding_runtime_ms, lm_head_runtime_ms = \