hao-ai-lab/FastVideo

Pattern Findings

The scanner identified 9106 distinct pattern matches across 26 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a detector match to inspect in context. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.

Reading the findings table: The Severity column indicates the configured finding weight (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — legacy comment-context matches receive a ×1.5 score weight. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier under the legacy scoring model.

Hyper-Verbose Identifiers4297 hits · 4443 pts

Severity	File	Line	Snippet	Context
LOW	collect_env.py	129	def run_and_parse_first_match(run_lambda, command, regex):	CODE
LOW	collect_env.py	140	def run_and_return_first_line(run_lambda, command):	CODE
LOW	collect_env.py	172	def get_nvidia_driver_version(run_lambda):	CODE
LOW	collect_env.py	281	def summarize_fastvideo_build_flags():	CODE
LOW	collect_env.py	496	def get_cachingallocator_config():	CODE
LOW	collect_env.py	501	def get_cuda_module_loading_config():	CODE
LOW	comfyui/video_generator/video_generator.py	127	def _monitor_for_interruption(self):	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/vsa_utils.py	32	def get_tile_partition_indices(	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/vsa_utils.py	58	def get_reverse_tile_partition_indices(	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/vsa_utils.py	68	def construct_variable_block_sizes(	CODE
LOW⚡	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	521	def block_sparse_attn_from_indices(	CODE
LOW	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	124	def _invert_indices_for_backward(	CODE
LOW	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	177	def _block_sparse_attn_triton_fake(	CODE
LOW	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	199	def block_sparse_attn_backward_triton(	CODE
LOW	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	234	def _block_sparse_attn_backward_triton_fake(	CODE
LOW	…/python/fastvideo_kernel/block_sparse_attn_cute_fwd.py	93	def _choose_q_sparse_block_size(q_len: int, q_tile_size: int = _FA4_Q_BLOCK_SIZE) -> int:	CODE
LOW	…/python/fastvideo_kernel/block_sparse_attn_cute_fwd.py	359	def block_sparse_attn_cute_fwd(	CODE
LOW	…/python/fastvideo_kernel/block_sparse_attn_cute_fwd.py	387	def block_sparse_attn_cute_fwd_bshd(	CODE
LOW	…o-kernel/python/fastvideo_kernel/turbodiffusion_ops.py	207	def _layer_norm_param_fwd_fused(	CODE
LOW	…o-kernel/python/fastvideo_kernel/turbodiffusion_ops.py	307	def _layer_norm_noparam_fwd_fused(	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/vmoba.py	59	def _select_threshold_query_head(gate: torch.Tensor, valid_gate_mask: torch.Tensor, gate_self_chunk_mask: torch.Tensor,	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/vmoba.py	185	def _select_threshold_overall(gate: torch.Tensor, valid_gate_mask: torch.Tensor, gate_self_chunk_mask: torch.Tensor,	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/vmoba.py	264	def _select_threshold_head_global(gate: torch.Tensor, valid_gate_mask: torch.Tensor, gate_self_chunk_mask: torch.Tensor,	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/vmoba.py	806	def test_attn_varlen_moba_speed(batch,	CODE
LOW	…fastvideo_kernel/triton_kernels/fused_compress_topk.py	63	def _fused_block_mean_bwd_kernel(	CODE
LOW	…fastvideo_kernel/triton_kernels/fused_compress_topk.py	290	def _pytorch_topk_mask_fallback(	CODE
LOW	…thon/fastvideo_kernel/triton_kernels/st_attn_triton.py	22	def get_common_autotune_config():	CODE
LOW	…thon/fastvideo_kernel/triton_kernels/st_attn_triton.py	241	def sliding_tile_attention_triton(q: torch.Tensor,	CODE
LOW	…ideo_kernel/triton_kernels/block_sparse_attn_triton.py	703	def triton_block_sparse_attn_forward(q, k, v, q2k_index, q2k_num, variable_block_sizes):	CODE
LOW	…ideo_kernel/triton_kernels/block_sparse_attn_triton.py	755	def triton_block_sparse_attn_backward(do, q, k, v, o, M, q2k_index, q2k_num, k2q_index, k2q_num, variable_block_sizes):	CODE
LOW	…thon/fastvideo_kernel/triton_kernels/attn_qat_train.py	72	def _consumer_blackwell_join_qat_pv_enabled():	CODE
LOW	…thon/fastvideo_kernel/triton_kernels/attn_qat_train.py	274	def _host_descriptor_pre_hook(nargs):	CODE
LOW	…thon/fastvideo_kernel/triton_kernels/attn_qat_train.py	679	def _compute_cross_attn_pointer_offsets(bhid, H, N_CTX_Q, stride_z_q, stride_z_kv, stride_h_q, stride_h_kv):	CODE
LOW	fastvideo-kernel/tests/test_sta.py	76	def test_sliding_tile_attention():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_infer_sparse.py	116	def test_full_lists_reproduce_dense_bitwise() -> None:	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_infer_sparse.py	128	def test_bshd_entry_matches_bhsd() -> None:	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_infer_sparse.py	136	def test_first_visited_block_check_rejects_unanchored_rows() -> None:	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_infer_sparse.py	147	def test_block_list_check_rejects_out_of_bounds_lists() -> None:	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_infer.py	137	def test_cross_attention_unequal_sequence_lengths(q_len: int, kv_len: int):	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	17	def test_smooth_q_fails_before_tensor_or_kernel_work():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	168	def test_qat_attention_forward():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	216	def test_qat_attention_backward():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	303	def test_qat_attention_different_shapes():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	350	def test_qat_attention_non_causal():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	392	def test_qat_attention_causal():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	434	def test_qat_attention_causal_backward():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	501	def test_qat_attention_different_seq_lengths():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	553	def test_qat_attention_different_seq_lengths_backward():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	643	def test_fused_attention_forward():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	690	def test_fused_attention_backward():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	776	def test_fused_attention_different_shapes():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	824	def test_fused_attention_non_causal():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	866	def test_fused_attention_causal():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	908	def test_fused_attention_causal_backward():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	975	def test_fused_vs_qat_attention_forward():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	1024	def test_fused_vs_qat_attention_backward():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	1109	def test_fused_vs_qat_attention_different_shapes():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	1157	def test_fused_vs_qat_attention_non_causal():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	1199	def test_fused_vs_qat_attention_causal():	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	1241	def test_fused_vs_qat_attention_causal_backward():	CODE
4237 more matches not shown…

Decorative Section Separators842 hits · 2840 pts

Severity	File	Line	Snippet	Context
MEDIUM	fastvideo-kernel/CMakeLists.txt	73	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	fastvideo-kernel/CMakeLists.txt	85	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	fastvideo-kernel/CMakeLists.txt	241	# ---------------------------	COMMENT
MEDIUM	fastvideo-kernel/CMakeLists.txt	243	# ---------------------------	COMMENT
MEDIUM	fastvideo-kernel/CMakeLists.txt	373	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	fastvideo-kernel/CMakeLists.txt	375	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	fastvideo-kernel/CMakeLists.txt	444	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	fastvideo-kernel/build.sh	16	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	fastvideo-kernel/build.sh	26	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	10	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	12	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	516	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	518	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	98	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	100	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	144	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	146	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	265	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	267	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	388	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	398	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…ideo_kernel/triton_kernels/block_sparse_attn_triton.py	15	# ──────────────────────────── SPARSE ADDITION BEGIN ───────────────────────────	COMMENT
MEDIUM	…ideo_kernel/triton_kernels/block_sparse_attn_triton.py	17	# ──────────────────────────── SPARSE ADDITION END ─────────────────────────────	COMMENT
MEDIUM	…ideo_kernel/triton_kernels/block_sparse_attn_triton.py	38	# ──────────────────────────── SPARSE ADDITION BEGIN ───────────────────────────	COMMENT
MEDIUM	…ideo_kernel/triton_kernels/block_sparse_attn_triton.py	169	# ──────────────────────────── SPARSE ADDITION END ─────────────────────────────	COMMENT
MEDIUM	…ideo_kernel/triton_kernels/block_sparse_attn_triton.py	702	# ──────────────────────────── SPARSE ADDITION BEGIN ───────────────────────────	COMMENT
MEDIUM	fastvideo-kernel/tests/test_attn_qat_infer.py	53	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	fastvideo-kernel/tests/test_attn_qat_infer.py	55	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	fastvideo-kernel/tests/test_attn_qat_infer.py	93	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	fastvideo-kernel/tests/test_attn_qat_infer.py	95	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1372	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1374	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1379	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1381	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1386	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1388	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1391	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1393	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1399	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1401	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1412	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1415	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1420	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1422	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1485	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1487	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1492	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1494	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1499	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1501	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1506	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1508	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1519	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1521	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1526	# -----------------------------	COMMENT
MEDIUM⚡	fastvideo-kernel/tests/test_attn_qat_train.py	1528	# -----------------------------	COMMENT
MEDIUM	fastvideo-kernel/tests/test_attn_qat_train.py	1454	# -----------------------------	COMMENT
MEDIUM	fastvideo-kernel/tests/test_attn_qat_train.py	1456	# -----------------------------	COMMENT
MEDIUM	fastvideo-kernel/tests/test_attn_qat_train.py	1560	# -----------------------------	COMMENT
MEDIUM	fastvideo-kernel/tests/test_attn_qat_train.py	1562	# -----------------------------	COMMENT
782 more matches not shown…

Unused Imports1915 hits · 1791 pts

Severity	File	Line	Context
LOW	comfyui/__init__.py	1	CODE
LOW	comfyui/__init__.py	1	CODE
LOW	comfyui/video_generator/video_generator.py	1	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/vsa_utils.py	9	CODE
LOW	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	3	CODE
LOW	…ernel/python/fastvideo_kernel/block_sparse_attn_256.py	19	CODE
LOW	…/python/fastvideo_kernel/block_sparse_attn_cute_fwd.py	18	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	1	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	3	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	3	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	3	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	9	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	9	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	14	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	14	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	14	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	20	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	20	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	20	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	20	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	20	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	28	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	31	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	31	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	31	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	31	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	31	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	31	CODE
LOW	…el/python/fastvideo_kernel/block_sparse_attn_varlen.py	8	CODE
LOW	…o-kernel/python/fastvideo_kernel/turbodiffusion_ops.py	1	CODE
LOW	…o-kernel/python/fastvideo_kernel/turbodiffusion_ops.py	3	CODE
LOW	…o-kernel/python/fastvideo_kernel/turbodiffusion_ops.py	6	CODE
LOW	…hon/fastvideo_kernel/triton_kernels/fused_attention.py	10	CODE
LOW	fastvideo-kernel/tests/test_vsa256_forward_vbs.py	3	CODE
LOW	fastvideo-kernel/tests/test_vsa_forward.py	1	CODE
LOW	fastvideo-kernel/tests/test_vsa_forward.py	2	CODE
LOW	fastvideo-kernel/tests/test_vsa128_backward.py	3	CODE
LOW	fastvideo-kernel/tests/test_turbodiffusion.py	3	CODE
LOW	fastvideo-kernel/tests/test_turbodiffusion.py	4	CODE
LOW	fastvideo-kernel/tests/test_vsa.py	2	CODE
LOW	fastvideo-kernel/tests/test_vsa.py	3	CODE
LOW	fastvideo-kernel/tests/test_vsa.py	4	CODE
LOW	fastvideo-kernel/tests/test_vsa.py	5	CODE
LOW	fastvideo-kernel/tests/test_vsa.py	6	CODE
LOW	fastvideo-kernel/tests/test_vsa256_backward.py	9	CODE
LOW	fastvideo-kernel/tests/test_vsa256_forward.py	3	CODE
LOW	fastvideo-kernel/tests/test_vsa_utils.py	11	CODE
LOW	fastvideo-kernel/tests/test_vsa256_triton.py	7	CODE
LOW	fastvideo-kernel/tests/test_vsa256_forward_cross.py	3	CODE
LOW	fastvideo-kernel/benchmarks/bench_vsa.py	12	CODE
LOW	fastvideo-kernel/benchmarks/device_specs.py	10	CODE
LOW	fastvideo-kernel/benchmarks/bench_vsa_combine.py	18	CODE
LOW	fastvideo-kernel/benchmarks/bench_vsa_tile_layout.py	16	CODE
LOW	…stvideo-kernel/benchmarks/bench_fused_compress_topk.py	20	CODE
LOW	fastvideo-kernel/benchmarks/benchmark_attn_qat_train.py	8	CODE
LOW	fastvideo-kernel/attn_qat_infer/__init__.py	16	CODE
LOW	tests/test_fasth3_packaging.py	2	CODE
LOW	tests/local_tests/test_trainer_managed_optimization.py	1	CODE
LOW	tests/local_tests/test_trainer_managed_optimization.py	1	CODE
LOW	…tests/kandinsky6/test_kandinsky6_transformer_parity.py	17	CODE
1855 more matches not shown…

Cross-File Repetition105 hits · 525 pts

Severity	File	Line	Snippet	Context
HIGH	fastvideo-kernel/attn_qat_infer/__init__.py	0	copyright (c) 2025 by sageattention team. licensed under the apache license, version 2.0 (the "license"); you may not us	STRING
HIGH	fastvideo-kernel/attn_qat_infer/api.py	0	copyright (c) 2025 by sageattention team. licensed under the apache license, version 2.0 (the "license"); you may not us	STRING
HIGH	…nel/attn_qat_infer/quantization/bench/bench_quant_q.py	0	copyright (c) 2025 by sageattention team. licensed under the apache license, version 2.0 (the "license"); you may not us	STRING
HIGH	…ernel/attn_qat_infer/quantization/bench/bench_utils.py	0	copyright (c) 2025 by sageattention team. licensed under the apache license, version 2.0 (the "license"); you may not us	STRING
HIGH	…nel/attn_qat_infer/quantization/bench/bench_quant_k.py	0	copyright (c) 2025 by sageattention team. licensed under the apache license, version 2.0 (the "license"); you may not us	STRING
HIGH	…nel/attn_qat_infer/quantization/bench/bench_quant_v.py	0	copyright (c) 2025 by sageattention team. licensed under the apache license, version 2.0 (the "license"); you may not us	STRING
HIGH	docs/inference/add_pipeline.md	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…lines/preprocess/preprocess_pipeline_ode_trajectory.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…video/pipelines/preprocess/preprocess_pipeline_text.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…video/pipelines/preprocess/preprocess_pipeline_base.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…ame2/matrixgame2_preprocess_pipeline_ode_trajectory.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/pipelines/basic/flux_2/flux_2_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…deo/pipelines/basic/hunyuan15/hunyuan15_sr_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…tvideo/pipelines/basic/hunyuan15/hunyuan15_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…eo/pipelines/basic/hunyuan15/hunyuan15_2sr_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…eo/pipelines/basic/hunyuan15/hunyuan15_i2v_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/pipelines/basic/cosmos/cosmos_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/pipelines/basic/gen3c/gen3c_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…elines/basic/turbodiffusion/turbodiffusion_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…es/basic/turbodiffusion/turbodiffusion_i2v_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…tvideo/pipelines/basic/gamecraft/gamecraft_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/pipelines/basic/hunyuan/hunyuan_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/pipelines/basic/wan/wan_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	…stvideo/pipelines/basic/wan/wan_causal_dmd_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/pipelines/basic/wan/wan_i2v_dmd_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/pipelines/basic/wan/wan_v2v_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/pipelines/basic/wan/wan_i2v_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/pipelines/basic/wan/wan_dmd_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/pipelines/basic/longcat/longcat_pipeline.py	0	set up pipeline stages with proper dependency injection.	STRING
HIGH	fastvideo/distributed/parallel_state.py	0	note: we assume that the input tensor is on the same device across all the ranks. note: `dst` is the local rank of the d	STRING
HIGH	…o/distributed/device_communicators/cpu_communicator.py	0	note: we assume that the input tensor is on the same device across all the ranks. note: `dst` is the local rank of the d	STRING
HIGH	…buted/device_communicators/base_device_communicator.py	0	note: we assume that the input tensor is on the same device across all the ranks. note: `dst` is the local rank of the d	STRING
HIGH	fastvideo/distributed/parallel_state.py	0	note: `dst` is the local rank of the destination rank.	STRING
HIGH	…/distributed/device_communicators/cuda_communicator.py	0	note: `dst` is the local rank of the destination rank.	STRING
HIGH	…buted/device_communicators/base_device_communicator.py	0	note: `dst` is the local rank of the destination rank.	STRING
HIGH	…o/distributed/device_communicators/npu_communicator.py	0	note: `dst` is the local rank of the destination rank.	STRING
HIGH	fastvideo/distributed/parallel_state.py	0	sends a tensor to the destination rank in a non-blocking way	STRING
HIGH	…/distributed/device_communicators/cuda_communicator.py	0	sends a tensor to the destination rank in a non-blocking way	STRING
HIGH	…buted/device_communicators/base_device_communicator.py	0	sends a tensor to the destination rank in a non-blocking way	STRING
HIGH	…o/distributed/device_communicators/npu_communicator.py	0	sends a tensor to the destination rank in a non-blocking way	STRING
HIGH	fastvideo/dataset/dataloader/record_schema.py	0	create a record for the parquet dataset with clip features.	STRING
HIGH	…tvideo/pipelines/preprocess/preprocess_pipeline_i2v.py	0	create a record for the parquet dataset with clip features.	STRING
HIGH	…process/matrixgame2/matrixgame2_preprocess_pipeline.py	0	create a record for the parquet dataset with clip features.	STRING
HIGH	fastvideo/training/training_pipeline.py	0	add visualization data to tracker logging and save frames to disk.	STRING
HIGH	…stvideo/training/self_forcing_distillation_pipeline.py	0	add visualization data to tracker logging and save frames to disk.	STRING
HIGH	fastvideo/training/distillation_pipeline.py	0	add visualization data to tracker logging and save frames to disk.	STRING
HIGH	…ning/matrixgame2_self_forcing_distillation_pipeline.py	0	override to properly handle i2v concatenation - call parent first, then concatenate image conditioning.	STRING
HIGH	fastvideo/training/wan_i2v_distillation_pipeline.py	0	override to properly handle i2v concatenation - call parent first, then concatenate image conditioning.	STRING
HIGH	fastvideo/training/wan_i2v_training_pipeline.py	0	override to properly handle i2v concatenation - call parent first, then concatenate image conditioning.	STRING
HIGH	fastvideo/training/matrixgame2_training_pipeline.py	0	override to properly handle i2v concatenation - call parent first, then concatenate image conditioning.	STRING
HIGH	fastvideo/pipelines/composed_pipeline_base.py	0	generate a video or image using the pipeline. args: batch: the batch to generate from. fastvideo_args: the inference arg	STRING
HIGH	…deo/pipelines/basic/hunyuan15/hunyuan15_sr_pipeline.py	0	generate a video or image using the pipeline. args: batch: the batch to generate from. fastvideo_args: the inference arg	STRING
HIGH	…eo/pipelines/basic/hunyuan15/hunyuan15_2sr_pipeline.py	0	generate a video or image using the pipeline. args: batch: the batch to generate from. fastvideo_args: the inference arg	STRING
HIGH	fastvideo/pipelines/stages/denoising.py	0	run the denoising loop. args: batch: the current batch information. fastvideo_args: the inference arguments. returns: th	STRING
HIGH	fastvideo/pipelines/stages/sr_denoising.py	0	run the denoising loop. args: batch: the current batch information. fastvideo_args: the inference arguments. returns: th	STRING
HIGH	fastvideo/pipelines/basic/wan/stages/dmd.py	0	run the denoising loop. args: batch: the current batch information. fastvideo_args: the inference arguments. returns: th	STRING
HIGH	…deo/pipelines/basic/hunyuan15/hunyuan15_sr_pipeline.py	0	wan video-to-video diffusion pipeline implementation. this module contains an implementation of the wan video-to-video d	STRING
HIGH	…tvideo/pipelines/basic/hunyuan15/hunyuan15_pipeline.py	0	wan video-to-video diffusion pipeline implementation. this module contains an implementation of the wan video-to-video d	STRING
HIGH	…eo/pipelines/basic/hunyuan15/hunyuan15_2sr_pipeline.py	0	wan video-to-video diffusion pipeline implementation. this module contains an implementation of the wan video-to-video d	STRING
HIGH	…eo/pipelines/basic/hunyuan15/hunyuan15_i2v_pipeline.py	0	wan video-to-video diffusion pipeline implementation. this module contains an implementation of the wan video-to-video d	STRING
45 more matches not shown…

Empty Error Handlers98 hits · 490 pts

Severity	File	Line	Snippet	Context
HIGH	…tests/kandinsky6/test_kandinsky6_transformer_parity.py	51	except ImportError:	CODE
HIGH	tests/local_tests/glm_image/test_pipeline_parity.py	119	except Exception:	CODE
HIGH	tests/local_tests/flux2/test_flux2_component_parity.py	137	except Exception:	CODE
HIGH	tests/local_tests/sd35/test_sd35_component_parity.py	76	except Exception:	CODE
HIGH	tests/local_tests/helpers/magi_human_upstream.py	137	except Exception:	CODE
HIGH	fastvideo/logger.py	282	except NameError:	CODE
HIGH	fastvideo/utils.py	1154	except Exception: # noqa: BLE001	CODE
HIGH	fastvideo/utils.py	1172	except Exception:	CODE
HIGH	fastvideo/utils.py	1182	except Exception:	CODE
HIGH	fastvideo/attention/utils/flash_attn_default.py	71	except ImportError:	CODE
HIGH	…tvideo/distributed/device_communicators/ulysses_a2a.py	64	except OSError:	CODE
HIGH	fastvideo/dataset/preprocessing_datasets.py	407	except (ValueError, OSError):	CODE
HIGH	fastvideo/training/training_pipeline.py	28	except Exception:	CODE
HIGH	fastvideo/training/training_pipeline.py	618	except Exception:	CODE
HIGH	fastvideo/training/training_utils.py	1656	except Exception:	CODE
HIGH	fastvideo/tests/utils.py	64	except (ImportError, AttributeError):	CODE
HIGH	…video/tests/attention/test_attn_qat_infer_arch_gate.py	203	except RuntimeError:	CODE
HIGH	fastvideo/tests/distributed/test_ulysses_a2a_parity.py	190	except Exception: # noqa: BLE001 - a missing optional backend is a valid fallback	CODE
HIGH	fastvideo/tests/ssim/ci_runner.py	168	except ProcessLookupError:	CODE
HIGH	fastvideo/tests/ssim/ci_runner.py	177	except ProcessLookupError:	CODE
HIGH	…video/tests/entrypoints/test_openai_api_integration.py	96	except requests.ConnectionError:	CODE
HIGH	fastvideo/tests/performance/dashboard.py	55	except (TypeError, ValueError):	CODE
HIGH	fastvideo/tests/performance/seed_baseline.py	349	except OSError:	CODE
HIGH	fastvideo/tests/performance/seed_baseline.py	334	except FileNotFoundError:	CODE
HIGH	fastvideo/tests/performance/seed_baseline.py	340	except FileNotFoundError:	CODE
HIGH	fastvideo/tests/performance/seed_baseline.py	345	except OSError:	CODE
HIGH	fastvideo/tests/modal/ssim_test.py	618	except ProcessLookupError:	CODE
HIGH	fastvideo/tests/modal/ssim_test.py	631	except ProcessLookupError:	CODE
HIGH	…tvideo/tests/mlx/test_mlx_wan22_real_weights_parity.py	61	except Exception:	CODE
HIGH	…ideo/pipelines/basic/magi_human/magi_human_pipeline.py	132	except (FileNotFoundError, json.JSONDecodeError):	CODE
HIGH	fastvideo/models/vision_utils.py	189	except EOFError:	CODE
HIGH	fastvideo/models/dits/matrixgame2/utils.py	328	except Exception:	CODE
HIGH	fastvideo/models/dits/matrixgame2/utils.py	358	except Exception:	CODE
HIGH	fastvideo/models/dits/matrixgame2/utils.py	385	except Exception:	CODE
HIGH	fastvideo/models/dits/matrixgame2/causal_model.py	1316	except AttributeError:	CODE
HIGH	fastvideo/models/loader/component_loader.py	332	except json.JSONDecodeError:	CODE
HIGH	fastvideo/models/loader/component_loader.py	754	except Exception:	CODE
HIGH	fastvideo/models/loader/weight_utils.py	54	except ImportError:	CODE
HIGH	fastvideo/models/loader/fsdp_load.py	647	except Exception: # pragma: no cover - attention stack not importable	CODE
HIGH	fastvideo/models/loader/fsdp_load.py	686	except Exception: # pragma: no cover - flash-attn stack not importable	CODE
HIGH	fastvideo/train/utils/config.py	514	except ValueError:	CODE
HIGH	fastvideo/train/utils/config.py	519	except ValueError:	CODE
HIGH	fastvideo/train/utils/config.py	525	except yaml.YAMLError:	CODE
HIGH	fastvideo/entrypoints/cli/bench_serving.py	602	except (aiohttp.ClientError, asyncio.TimeoutError):	CODE
HIGH	fastvideo/entrypoints/streaming/gpu_pool.py	383	except Exception: # pragma: no cover - best-effort cleanup	CODE
HIGH	fastvideo/entrypoints/streaming/stream.py	183	except BrokenPipeError:	CODE
HIGH	fastvideo/api/overrides.py	89	except ValueError:	CODE
HIGH	fastvideo/api/overrides.py	94	except ValueError:	CODE
HIGH	fastvideo/api/overrides.py	100	except yaml.YAMLError:	CODE
HIGH	fastvideo/eval/metrics/vbench/_grit_helper.py	37	except ImportError:	CODE
HIGH	fastvideo/eval/metrics/vbench/_grit_helper.py	53	except (ImportError, AttributeError):	CODE
HIGH	fastvideo/eval/metrics/vbench/__init__.py	46	except ImportError:	CODE
HIGH	fastvideo/eval/metrics/vbench/__init__.py	61	except ImportError:	CODE
HIGH	fastvideo/eval/io/video.py	55	except (ImportError, OSError, ValueError):	CODE
HIGH	fastvideo/eval/io/video.py	89	except ImportError:	CODE
HIGH	fastvideo/eval/io/video.py	93	except ImportError:	CODE
HIGH	fastvideo/eval/v2a/_av_benchmark_runner.py	85	except ValueError:	CODE
HIGH	fastvideo/worker/multiproc_executor.py	685	except Exception:	CODE
HIGH	fastvideo/mlx_runtime/prompt_enhance.py	388	except (OSError, KeyError, json.JSONDecodeError):	CODE
HIGH	examples/inference/basic/basic_sd35_t2i.py	45	except FileNotFoundError:	CODE
38 more matches not shown…

Excessive Try-Catch Wrapping464 hits · 483 pts

Severity	File	Line	Snippet	Context
LOW	comfyui/video_generator/video_generator.py	150	except Exception as e:	CODE
MEDIUM	comfyui/video_generator/video_generator.py	151	print(f"Error sending interrupt signal: {e}")	CODE
LOW	comfyui/video_generator/video_generator.py	166	except Exception as e:	CODE
MEDIUM	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	15	def _get_sm90_ops():	CODE
LOW⚡	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	18	except Exception:	CODE
LOW	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	85	except Exception:	CODE
LOW	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	117	except Exception as e: # pragma: no cover - environment issue	CODE
LOW	fastvideo-kernel/tests/test_vmoba_correctness.py	85	except Exception as e:	CODE
LOW	fastvideo-kernel/benchmarks/bench_vsa.py	24	except Exception as e: # pragma: no cover	CODE
LOW	fastvideo-kernel/benchmarks/bench_vsa_tile_layout.py	27	except Exception as e: # pragma: no cover	CODE
LOW	…tests/kandinsky6/test_kandinsky6_transformer_parity.py	88	except Exception as exc:	CODE
LOW	tests/local_tests/ltx2/test_ltx2_gemma_parity.py	147	except Exception:	CODE
LOW	tests/local_tests/ltx2/test_ltx2_gemma_encoder.py	61	except Exception as exc:	CODE
LOW⚡	…ests/stable_audio/test_stable_audio_pipeline_parity.py	62	except Exception:	CODE
MEDIUM	…ests/stable_audio/test_stable_audio_pipeline_parity.py	55	def _stable_audio_tools_inference_available() -> bool:	CODE
LOW	…ts/local_tests/stable_audio/test_oobleck_vae_parity.py	49	except Exception:	CODE
MEDIUM	…ts/local_tests/stable_audio/test_oobleck_vae_parity.py	44	def _stable_audio_tools_available() -> bool:	CODE
LOW	…cal_tests/stable_audio/test_stable_audio_a2a_parity.py	34	except Exception:	CODE
MEDIUM	…cal_tests/stable_audio/test_stable_audio_a2a_parity.py	29	def _stable_audio_tools_inference_available() -> bool:	CODE
LOW	…tests/stable_audio/test_oobleck_vae_official_parity.py	45	except Exception:	CODE
MEDIUM	…tests/stable_audio/test_oobleck_vae_official_parity.py	41	def _stable_audio_tools_available() -> bool:	CODE
LOW	tests/local_tests/stable_audio/_stable_audio_helpers.py	39	except Exception:	CODE
LOW	tests/local_tests/glm_image/test_pipeline_parity.py	119	except Exception:	CODE
LOW	tests/local_tests/glm_image/test_pipeline_parity.py	131	except Exception:	CODE
MEDIUM	tests/local_tests/glm_image/test_pipeline_parity.py	30	def _upstream_glm_image_available() -> bool:	CODE
LOW	tests/local_tests/flux2/test_flux2_component_parity.py	137	except Exception:	CODE
LOW	tests/local_tests/flux2/test_flux2_component_parity.py	181	except Exception:	CODE
LOW	tests/local_tests/flux2/test_flux2_component_parity.py	188	except Exception:	CODE
LOW	…sts/local_tests/zimage/test_zimage_scheduler_parity.py	61	except Exception as exc:	CODE
LOW	tests/local_tests/zimage/test_zimage_vae_parity.py	67	except Exception as exc:	CODE
LOW	…_world/test_dreamx_world_camera_conditioning_parity.py	39	except Exception as exc: # noqa: BLE001 - local parity should skip missing reference deps.	CODE
LOW	…reamx_world/test_dreamx_world_ar_transformer_parity.py	59	except Exception as exc: # noqa: BLE001	CODE
LOW⚡	…/dreamx_world/test_dreamx_world_text_encoder_parity.py	98	except Exception as exc: # noqa: BLE001	CODE
LOW	…cal_tests/dreamx_world/test_dreamx_world_vae_parity.py	161	except Exception as exc: # noqa: BLE001	CODE
LOW⚡	…s/dreamx_world/test_dreamx_world_transformer_parity.py	114	except Exception as exc: # noqa: BLE001 - local parity should skip missing refs.	CODE
LOW	…s/dreamx_world/test_dreamx_world_transformer_parity.py	209	except Exception as exc: # noqa: BLE001	CODE
LOW	tests/local_tests/kandinsky6_sr/reference.py	157	except Exception as exc: # missing optional dependency of the reference on this machine	CODE
LOW	tests/local_tests/gen3c/test_gen3c.py	383	except Exception as e:	CODE
LOW	…magi_human/test_magi_human_sa_audio_official_parity.py	71	except Exception:	CODE
LOW	…cal_tests/magi_human/test_magi_human_distill_parity.py	55	except Exception:	CODE
LOW	tests/local_tests/magi_human/test_magi_human_parity.py	60	except Exception:	CODE
LOW	…al_tests/magi_human/test_magi_human_sa_audio_parity.py	50	except Exception:	CODE
LOW	…al_tests/magi_human/test_magi_human_pipeline_parity.py	84	except Exception:	CODE
LOW	…al_tests/magi_human/test_magi_human_pipeline_parity.py	119	except Exception as exc:	CODE
LOW	…al_tests/magi_human/test_magi_human_pipeline_parity.py	167	except Exception:	CODE
LOW	…s/local_tests/magi_human/test_magi_human_vae_parity.py	64	except Exception as exc:	CODE
LOW	…cal_tests/magi_human/test_magi_human_pipeline_smoke.py	176	except Exception:	CODE
LOW	…cal_tests/magi_human/test_magi_human_t5gemma_parity.py	48	except Exception:	CODE
LOW	…/magi_human/test_magi_human_sr1080p_pipeline_parity.py	60	except Exception:	CODE
LOW	…s/magi_human/test_magi_human_sr540p_pipeline_parity.py	49	except Exception:	CODE
LOW	tests/local_tests/gamecraft/test_gamecraft_parity.py	141	except Exception as e:	CODE
LOW⚡	…ocal_tests/gamecraft/test_gamecraft_pipeline_parity.py	300	except Exception as e:	CODE
LOW	…sts/local_tests/gamecraft/test_gamecraft_vae_parity.py	161	except Exception:	CODE
LOW	…ocal_tests/gamecraft/test_gamecraft_encoders_parity.py	99	except Exception as e:	CODE
LOW	…ocal_tests/gamecraft/test_gamecraft_encoders_parity.py	166	except Exception as e:	CODE
LOW	…ocal_tests/gamecraft/test_gamecraft_encoders_parity.py	237	except Exception as e:	CODE
LOW	tests/local_tests/sd35/test_sd35_component_parity.py	76	except Exception:	CODE
LOW	tests/local_tests/sd35/test_sd35_component_parity.py	120	except Exception:	CODE
LOW	tests/local_tests/sd35/test_sd35_component_parity.py	127	except Exception:	CODE
LOW	…/kandinsky5/test_kandinsky5_lite_transformer_parity.py	42	except Exception as exc:	CODE
404 more matches not shown…

Deep Nesting429 hits · 406 pts

Severity	File	Line	Context
LOW	collect_env.py	180	CODE
LOW	collect_env.py	394	CODE
LOW	comfyui/video_generator/inference_args.py	46	CODE
LOW	comfyui/video_generator/video_generator.py	127	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/vmoba.py	508	CODE
LOW	…thon/fastvideo_kernel/triton_kernels/st_attn_triton.py	97	CODE
LOW	fastvideo-kernel/tests/utils.py	27	CODE
LOW	fastvideo-kernel/tests/test_block_sparse_bwd_sm100a.py	101	CODE
LOW	…ernel/attn_qat_infer/quantization/bench/bench_utils.py	95	CODE
LOW	tests/test_block_sparse_sm100a.py	61	CODE
LOW	tests/local_tests/ltx2/test_ltx2_vae.py	34	CODE
LOW	tests/local_tests/ltx2/test_ltx2_gemma_encoder.py	23	CODE
LOW	tests/local_tests/ltx2/test_ltx2_audio_vae.py	38	CODE
LOW	tests/local_tests/flux2/debug_flux2_klein_tp_trace.py	209	CODE
LOW	…ts/local_tests/pipelines/test_flux2_pipeline_parity.py	204	CODE
LOW	…al_tests/magi_human/test_magi_human_pipeline_parity.py	311	CODE
LOW	tests/local_tests/gamecraft/test_gamecraft_parity.py	207	CODE
LOW	tests/local_tests/gamecraft/test_gamecraft_parity.py	453	CODE
LOW	…ts/local_tests/minimax_h3/test_mlx_audio_vae_parity.py	58	CODE
LOW	…ills/add-model-08-trace/templates/block_trace_debug.py	86	CODE
LOW	…ills/add-model-08-trace/templates/block_trace_debug.py	242	CODE
LOW	…d-model-07-conversion/templates/family_to_diffusers.py	94	CODE
LOW	docs/generate_examples.py	91	CODE
LOW	docs/generate_examples.py	659	CODE
LOW	docs/generate_examples.py	747	CODE
LOW	docs/generate_examples.py	487	CODE
LOW	fastvideo/fastvideo_args.py	97	CODE
LOW	fastvideo/fastvideo_args.py	989	CODE
LOW	fastvideo/fastvideo_args.py	1411	CODE
LOW	fastvideo/logger.py	249	CODE
LOW	fastvideo/utils.py	1013	CODE
LOW	fastvideo/utils.py	1237	CODE
LOW	fastvideo/utils.py	158	CODE
LOW	fastvideo/utils.py	199	CODE
LOW	fastvideo/utils.py	327	CODE
LOW	fastvideo/utils.py	364	CODE
LOW	fastvideo/forward_context.py	54	CODE
LOW	fastvideo/attention/backends/bsa_attn.py	206	CODE
LOW	fastvideo/attention/backends/bsa_attn.py	485	CODE
LOW	fastvideo/attention/backends/video_sparse_attn_h3.py	723	CODE
LOW	fastvideo/platforms/rocm.py	79	CODE
LOW	fastvideo/platforms/cuda.py	169	CODE
LOW	fastvideo/platforms/cuda.py	408	CODE
LOW	fastvideo/layers/linear.py	526	CODE
LOW	fastvideo/layers/linear.py	830	CODE
LOW	fastvideo/layers/quantization/nvfp4_config.py	669	CODE
LOW	fastvideo/layers/quantization/nvfp4_config.py	756	CODE
LOW	fastvideo/distributed/parallel_state.py	1071	CODE
LOW	fastvideo/distributed/parallel_state.py	450	CODE
LOW	…tvideo/distributed/device_communicators/ulysses_a2a.py	482	CODE
LOW	…tvideo/distributed/device_communicators/ulysses_a2a.py	214	CODE
LOW	fastvideo/dataset/v2a_feature_cache.py	51	CODE
LOW	fastvideo/dataset/validation_dataset.py	19	CODE
LOW	fastvideo/dataset/validation_dataset.py	99	CODE
LOW	fastvideo/dataset/utils.py	83	CODE
LOW	fastvideo/dataset/utils.py	164	CODE
LOW	fastvideo/dataset/parquet_dataset_map_style.py	113	CODE
LOW	fastvideo/dataset/parquet_dataset_iterable_style.py	124	CODE
LOW	fastvideo/dataset/parquet_dataset_iterable_style.py	31	CODE
LOW	fastvideo/training/training_pipeline.py	516	CODE
369 more matches not shown…

Modern Structural Boilerplate300 hits · 302 pts

Severity	File	Line	Snippet	Context
LOW	comfyui/__init__.py	4	__all__ = ['NODE_CLASS_MAPPINGS', 'NODE_DISPLAY_NAME_MAPPINGS', 'WEB_DIRECTORY']	CODE
LOW	comfyui/video_generator/video_generator.py	32	def update_config_from_args(config: Any, args_dict: dict[str, Any]) -> None:	CODE
LOW	fastvideo-kernel/python/fastvideo_kernel/__init__.py	40	__all__ = [	CODE
LOW	…ython/fastvideo_kernel/block_sparse_attn_bwd_sm100a.py	38	def set_extension(module) -> None:	CODE
LOW	…fastvideo_kernel/triton_kernels/fused_compress_topk.py	19	logger = logging.getLogger(__name__)	CODE
LOW	…hon/fastvideo_kernel/triton_kernels/fused_attention.py	55	__all__ = ["attention"]	CODE
LOW	fastvideo-kernel/benchmarks/bench_vsa.py	28	def set_seed(seed: int = 42) -> None:	CODE
LOW	fastvideo-kernel/benchmarks/bench_vsa_tile_layout.py	35	def set_seed(seed: int = 42) -> None:	CODE
LOW	…stvideo-kernel/benchmarks/bench_fused_compress_topk.py	36	def set_seed(seed: int = 42) -> None:	CODE
LOW	tests/local_tests/flux2/debug_flux2_klein_tp_trace.py	105	def _set_submodule(model: torch.nn.Module, name: str, module: torch.nn.Module) -> None:	CODE
LOW	…l_tests/transformers/test_ltx2_5_transformer_parity.py	40	def _setup_test_environment(monkeypatch: pytest.MonkeyPatch) -> None:	CODE
LOW	fastvideo/envs.py	210	def set_external(name: str, value: str) -> None:	CODE
LOW	fastvideo/registry.py	1606	__all__ = [	CODE
LOW	fastvideo/__init__.py	6	__all__ = ["VideoGenerator", "PipelineConfig", "SamplingParam", "__version__"]	CODE
LOW	fastvideo/utils.py	991	def set_random_seed(seed: int) -> None:	CODE
LOW	fastvideo/profiler.py	128	def set_global_controller(controller: TorchProfilerController \| None) -> None:	CODE
LOW	fastvideo/profiler.py	360	def _set_collection(self, enabled: bool) -> None:	CODE
LOW	fastvideo/attention/__init__.py	7	__all__ = [	CODE
LOW	fastvideo/attention/layer.py	104	def _set_compile_forward_enabled(self, enabled: bool) -> None:	CODE
LOW	…deo/attention/backends/video_sparse_attn_h3_scatter.py	70	__all__ = ["HAVE_TRITON", "scatter_rows_heads_first", "supports_heads_first_scatter"]	CODE
LOW	fastvideo/platforms/__init__.py	219	__all__ = ['Platform', 'PlatformEnum', 'current_platform', "_init_trace"]	CODE
LOW	fastvideo/layers/mxfp8linear.py	289	__all__ = [	CODE
LOW	fastvideo/layers/pdd.py	300	__all__ = [	CODE
LOW	fastvideo/layers/quantization/fp8_qat_train_config.py	27	logger = logging.getLogger(__name__)	CODE
LOW	fastvideo/layers/quantization/mxfp8_config.py	143	__all__ = [	CODE
LOW	fastvideo/layers/quantization/nvfp4_qat_train_config.py	29	logger = logging.getLogger(__name__)	CODE
LOW	fastvideo/layers/quantization/fp8_config.py	30	logger = logging.getLogger(__name__)	CODE
LOW	fastvideo/layers/quantization/fp8_config.py	291	__all__ = [	CODE
LOW	fastvideo/layers/quantization/nvfp4_config.py	40	logger = logging.getLogger(__name__)	CODE
LOW	fastvideo/layers/quantization/nvfp4_config.py	814	__all__ = [	CODE
LOW	fastvideo/layers/quantization/nvfp4_qat_config.py	46	logger = logging.getLogger(__name__)	CODE
LOW	fastvideo/layers/quantization/nvfp4_qat_config.py	236	__all__ = [	CODE
LOW	fastvideo/distributed/__init__.py	11	__all__ = [	CODE
LOW	…deo/distributed/device_communicators/pynccl_wrapper.py	306	__all__ = [	CODE
LOW	…deo/distributed/device_communicators/pyhccl_wrapper.py	192	__all__ = [	CODE
LOW	fastvideo/logging_utils/__init__.py	5	__all__ = [	CODE
LOW	fastvideo/dataset/v2a_feature_cache.py	126	__all__ = ["V2AFeatureShardWriter"]	CODE
LOW	fastvideo/dataset/__init__.py	45	__all__ = [	CODE
LOW	fastvideo/dataset/minimax_h3_ref2va_dataset.py	480	__all__ = [	CODE
LOW	fastvideo/dataset/mmaudio_feature_dataset.py	322	__all__ = [	CODE
LOW	fastvideo/dataset/vggsound.py	118	__all__ = ["VGGSoundDataset"]	CODE
LOW	fastvideo/training/training_pipeline.py	89	def set_schemas(self) -> None:	CODE
LOW	fastvideo/training/__init__.py	6	__all__ = [	CODE
LOW	fastvideo/tests/contract/test_dynamo_shape.py	318	__all__ = [	CODE
LOW	fastvideo/tests/golden_gate/test_longcat.py	25	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_gamecraft.py	23	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_zimage.py	23	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_wan_t2v.py	19	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_matrixgame.py	34	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_sd35.py	23	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_kandinsky5.py	26	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_lingbot.py	23	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_wan_vae.py	14	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_glm_image.py	23	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_gen3c.py	20	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_wan_denoising.py	15	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_flux.py	22	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_flux2_klein.py	25	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_stable_audio.py	23	__all__ = ["distributed_runtime"]	CODE
LOW	fastvideo/tests/golden_gate/test_wan_causal.py	17	__all__ = ["distributed_runtime"]	CODE
240 more matches not shown…

Cross-Language Confusion53 hits · 260 pts

Severity	File	Line	Snippet	Context
HIGH	docs/check_cookbook_assets.py	6	PREPARE = "Run npm ci --prefix docs && npm run build:catalog --prefix docs from the repository root."	CODE
HIGH⚡	fastvideo/tests/contract/test_ci_aggregate_status.py	15	assert "s.context === context && s.state === 'failure'" in source	CODE
HIGH	fastvideo/tests/contract/test_ci_test_collection.py	169	assert "group: merge-gate-${{ inputs.pr_number \|\| github.event.pull_request.number }}" in trigger_job	CODE
HIGH	fastvideo/tests/stages/kandinsky6_sr/k6_sr_release.py	15	FLOW_MATCHING_SCHEDULER_CONFIG = json.loads(r'''{"_class_name":"FlowMatchEulerDiscreteScheduler","_diffusers_version":"0	CODE
HIGH	fastvideo/tests/stages/kandinsky6_sr/k6_sr_release.py	33	RELEASE_LATENT_UPSCALER_CONFIG = json.loads(r'''{"models":[{"model":{"architecture":"multi_scale","bare_stem":true,"bott	CODE
HIGH	…stvideo/tests/dataset/test_hunyuan15_parquet_schema.py	58	assert table.schema.equals(pyarrow_schema_t2v)	CODE
HIGH	fastvideo/tests/dataset/test_preprocess_hunyuan15.py	427	assert table.schema.equals(pyarrow_schema_t2v_dual_text)	CODE
HIGH	fastvideo/tests/performance/dashboard.py	314	f"({item['non_null']}/{item['records']} non-null)")	CODE
HIGH⚡	fastvideo/tests/modal/launch_l40s_job.py	331	"source $HOME/.local/bin/env 2>/dev/null \|\| true; "	CODE
HIGH⚡	fastvideo/tests/modal/launch_l40s_job.py	332	"source /opt/venv/bin/activate 2>/dev/null \|\| true; "	CODE
HIGH⚡	fastvideo/tests/modal/launch_l40s_job.py	341	"source $HOME/.local/bin/env 2>/dev/null \|\| true; "	CODE
HIGH⚡	fastvideo/tests/modal/launch_l40s_job.py	342	"source /opt/venv/bin/activate 2>/dev/null \|\| true; "	CODE
HIGH	fastvideo/tests/modal/pr_test.py	62	"nodejs").run_commands("node --version && npm --version"))	CODE
HIGH	fastvideo/tests/modal/pr_test.py	253	"export HF_HOME='/root/data/.cache' && hf auth login --token $HF_API_KEY && bash .buildkite/scripts/lanes/encode	CODE
HIGH	fastvideo/tests/modal/pr_test.py	264	"export HF_HOME='/root/data/.cache' && hf auth login --token $HF_API_KEY && bash .buildkite/scripts/lanes/vae.sh	CODE
HIGH	fastvideo/tests/modal/pr_test.py	278	"export HF_HOME='/root/data/.cache' && hf auth login --token $HF_API_KEY && bash .buildkite/scripts/lanes/golden	CODE
HIGH	fastvideo/tests/modal/pr_test.py	318	run_test("wandb login $WANDB_API_KEY && FASTVIDEO_FA4=0 pytest ./fastvideo/tests/training/VSA -srP")	CODE
HIGH	fastvideo/tests/modal/pr_test.py	381	trap "kill $MOCK_SERVER_PID 2>/dev/null \|\| true" EXIT	CODE
HIGH⚡	fastvideo/tests/modal/pr_test.py	453	"export HF_HOME='/root/data/.cache' && hf auth login --token $HF_API_KEY && bash .buildkite/scripts/lanes/eval.s	STRING
HIGH⚡	fastvideo/tests/modal/pr_test.py	460	run_test("hf auth login --token $HF_API_KEY && pytest ./fastvideo/tests/lora_extraction/test_lora_extraction.py")	STRING
HIGH	fastvideo/tests/modal/pr_test.py	499	"python ./fastvideo/tests/performance/dashboard.py \|\| true; "	STRING
HIGH	fastvideo/tests/modal/pr_test.py	501	"cat /tmp/gpu_telemetry.csv \|\| true; "	STRING
HIGH	fastvideo/tests/modal/pr_test.py	514	"export HF_HOME='/root/data/.cache' && hf auth login --token $HF_API_KEY && pytest ./fastvideo/tests/entrypoints	STRING
HIGH	fastvideo/tests/modal/ssim_test.py	457	return f"git fetch --prune origin refs/pull/{pr_id}/head && git checkout FETCH_HEAD"	CODE
HIGH	fastvideo/tests/modal/ssim_test.py	553	command = f"set -euo pipefail && source $HOME/.local/bin/env && source /opt/venv/bin/activate && {pytest_command}"	STRING
HIGH	…o/pipelines/preprocess/preprocess_minimax_h3_ref2va.py	470	if not table.schema.equals(pyarrow_schema_minimax_h3_ref2va, check_metadata=False):	CODE
HIGH	fastvideo/pipelines/basic/mmaudio/stages.py	144	source.push(frame)	CODE
HIGH	fastvideo/pipelines/basic/mmaudio/stages.py	149	source.push(None)	CODE
HIGH	fastvideo/models/dits/wan_animate_face.py	53	# motion_encoder_channel_sizes: null, which means "use this table" (it is the	COMMENT
HIGH	fastvideo/models/vaes/minimax_h3_parallel.py	373	assembler.push(clip[:, :, vae.frame_pre_padding:].contiguous())	CODE
HIGH	fastvideo/models/vaes/minimax_h3_parallel.py	431	assembler.push(gathered.narrow(2, slot * segment_frames, segment_frames))	CODE
HIGH	fastvideo/eval/metrics/audio/kl_divergence/metric.py	1	"""PaSST KL divergence ``KL(gt \|\| pred)`` on AudioSet-527 logits.	STRING
HIGH	fastvideo/eval/metrics/audio/kl_divergence/metric.py	64	"""``KL(gt \|\| pred)`` over class-probability softmax. Matches av-benchmark."""	STRING
HIGH	fastvideo/eval/metrics/audio/kl_divergence/metric.py	75	"""``KL(gt \|\| pred)`` over per-class Bernoulli sigmoid (multi-label variant)."""	STRING
HIGH	fastvideo/eval/metrics/audio/kl_divergence/metric.py	86	"""PaSST KL divergence ``KL(gt \|\| pred)`` on AudioSet-527 logits.	STRING
HIGH	fastvideo/eval/metrics/optical_flow/_shared.py	76	"""KL(P_a \|\| P_b) over a joint (angle, log-magnitude) histogram."""	STRING
HIGH	fastvideo/mlx_runtime/minimax_h3_vsa_simd.py	83	if (active && g0 < q_valid) {	CODE
HIGH	fastvideo/mlx_runtime/minimax_h3_vsa_simd.py	86	if (active && g1 < q_valid) {	CODE
HIGH	fastvideo/mlx_runtime/minimax_h3_vsa_simd.py	113	if (gtok < k_valid && gtok < TILE && col < D) {	CODE
HIGH	fastvideo/mlx_runtime/minimax_h3_vsa_simd.py	136	if (grow < q_valid && gtok < k_valid && gtok < TILE) {	CODE
HIGH	fastvideo/mlx_runtime/minimax_h3_vsa_simd.py	159	if (gtok < k_valid && gtok < TILE && col < D) {	CODE
HIGH	examples/inference/basic/basic_gen3c.py	12	sudo apt-get update && sudo apt-get install -y libgl1 libglib2.0-0 libsm6 libxext6 libxrender1	STRING
HIGH	examples/inference/basic/basic_fasth3.py	273	"(`cd fastvideo-kernel && CMAKE_ARGS='-DFASTVIDEO_KERNEL_BUILD_TK=ON' ./build.sh`), or pass "	CODE
HIGH	…/inference/gradio/local/gradio_local_demo_ltx2_3/ui.py	1478	if (document.activeElement && document.activeElement !== document.body) return true;	CODE
HIGH	…/inference/gradio/local/gradio_local_demo_ltx2_3/ui.py	1489	let releaseTimer = null;	CODE
HIGH⚡	…/inference/gradio/local/gradio_local_demo_ltx2_3/ui.py	1502	const scroller = document.scrollingElement \|\| document.documentElement;	CODE
HIGH⚡	…/inference/gradio/local/gradio_local_demo_ltx2_3/ui.py	1504	const atTop = scroller.scrollTop <= 0 && event.deltaY < 0;	CODE
HIGH⚡	…/inference/gradio/local/gradio_local_demo_ltx2_3/ui.py	1505	const atBottom = scroller.scrollTop >= maxScroll - 1 && event.deltaY > 0;	CODE
HIGH	scripts/benchmarks/minimax_h3_pro6000/app.py	25	.run_commands("cd /src/fastvideo && UV_TORCH_BACKEND=cu130 uv pip install --system -e . --no-sources")	CODE
HIGH	scripts/benchmarks/minimax_h3_pro6000/app.py	49	"ln -s /cutlass /src/fastvideo/fastvideo-kernel/include/cutlass && mkdir -p /vol/wheels/cu130")	CODE
HIGH	scripts/benchmarks/minimax_h3_pro6000/app.py	54	_sh("cd /src/fastvideo/fastvideo-kernel && pip wheel . --no-build-isolation --no-deps -w /vol/wheels/cu130",	CODE
HIGH	scripts/benchmarks/minimax_h3_pro6000/app.py	105	"cd /src/fastvideo && python scripts/checkpoint_conversion/convert_minimax_h3_text_encoder_nvfp4.py "	CODE
HIGH	scripts/benchmarks/minimax_h3_pro6000/app.py	118	"cd /src/fastvideo && python scripts/checkpoint_conversion/convert_minimax_h3_modelopt_nvfp4_dit.py "	CODE

Docstring Block Structure40 hits · 200 pts

Severity	File	Line	Snippet	Context
HIGH	fastvideo/pipelines/stages/validators.py	338	Add a validation check for a field. Args: field_name: Name of the field being chec	STRING
HIGH	…video/pipelines/preprocess/preprocess_pipeline_base.py	67	Create a record for the Parquet dataset using a generic schema-based approach. Args: prepro	STRING
HIGH	fastvideo/models/vision_utils.py	197	Load frames from a video file using ffmpeg. Args: video_path: Path to the video file Returns:	STRING
HIGH	fastvideo/models/loader/fsdp_load.py	911	Converting full state dict into a sharded state dict and loading it into FSDP model (if training) or normal hug	STRING
HIGH	fastvideo/models/encoders/qwen2_5_vl_custom.py	1534	Calculate the 3D rope index based on image and video's temporal, height and width in LLM. Explanation:	STRING
HIGH	fastvideo/models/encoders/qwen2_5_vl_custom.py	1973	Calculate the 3D rope index based on image and video's temporal, height and width in LLM. Explanation:	STRING
HIGH	fastvideo/mlx_runtime/checkpoint.py	53	Return the manifest name for a supported MLX data type. Parameters: dtype: The MLX data type to convert.	STRING
HIGH	fastvideo/mlx_runtime/checkpoint.py	180	Reconstruct an MLXWanDiT model from a versioned checkpoint. Parameters: checkpoint_dir (str \| Path): D	STRING
HIGH	fastvideo/mlx_runtime/memory.py	60	Convert a positive memory limit from GiB to bytes. Parameters: value (float \| None): Memory limit in G	STRING
HIGH	fastvideo/mlx_runtime/memory.py	101	Set a PyTorch MPS watermark environment variable. Parameters: name (str): Name of the environment variable	STRING
HIGH	fastvideo/mlx_runtime/refine.py	255	Upsample 5-D latent arrays spatially using nearest-neighbor or bilinear interpolation. Parameters: latents	STRING
HIGH	fastvideo/mlx_runtime/refine.py	340	Upsample clean latents spatially and optionally mix them with Gaussian noise. Parameters: clean_latent	STRING
HIGH	fastvideo/mlx_runtime/refine.py	374	Derive the refinement noise level from the first refinement timestep. Parameters: schedule (MLXDMDSchedule)	STRING
HIGH	fastvideo/mlx_runtime/refine.py	395	Derive stage-2 timesteps from the stage-1 DMD grid. The stage-2 pass must start below full noise, otherwise the h	STRING
HIGH	fastvideo/mlx_runtime/refine.py	526	Run base denoising and, when enabled, spatial refinement denoising. Parameters: dit: DiT callable used	STRING
HIGH	fastvideo/mlx_runtime/sampling.py	112	Compute one DMD sampling update, optionally re-noising the clean latent prediction. Args: latents: Ret	STRING
HIGH	fastvideo/mlx_runtime/fast_spatial.py	130	Build a plan for reduced-resolution denoising followed by pixel-space upsampling. Parameters: upsample	STRING
HIGH	fastvideo/mlx_runtime/wan_vae.py	89	Return a validated TAEHV checkpoint for the specified latent channel count. Parameters: z_dim (int): N	STRING
HIGH	fastvideo/mlx_runtime/wan_vae.py	411	Decode latent representations with the MLX TAEHV decoder. Parameters: latents_np (np.ndarray): Latents	STRING
HIGH	fastvideo/mlx_runtime/wan_vae.py	484	Decode latent video frames and export them as an MP4 file. Parameters: latents_np (np.ndarray): Latent vide	STRING
HIGH	fastvideo/mlx_runtime/prompt_enhance.py	104	Normalize a user prompt for enhancement. Parameters: prompt (str): User-provided prompt text. Ret	STRING
HIGH	fastvideo/mlx_runtime/prompt_enhance.py	179	Enhance a user prompt with a locally hosted mlx-lm instruction model. Parameters: prompt (str): The pr	STRING
HIGH	fastvideo/mlx_runtime/prompt_enhance.py	277	Enhance a prompt using the selected backend, falling back to a deterministic template when configured for automatic sele	STRING
HIGH	fastvideo/mlx_runtime/frame_upsample.py	39	Map a pixel upsample mode name onto its OpenCV interpolation flag. Parameters: mode (str): One of :dat	STRING
HIGH	fastvideo/mlx_runtime/frame_upsample.py	92	Resample one decoded RGB frame to the target pixel size. Parameters: frame (np.ndarray): HxWx3 uint8 R	STRING
HIGH	fastvideo/mlx_runtime/windowed_attention.py	35	Determine the attention scaling factor from an explicit value or head dimension. Parameters: head_dim	STRING
HIGH	fastvideo/mlx_runtime/windowed_attention.py	56	Validate compatible rank-4 query, key, and value tensors. Parameters: q (mx.array): Query tensor shape	STRING
HIGH	fastvideo/mlx_runtime/windowed_attention.py	163	Apply symmetric sliding-window self-attention with optional global sink positions. Parameters: q (mx.a	STRING
HIGH	fastvideo/mlx_runtime/taehv_decode.py	74	Ensure the TAEW2.1 checkpoint is available locally. A caller-provided path is treated as trusted and is only c	STRING
HIGH	fastvideo/mlx_runtime/taehv_decode.py	132	Load the TAEHV class from the vendored implementation or a local source override. Parameters: source_path (	STRING
HIGH	fastvideo/mlx_runtime/quant_backends.py	82	Normalize a quantization backend name and validate that it is supported. Parameters: backend (str): Backend	STRING
HIGH	fastvideo/mlx_runtime/quant_backends.py	168	Quantize a two-dimensional weight matrix using the specified native MLX backend. Parameters: w (mx.arr	STRING
HIGH	fastvideo/mlx_runtime/quant_backends.py	256	Measure the effective storage cost of a quantized weight. Parameters: backend (str): Quantization back	STRING
HIGH	examples/inference/basic/mlx_wan_quant_benchmark.py	23	Parse and validate a comma-separated list of quantization modes. Parameters: raw (str): Comma-separate	STRING
HIGH	examples/inference/basic/mlx_wan_quant_benchmark.py	97	Decode latents to a video and collect export timing and PyTorch MPS memory metrics. Parameters: args:	STRING
HIGH	examples/inference/basic/mlx_wan_quant_benchmark.py	144	Run denoising for one quantization mode and collect performance and memory metrics. Parameters: mode (	STRING
HIGH	examples/inference/basic/mlx_wan_video_quality.py	13	Read all frames from a video file as an RGB NumPy array. Parameters: path (Path): Path to the video fi	STRING
HIGH	examples/inference/basic/mlx_wan_video_quality.py	44	Compute pixel-level comparison metrics between candidate and reference video frames. Parameters: candi	STRING
HIGH	examples/inference/basic/mlx_wan_decode_benchmark.py	45	Parse and validate a comma-separated list of decoding backends. Parameters: raw (str): Comma-separated	STRING
HIGH	apps/fastvideo_studio/job_runner.py	763	Return log lines for a job. Args: job_id: The job ID after: Return only lines a	STRING

Over-Commented Block201 hits · 198 pts

Severity	File	Line	Snippet	Context
LOW	collect_env.py	301	# example outputs of CPU infos	COMMENT
LOW	collect_env.py	321	# pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand	COMMENT
LOW	collect_env.py	341	# L1tf: Not affected	COMMENT
LOW	collect_env.py	361	# ProcessorType=3	COMMENT
LOW	pyproject.toml	81		COMMENT
LOW	pyproject.toml	121	# GB10 (sm_121), so on ARM build from the in-tree source instead. The FastH3	COMMENT
LOW	pyproject.toml	181	# are opt-in and intentionally kept out of the default [eval] set.	COMMENT
LOW	requirements-mkdocs.txt	1	# This file was autogenerated by uv via the following command:	COMMENT
LOW	requirements-mkdocs.txt	21	gitdb==4.0.12	COMMENT
LOW	requirements-mkdocs.txt	41	# mkdocs	COMMENT
LOW	requirements-mkdocs.txt	61	# mkdocs-api-autonav	COMMENT
LOW	requirements-mkdocs.txt	81	# via -r requirements-mkdocs.in	COMMENT
LOW	requirements-mkdocs.txt	101	# via mkdocs-material	COMMENT
LOW	fastvideo-kernel/CMakeLists.txt	441		COMMENT
LOW	fastvideo-kernel/build.sh	1	#!/bin/bash	COMMENT
LOW	…eo-kernel/python/fastvideo_kernel/block_sparse_attn.py	381	q, k, v, o, lse, q2k_idx, q2k_num, variable_block_sizes = ctx.saved_tensors	COMMENT
LOW	…o-kernel/python/fastvideo_kernel/turbodiffusion_ops.py	481	# But ctx.save_for_backward saves the original x (3D if input was 3D).	COMMENT
LOW	…o-kernel/python/fastvideo_kernel/turbodiffusion_ops.py	501	# dy = grad_output	COMMENT
LOW	…o-kernel/python/fastvideo_kernel/turbodiffusion_ops.py	521	# Using PyTorch autograd for reference logic:	COMMENT
LOW	…fastvideo_kernel/triton_kernels/fused_compress_topk.py	241	scores_f32 = scores.to(tl.float32)	COMMENT
LOW	…l/python/fastvideo_kernel/triton_kernels/sla_triton.py	1	# SPDX-License-Identifier: Apache-2.0	COMMENT
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	1441	print(	COMMENT
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	1601	# test_fused_attention_backward()	COMMENT
LOW	fastvideo-kernel/attn_qat_infer/blackwell/utils.h	21	#include <stdlib.h>	COMMENT
LOW	…ideo-kernel/attn_qat_infer/blackwell/epilogue_tma_ws.h	141	Tensor sO = cute::as_position_independent_swizzle_tensor(make_tensor(make_smem_ptr(shared_storage.smem_o.begin()	COMMENT
LOW	…ideo-kernel/attn_qat_infer/blackwell/epilogue_tma_ws.h	161	// #pragma unroll	COMMENT
LOW	…tvideo-kernel/attn_qat_infer/blackwell/static_switch.h	1	// Inspired by	COMMENT
LOW	fastvideo-kernel/attn_qat_infer/blackwell/kernel_ws.h	21		COMMENT
LOW	…tvideo-kernel/attn_qat_infer/quantization/cuda_utils.h	1	#pragma once	COMMENT
LOW	tests/local_tests/zimage/test_zimage_encoder_parity.py	221	model = SimpleNamespace(	COMMENT
LOW	tests/local_tests/dreamx_world/README.md	201	# decoded: 9 frames, (64, 64, 3), uint8	COMMENT
LOW	tests/local_tests/magi_human/test_magi_human_parity.py	241		COMMENT
LOW	.agents/scripts/sync-skills.sh	1	#!/usr/bin/env bash	COMMENT
LOW	…add-model-02-parity/templates/component_parity_test.py	181	f"diff_max={diff.max().item():.6f} diff_mean={diff.mean().item():.6f}")	COMMENT
LOW	fastvideo/envs.py	521	FASTVIDEO_COSMOS25_LOG_KNOBS = EnvBool(False,	COMMENT
LOW	fastvideo/registry.py	961	hf_model_paths=["kandinskylab/Kandinsky-6.0-VSR-5s-Diffusers"],	COMMENT
LOW	fastvideo/fastvideo_args.py	181	# can restrict layers to adapt, e.g. ["q_proj"]	COMMENT
LOW	fastvideo/fastvideo_args.py	201	use_fsdp_inference: bool = False	COMMENT
LOW	fastvideo/fastvideo_args.py	221	# ``None`` (auto) folds FASTVIDEO_H3_ENCODER_SPLIT; requires the Ray	COMMENT
LOW	fastvideo/fastvideo_args.py	241	# probe. A released component is re-read from disk on the next generation,	COMMENT
LOW	fastvideo/fastvideo_args.py	281	# Regional (per-transformer-block) fullgraph torch.compile of the DiT at	COMMENT
LOW	fastvideo/utils.py	661	# encoders and avoid duplicating a DiT that will never be loaded.	COMMENT
LOW	fastvideo/attention/backends/video_sparse_attn_h3.py	561	# tensor is prepared after weights load and drives only the compiled	COMMENT
LOW	fastvideo/attention/backends/video_sparse_attn_h3.py	661	# holder outlives the training step -- one builder serves the whole run,	COMMENT
LOW	fastvideo/attention/backends/sla.py	1	# SPDX-License-Identifier: Apache-2.0	COMMENT
LOW	fastvideo/attention/backends/attn_qat_infer.py	61	# DIFFERENT quantization scheme from the sm_12x CUTLASS extension above --	COMMENT
LOW	fastvideo/attention/utils/flash_attn_no_pad.py	1	# Licensed under the TENCENT HUNYUAN COMMUNITY LICENSE AGREEMENT (the "License");	COMMENT
LOW	fastvideo/attention/utils/flash_attn_no_pad.py	201	return output	COMMENT
LOW	fastvideo/attention/utils/flash_attn_default.py	181	setup_context=_flash_attn_default_setup_context,	COMMENT
LOW	fastvideo/layers/linear.py	641		COMMENT
LOW	fastvideo/layers/rotary_embedding.py	1	# SPDX-License-Identifier: Apache-2.0	COMMENT
LOW	…deo/distributed/device_communicators/pynccl_wrapper.py	1	# SPDX-License-Identifier: Apache-2.0	COMMENT
LOW	…deo/distributed/device_communicators/pynccl_wrapper.py	141	# is a pointer to a pointer	COMMENT
LOW	…deo/distributed/device_communicators/pynccl_wrapper.py	161		COMMENT
LOW	…deo/distributed/device_communicators/pynccl_wrapper.py	181	[buffer_type, ctypes.c_size_t, ncclDataType_t, ctypes.c_int, ncclComm_t, cudaStream_t]),	COMMENT
LOW	…o/tests/nightly/test_e2e_kandinsky5_dmd_t2v_overfit.py	81		COMMENT
LOW	…o/tests/nightly/test_e2e_kandinsky5_dmd_t2v_overfit.py	101	# the checkpoint instead, feeding an already-exported diffusers directory	COMMENT
LOW	…o/tests/nightly/test_e2e_kandinsky5_dmd_t2v_overfit.py	121	# produced a grey-static reference that was visually indistinguishable	COMMENT
LOW	…o/tests/nightly/test_e2e_kandinsky5_dmd_t2v_overfit.py	141	# - MS-SSIM alone CANNOT separate collapsed clips from good ones with	COMMENT
LOW	fastvideo/tests/golden_gate/test_kandinsky5.py	101	# module-level os.environ.setdefault here leaked into every other	COMMENT
141 more matches not shown…

AI Structural Patterns147 hits · 138 pts

Severity	File	Line	Context
LOW	…thon/fastvideo_kernel/triton_kernels/st_attn_triton.py	56	CODE
LOW	…thon/fastvideo_kernel/triton_kernels/attn_qat_train.py	1230	CODE
LOW	fastvideo-kernel/tests/test_sta.py	15	CODE
LOW	fastvideo-kernel/tests/test_attn_qat_train.py	165	CODE
LOW	fastvideo-kernel/tests/test_vsa_forward.py	71	CODE
LOW	fastvideo/attention/layer.py	344	CODE
LOW	fastvideo/attention/backends/sage_attn3.py	82	CODE
LOW	fastvideo/attention/backends/sdpa.py	147	CODE
LOW	fastvideo/attention/backends/sla.py	189	CODE
LOW	fastvideo/attention/backends/sla.py	328	CODE
LOW	fastvideo/attention/backends/sla.py	561	CODE
LOW	fastvideo/attention/backends/vmoba.py	69	CODE
LOW	fastvideo/attention/backends/sage_attn.py	64	CODE
LOW	fastvideo/attention/utils/flash_attn_no_pad.py	100	CODE
LOW	fastvideo/attention/utils/flash_attn_no_pad.py	152	CODE
LOW	fastvideo/attention/utils/flash_attn_no_pad.py	201	CODE
LOW	fastvideo/attention/utils/flash_attn_cute.py	458	CODE
LOW	fastvideo/layers/linear.py	156	CODE
LOW	fastvideo/layers/visual_embedding.py	27	CODE
LOW	fastvideo/layers/vocab_parallel_embedding.py	356	CODE
LOW	fastvideo/layers/rotary_embedding.py	349	CODE
LOW	fastvideo/layers/quantization/nvfp4_config.py	588	CODE
LOW	fastvideo/layers/quantization/nvfp4_qat_config.py	141	CODE
LOW	fastvideo/dataset/preprocessing_datasets.py	667	CODE
LOW	fastvideo/dataset/parquet_dataset_iterable_style.py	62	CODE
LOW	fastvideo/training/training_utils.py	181	CODE
LOW	fastvideo/training/training_utils.py	542	CODE
LOW	fastvideo/tests/train/callbacks/test_validation.py	42	CODE
LOW	fastvideo/tests/ops/test_rotary_pos_embed_cache.py	20	CODE
LOW	fastvideo/tests/modal/launch_l40s_job.py	512	CODE
LOW	fastvideo/tests/modal/ssim_test.py	871	CODE
LOW	fastvideo/tests/modal/ssim_test.py	946	CODE
LOW	fastvideo/pipelines/stages/text_encoding.py	136	CODE
LOW	fastvideo/pipelines/stages/gamecraft_image_encoding.py	168	CODE
LOW	fastvideo/pipelines/stages/gamecraft_image_encoding.py	173	CODE
LOW	fastvideo/pipelines/basic/dreamx_world/stages.py	156	CODE
LOW	fastvideo/pipelines/basic/gen3c/camera_utils.py	83	CODE
LOW	…stvideo/pipelines/basic/magi_human/stages/denoising.py	76	CODE
LOW	…ipelines/basic/magi_human/stages/latent_preparation.py	103	CODE
LOW	…ideo/pipelines/basic/magi_human/stages/sr_denoising.py	27	CODE
LOW	fastvideo/models/mask_utils.py	71	CODE
LOW	fastvideo/models/upsamplers/ltx2_upsampler.py	162	CODE
LOW	fastvideo/models/dits/longcat.py	1073	CODE
LOW	fastvideo/models/dits/dreamx_world_ar.py	779	CODE
LOW	fastvideo/models/dits/sd3.py	103	CODE
LOW	fastvideo/models/dits/kandinsky6.py	946	CODE
LOW	fastvideo/models/dits/flux_2.py	181	CODE
LOW	fastvideo/models/dits/flux_2.py	350	CODE
LOW	fastvideo/models/dits/dreamx_world.py	193	CODE
LOW	fastvideo/models/dits/dreamx_world.py	449	CODE
LOW	fastvideo/models/dits/ltx2.py	1774	CODE
LOW	fastvideo/models/dits/ltx2.py	2318	CODE
LOW	fastvideo/models/dits/ltx2.py	2804	CODE
LOW	fastvideo/models/dits/flux.py	463	CODE
LOW	fastvideo/models/dits/matrixgame3/action_module.py	70	CODE
LOW	fastvideo/models/dits/matrixgame3/model.py	199	CODE
LOW	fastvideo/models/dits/matrixgame3/model.py	285	CODE
LOW	fastvideo/models/dits/matrixgame3/model.py	555	CODE
LOW	fastvideo/models/dits/matrixgame2/action_module.py	210	CODE
LOW	fastvideo/models/dits/matrixgame2/action_module.py	625	CODE
87 more matches not shown…

Self-Referential Comments39 hits · 109 pts

Severity	File	Line	Snippet	Context
MEDIUM	tests/local_tests/gamecraft/test_gamecraft_parity.py	100	# Create a minimal args object	COMMENT
MEDIUM	…ocal_tests/gamecraft/test_gamecraft_pipeline_parity.py	196	# Create the sampler	COMMENT
MEDIUM	tests/local_tests/ltx2_5/test_ltx2_5_conversion.py	279	# Create a fresh fixture with all 2.5 gates enabled	COMMENT
MEDIUM	tests/local_tests/ltx2_5/test_ltx2_5_sampler.py	89	# Create a minimal stage with a mock transformer	COMMENT
MEDIUM	tests/local_tests/ltx2_5/test_ltx2_5_sampler.py	98	# Create a batch with video and audio latents matching expected shapes	COMMENT
MEDIUM	tests/local_tests/ltx2_5/test_ltx2_5_registry.py	47	# Create a local directory with LTX2Pipeline _class_name and distilled variant metadata	COMMENT
MEDIUM	tests/local_tests/ltx2_5/test_ltx2_5_registry.py	60	# Create a local directory with LTX2Pipeline _class_name and dev variant metadata	COMMENT
MEDIUM	docs/generate_examples.py	809	# Create the main examples index only if requested	COMMENT
MEDIUM	fastvideo/fastvideo_args.py	994	# Create a dictionary of attribute values, with defaults for missing attributes	COMMENT
MEDIUM	fastvideo/fastvideo_args.py	1416	# Create a dictionary of attribute values, with defaults for missing attributes	COMMENT
MEDIUM	fastvideo/distributed/parallel_state.py	287	# Create a dummy context for non-CUDA platforms	COMMENT
MEDIUM	…deo/distributed/device_communicators/pynccl_wrapper.py	4	# This file is a pure Python wrapper for the NCCL library.	COMMENT
MEDIUM	fastvideo/dataset/preprocessing_datasets.py	166	# Create a seeded random generator for deterministic sampling	COMMENT
MEDIUM	fastvideo/dataset/preprocessing_datasets.py	317	# Create a seeded random generator for deterministic CFG	COMMENT
MEDIUM	fastvideo/dataset/preprocessing_datasets.py	605	# Create a text-only batch with dummy path	COMMENT
MEDIUM	fastvideo/dataset/parquet_dataset_map_style.py	51	# Create a random permutation of all indices	COMMENT
MEDIUM	fastvideo/dataset/parquet_dataset_map_style.py	304	# Create a seeded random generator for deterministic CFG	COMMENT
MEDIUM	fastvideo/dataset/latent_datasets.py	23	# Create a seeded random generator for deterministic CFG	COMMENT
MEDIUM	…stvideo/training/self_forcing_distillation_pipeline.py	588	# Create a new batch with detached tensors	COMMENT
MEDIUM	…stvideo/training/self_forcing_distillation_pipeline.py	642	# Create a new batch with detached tensors	COMMENT
MEDIUM	fastvideo/tests/hooks/test_hooks_api.py	29	# Create a simple model	COMMENT
MEDIUM	fastvideo/pipelines/stages/latent_preparation.py	79	# Create a dummy zero-length text embedding to satisfy downstream checks.	COMMENT
MEDIUM	…lines/preprocess/preprocess_pipeline_ode_trajectory.py	275	# Initialize class variables for data sharing	COMMENT
MEDIUM	…video/pipelines/preprocess/preprocess_pipeline_text.py	168	# Initialize class variables for data sharing	COMMENT
MEDIUM	…video/pipelines/preprocess/preprocess_pipeline_base.py	46	# Initialize class variables for data sharing	COMMENT
MEDIUM	…ame2/matrixgame2_preprocess_pipeline_ode_trajectory.py	379	# Initialize class variables for data sharing	COMMENT
MEDIUM	fastvideo/models/mask_utils.py	120	## Creating a simple causal mask:	STRING
MEDIUM	fastvideo/models/mask_utils.py	141	## Creating a sliding window mask:	STRING
MEDIUM	fastvideo/models/mask_utils.py	162	## Creating a chunked attention mask	STRING
MEDIUM	fastvideo/models/dits/mmaudio.py	481	# This method is also called by inference pipeline stages. Persistent	COMMENT
MEDIUM	fastvideo/models/wan/vae.py	524	# Create the components	STRING
MEDIUM	fastvideo/models/vaes/hunyuanvae.py	56	# Define the projection layers	COMMENT
MEDIUM	fastvideo/worker/ray_env.py	51	# This file contains a list of env vars that should not be copied	COMMENT
MEDIUM	fastvideo/worker/ray_distributed_executor.py	173	# Create the workers.	COMMENT
MEDIUM	fastvideo/worker/ray_utils.py	216	# Create a new placement group	COMMENT
MEDIUM	examples/inference/basic/basic_longcat_vc.py	159	# Create a new generator with refinement LoRA and BSA enabled	COMMENT
MEDIUM	examples/inference/basic/basic_longcat_t2v.py	134	# Create a new generator with refinement LoRA and BSA enabled	COMMENT
MEDIUM	examples/inference/basic/basic_longcat_i2v.py	143	# Create a new generator with refinement LoRA and BSA enabled	COMMENT
MEDIUM	scripts/dataset_preparation/generate_validation_json.py	33	# Create the final validation structure	COMMENT

Redundant / Tautological Comments67 hits · 100 pts

Severity	File	Line	Snippet	Context
LOW	fastvideo-kernel/python/fastvideo_kernel/ops.py	30	# Check if the specific op is available	COMMENT
LOW	fastvideo-kernel/python/fastvideo_kernel/vmoba.py	254	if selected_other_indices.numel() > 0: # Check if any 'other' indices were selected	CODE
LOW	tests/local_tests/gen3c/test_gen3c.py	82	# Check if distributed is initialized	COMMENT
LOW	tests/local_tests/gen3c/test_gen3c.py	182	# Check if distributed is initialized	COMMENT
LOW	fastvideo/fastvideo_args.py	1437	# Check if the attribute is provided in args	COMMENT
LOW	fastvideo/attention/backends/bsa_attn.py	265	# Check if all heads share the same mask for this batch element	COMMENT
LOW	fastvideo/attention/backends/sla.py	331	# Check if spas_sage_attn is available for SageSLA	COMMENT
LOW	fastvideo/layers/rotary_embedding.py	126	# Check if cos/sin are full head_dim (rotate_half style) or half (traditional style)	COMMENT
LOW	fastvideo/dataset/preprocessing_datasets.py	187	# Check if video is too long	COMMENT
LOW	fastvideo/dataset/parquet_dataset_map_style.py	135	# Check if cached info exists	COMMENT
LOW	fastvideo/dataset/parquet_dataset_iterable_style.py	92	# Assign files to current rank's SP group	COMMENT
LOW	fastvideo/dataset/parquet_dataset_iterable_style.py	145	# Check if sharding plan already exists	COMMENT
LOW	…stvideo/training/self_forcing_distillation_pipeline.py	48	# Check if FSDP2 auto wrap is enabled - not supported for self-forcing distillation	COMMENT
LOW	fastvideo/training/distillation_pipeline.py	275	# Set flag to prevent custom weight loading for teacher/critic models	COMMENT
LOW	fastvideo/tests/training/distill/test_distill_dmd.py	85	# Check if the process failed	COMMENT
LOW	…video/tests/training/self-forcing/test_self_forcing.py	184	# Check if the process failed	COMMENT
LOW	fastvideo/tests/training/Vanilla/test_training_loss.py	86	# Check if the process failed	COMMENT
LOW	fastvideo/tests/training/Vanilla/mfu_calculation.py	158	# Check if the process failed	COMMENT
LOW	fastvideo/tests/transformers/test_cosmos.py	132	# Check if outputs have the same shape	COMMENT
LOW	fastvideo/tests/transformers/test_cosmos.py	136	# Check if outputs are similar (allowing for small numerical differences)	COMMENT
LOW	fastvideo/tests/transformers/test_cosmos2_5.py	295	# Set models to eval mode	COMMENT
LOW	fastvideo/tests/transformers/test_cosmos2_5.py	398	# Check if outputs have the same shape	COMMENT
LOW	fastvideo/tests/transformers/test_cosmos2_5.py	404	# Check if outputs are similar	COMMENT
LOW	fastvideo/tests/encoders/test_hyt5_encoder.py	123	# Check if outputs are similar (allowing for small numerical differences)	COMMENT
LOW	fastvideo/tests/ssim/test_longcat_similarity.py	185	# Check if absolute or already exists	COMMENT
LOW	…video/tests/entrypoints/test_openai_api_integration.py	98	# Check if the process died.	COMMENT
LOW⚡	fastvideo/pipelines/composed_pipeline_base.py	643	# Check if all required modules were loaded	COMMENT
LOW	fastvideo/pipelines/composed_pipeline_base.py	783	# Return the output	COMMENT
LOW	fastvideo/pipelines/stages/longcat_kv_cache_init.py	41	# Check if KV cache is enabled	COMMENT
LOW	fastvideo/pipelines/stages/base.py	141	# Check if verification is enabled (simple approach for prototype)	COMMENT
LOW	…lines/preprocess/preprocess_pipeline_ode_trajectory.py	90	if text and text.strip(): # Check if text is not empty	CODE
LOW	…video/pipelines/preprocess/preprocess_pipeline_text.py	64	if text and text.strip(): # Check if text is not empty	CODE
LOW	…video/pipelines/preprocess/preprocess_pipeline_base.py	271	if not torch.all(pixel_values == 0): # Check if all values are zero	CODE
LOW	…ame2/matrixgame2_preprocess_pipeline_ode_trajectory.py	202	if not torch.all(pixel_values == 0): # Check if all values are zero	CODE
LOW	…deo/pipelines/basic/hunyuan15/hunyuan15_sr_pipeline.py	117	# Return the output	COMMENT
LOW	…eo/pipelines/basic/hunyuan15/hunyuan15_2sr_pipeline.py	140	# Return the output	COMMENT
LOW	…stvideo/pipelines/basic/longcat/longcat_vc_pipeline.py	147	# Check if we have video_latent (from VC encoding stage)	COMMENT
LOW	…stvideo/pipelines/basic/longcat/longcat_vc_pipeline.py	150	# Set image_latent to video_latent for parent class compatibility	COMMENT
LOW	fastvideo/models/hf_transformer_utils.py	91	# Check if the model path exists	COMMENT
LOW	fastvideo/models/dits/matrixgame2/action_module.py	149	# Check if we need to evict tokens	COMMENT
LOW	fastvideo/models/dits/hyworld/retrieval_context.py	134	# Check if within FOV	COMMENT
LOW	fastvideo/models/loader/component_loader.py	1243	# Check if we should use custom initialization weights	COMMENT
LOW	fastvideo/models/loader/component_loader.py	1362	# Check if model has nn.ModuleList for layerwise offload compatibility	COMMENT
LOW	fastvideo/models/loader/fsdp_load.py	350	# Check if we should use FSDP	COMMENT
LOW	fastvideo/models/loader/fsdp_load.py	813	# Check if we should use size-based filtering	COMMENT
LOW	fastvideo/models/vaes/gamecraftvae.py	314	# Check if temporal tiling needed (>64 output frames)	COMMENT
LOW	fastvideo/models/vaes/gamecraftvae.py	317	# Check if spatial tiling needed (large H or W)	COMMENT
LOW⚡	fastvideo/models/schedulers/base.py	14	# Check if subclass has defined all required properties	COMMENT
LOW	fastvideo/entrypoints/cli/bench_serving.py	337	# Check if we need to use multipart (for image edits with input images)	COMMENT
LOW	fastvideo/entrypoints/cli/bench_serving.py	431	# Check if we need to upload images (Multipart) or just send JSON	COMMENT
LOW	fastvideo/worker/multiproc_executor.py	144	# Check if master_port is provided in fastvideo_args	COMMENT
LOW	fastvideo/worker/multiproc_executor.py	395	# Check if workers were initialized (they might not be if initialization failed)	COMMENT
LOW	examples/inference/basic/basic_matrixgame2.py	59	# Set pin_cpu_memory to false if CPU RAM is limited and there're no frequent CPU-GPU transfer	COMMENT
LOW	examples/inference/basic/basic_longcat_vc.py	55	# Check if video exists	COMMENT
LOW	examples/inference/basic/basic_longcat_vc.py	104	# Check if video exists	COMMENT
LOW	examples/inference/basic/basic_lingbotworld_base_cam.py	21	# Set pin_cpu_memory to false if CPU RAM is limited and there're no frequent CPU-GPU transfer	COMMENT
LOW	examples/inference/basic/basic_wan2_2_i2v.py	21	# Set pin_cpu_memory to false if CPU RAM is limited and there're no frequent CPU-GPU transfer	COMMENT
LOW	…nference/basic/basic_self_forcing_causal_wan2_2_i2v.py	25	# Set pin_cpu_memory to false if CPU RAM is limited and there're no frequent CPU-GPU transfer	COMMENT
LOW	examples/inference/basic/basic_matrixgame2_streaming.py	66	# Set pin_cpu_memory to false if CPU RAM is limited and there're no frequent CPU-GPU transfer	COMMENT
LOW	…nference/basic/basic_self_forcing_causal_wan2_2_t2v.py	24	# Set pin_cpu_memory to false if CPU RAM is limited and there're no frequent CPU-GPU transfer	COMMENT
7 more matches not shown…

Structural Annotation Overuse31 hits · 56 pts

Severity	File	Line	Snippet	Context
LOW	tests/local_tests/lingbot_video/README.md	76	### Step 1: Clone The Official Reference Implementation	COMMENT
LOW	tests/local_tests/lingbot_video/README.md	90	### Step 2: Prepare The Virtual Environment	COMMENT
LOW	tests/local_tests/lingbot_video/README.md	111	### Step 3: Configure The Hugging Face Cache	COMMENT
LOW	tests/local_tests/lingbot_video/README.md	137	### Step 4: Understand Checkpoint Conversion	COMMENT
LOW	docs/contributing/developer_env/runpod.md	11	## Step 1: Create a Pod	COMMENT
LOW	docs/contributing/developer_env/runpod.md	50	## Step 2: Connect via SSH	COMMENT
LOW	docs/contributing/developer_env/runpod.md	84	## Step 3: Set Up FastVideo	COMMENT
LOW	docs/inference/add_pipeline.md	16	## Step 1: Pipeline Modules	COMMENT
LOW	docs/inference/add_pipeline.md	181	## Step 2: Directory Structure	COMMENT
LOW	docs/inference/add_pipeline.md	192	## Step 3: Implement Pipeline Class	COMMENT
LOW	docs/inference/add_pipeline.md	333	## Step 4: Register Your Pipeline	COMMENT
LOW⚡	…stvideo/training/self_forcing_distillation_pipeline.py	210	# Step 1: Initialize KV cache to all zeros	COMMENT
LOW⚡	…stvideo/training/self_forcing_distillation_pipeline.py	217	# Step 2: Cache context feature	COMMENT
LOW	…stvideo/training/self_forcing_distillation_pipeline.py	239	# Step 3: Temporal denoising loop	COMMENT
LOW⚡	…ning/matrixgame2_self_forcing_distillation_pipeline.py	360	# Step 1: Initialize KV cache to all zeros	COMMENT
LOW⚡	…ning/matrixgame2_self_forcing_distillation_pipeline.py	364	# Step 2: Temporal denoising loop	COMMENT
LOW⚡	fastvideo/training/distillation_pipeline.py	531	# Step 1: Randomly sample a target timestep index from denoising_step_list	COMMENT
LOW⚡	fastvideo/training/distillation_pipeline.py	536	# Step 2: Simulate the multi-step inference process up to the target timestep	COMMENT
LOW⚡	fastvideo/training/distillation_pipeline.py	572	# Step 3: Use the simulated noisy input for the final training step	COMMENT
LOW⚡	fastvideo/training/distillation_pipeline.py	581	# Step 4: Final student prediction (this is what we train on)	COMMENT
LOW⚡	fastvideo/tests/train/callbacks/test_ema.py	135	# Step 0: re-init at 2.0, then update against 2.0 → still 2.0.	COMMENT
LOW⚡	fastvideo/tests/train/callbacks/test_ema.py	137	# Step 1: drift transformer to 12.0, expect	COMMENT
LOW	fastvideo/models/dits/cosmos2_5.py	448	# Step 1: Compute ALL modulation parameters once (matches official model)	COMMENT
LOW	fastvideo/models/dits/cosmos2_5.py	476	# Step 2: Self-attention block	COMMENT
LOW	fastvideo/models/dits/cosmos2_5.py	487	# Step 3: Cross-attention block	COMMENT
LOW	fastvideo/models/dits/cosmos2_5.py	500	# Step 4: MLP block	COMMENT
LOW	scripts/checkpoint_conversion/longcat_to_fastvideo.py	591	# Step 1: Convert transformer weights	COMMENT
LOW⚡	scripts/checkpoint_conversion/longcat_to_fastvideo.py	629	# Step 2: Copy other components	COMMENT
LOW⚡	scripts/checkpoint_conversion/longcat_to_fastvideo.py	639	# Step 3: Convert LoRA weights	COMMENT
LOW⚡	scripts/checkpoint_conversion/longcat_to_fastvideo.py	644	# Step 4: Update transformer config	COMMENT
LOW	scripts/checkpoint_conversion/longcat_to_fastvideo.py	657	# Step 5: Create model_index.json	COMMENT

Verbosity Indicators21 hits · 41 pts

Severity	File	Line	Snippet	Context
LOW	fastvideo/platforms/__init__.py	28	# we need to check if fastvideo is built with cpu too.	COMMENT
LOW⚡	…stvideo/training/self_forcing_distillation_pipeline.py	210	# Step 1: Initialize KV cache to all zeros	COMMENT
LOW⚡	…stvideo/training/self_forcing_distillation_pipeline.py	217	# Step 2: Cache context feature	COMMENT
LOW	…stvideo/training/self_forcing_distillation_pipeline.py	239	# Step 3: Temporal denoising loop	COMMENT
LOW⚡	…ning/matrixgame2_self_forcing_distillation_pipeline.py	360	# Step 1: Initialize KV cache to all zeros	COMMENT
LOW⚡	…ning/matrixgame2_self_forcing_distillation_pipeline.py	364	# Step 2: Temporal denoising loop	COMMENT
LOW⚡	fastvideo/training/distillation_pipeline.py	531	# Step 1: Randomly sample a target timestep index from denoising_step_list	COMMENT
LOW⚡	fastvideo/training/distillation_pipeline.py	536	# Step 2: Simulate the multi-step inference process up to the target timestep	COMMENT
LOW⚡	fastvideo/training/distillation_pipeline.py	572	# Step 3: Use the simulated noisy input for the final training step	COMMENT
LOW⚡	fastvideo/training/distillation_pipeline.py	581	# Step 4: Final student prediction (this is what we train on)	COMMENT
LOW⚡	fastvideo/tests/train/callbacks/test_ema.py	135	# Step 0: re-init at 2.0, then update against 2.0 → still 2.0.	COMMENT
LOW⚡	fastvideo/tests/train/callbacks/test_ema.py	137	# Step 1: drift transformer to 12.0, expect	COMMENT
LOW	fastvideo/models/dits/cosmos2_5.py	448	# Step 1: Compute ALL modulation parameters once (matches official model)	COMMENT
LOW	fastvideo/models/dits/cosmos2_5.py	476	# Step 2: Self-attention block	COMMENT
LOW	fastvideo/models/dits/cosmos2_5.py	487	# Step 3: Cross-attention block	COMMENT
LOW	fastvideo/models/dits/cosmos2_5.py	500	# Step 4: MLP block	COMMENT
LOW	scripts/checkpoint_conversion/longcat_to_fastvideo.py	591	# Step 1: Convert transformer weights	COMMENT
LOW⚡	scripts/checkpoint_conversion/longcat_to_fastvideo.py	629	# Step 2: Copy other components	COMMENT
LOW⚡	scripts/checkpoint_conversion/longcat_to_fastvideo.py	639	# Step 3: Convert LoRA weights	COMMENT
LOW⚡	scripts/checkpoint_conversion/longcat_to_fastvideo.py	644	# Step 4: Update transformer config	COMMENT
LOW	scripts/checkpoint_conversion/longcat_to_fastvideo.py	657	# Step 5: Create model_index.json	COMMENT

AI Slop Vocabulary14 hits · 36 pts

Severity	File	Line	Snippet	Context
MEDIUM	…stvideo-kernel/tests/test_vsa_triton_backward_scale.py	89	# Forward is exact at any scale; this pins the harness itself.	COMMENT
LOW	fastvideo/dataset/preprocessing_datasets.py	212	# For images, just add sample info	COMMENT
MEDIUM	fastvideo/tests/train/methods/test_wan_causal_dfsft.py	116	# 5a-ii: device-keyed grad-norm regression on top of the same harness.	COMMENT
MEDIUM	fastvideo/tests/train/methods/test_wan_causal_dfsft.py	118	# rtol above the harness default: the causal model compiles flex_attention	COMMENT
MEDIUM	…video/tests/train/methods/test_matrixgame2_finetune.py	147	# 5a-ii: device-keyed grad-norm regression on top of the same harness.	COMMENT
MEDIUM	fastvideo/tests/train/methods/test_wan_finetune.py	133	# 5a-ii: device-keyed grad-norm regression on top of the same harness.	COMMENT
MEDIUM	fastvideo/tests/train/methods/test_ltx2_finetune.py	189	# Device-keyed grad-norm regression on top of the same harness.	COMMENT
MEDIUM	fastvideo/tests/train/methods/test_cosmos_finetune.py	115	# 5a-ii: device-keyed grad-norm regression on top of the same harness.	COMMENT
MEDIUM	fastvideo/pipelines/stages/validators.py	472	"""Get a comprehensive summary of all validation failures."""	STRING
MEDIUM	…tvideo/pipelines/basic/stable_audio/stages/decoding.py	46	# so the LatentSimilarityUtils harness can compare on pre-VAE	COMMENT
MEDIUM	fastvideo/models/dits/cosmos.py	548	# dtype so the model forward is robust regardless of caller.	COMMENT
LOW	…eo/third_party/eval/synchformer/hf_src/modeling_ast.py	556	# since all values in the D-dimension (latent) will also be nans, we can just use the first el	STRING
MEDIUM	examples/inference/basic/mlx_wan_prompt_to_video.py	87	# benchmark harness). New code should prefer resolve_model_root(None), which	COMMENT
MEDIUM	apps/fastvideo_studio/job_runner.py	720	# that might escape (though they shouldn't with our comprehensive handling)	COMMENT

Modern AI Meta-Vocabulary10 hits · 27 pts

Severity	File	Line	Snippet	Context
MEDIUM	…dd-model-09-pipeline/templates/pipeline_parity_test.py	63	except Exception as exc: # pragma: no cover - scaffold guard	CODE
MEDIUM	fastvideo/tests/performance/worker_log_capture.py	44	# Also persist parent-process logs (run markers, load orchestration)	COMMENT
MEDIUM	fastvideo/tests/mlx/test_mlx_minimax_h3_fast_spatial.py	85	# -- generation-level orchestration contracts (heavyweight phases mocked) ----	COMMENT
MEDIUM	fastvideo/hooks/layerwise_offload.py	140	# These hook entry points only orchestrate host-side parameter	COMMENT
MEDIUM	fastvideo/configs/pipelines/kandinsky6.py	19	# Same Qwen2.5-VL "prompt engineer" system template and 129-token crop as	COMMENT
MEDIUM	fastvideo/eval/metrics/videoscore2/metric.py	52	# to avoid matching digits inside the chain-of-thought reasoning.	STRING
MEDIUM	fastvideo/eval/metrics/videoscore2/metric.py	118	# Anchor on the numbered list to skip the chain-of-thought.	COMMENT
MEDIUM	fastvideo/eval/metrics/videoscore2/metric.py	165	# chain-of-thought outside the regex anchor (yielding score=0).	COMMENT
MEDIUM	scripts/checkpoint_conversion/convert_ltx2_weights.py	245	# the Gemma backbone's much larger context window.	COMMENT
MEDIUM	.github/workflows/infra-build-image.yml	23	# metadata/key helper, ABI dependency metadata, and build orchestration.	COMMENT

Example Usage Blocks14 hits · 21 pts

Severity	File	Line	Snippet	Context
LOW	fastvideo-kernel/build.sh	5	# Usage:	COMMENT
LOW	fastvideo-kernel/tests/test_sta.py	75	# Example usage	COMMENT
LOW	.agents/scripts/sync-skills.sh	11	# Usage:	COMMENT
LOW	fastvideo/dataset/ltx2_precomputed_dataset.py	4	# Usage:	COMMENT
LOW	…raining/finetune/cosmos2_5/preprocess_cosmos2_5_t2w.sh	24	# Usage:	COMMENT
LOW	examples/train/run_slurm.sh	4	# Usage:	COMMENT
LOW	examples/train/run.sh	4	# Usage:	COMMENT
LOW	…nario/ode_init_self_forcing_wan_causal/step2_export.sh	4	# Usage:	COMMENT
LOW	…gs/distribution_matching/kandinsky5/distill_dmd_qat.sh	4	# Usage:	COMMENT
LOW	scripts/inference/run.sh	4	# Usage:	COMMENT
LOW	apps/dreamverse/scripts/install_native_ffmpeg.sh	20	# Usage:	COMMENT
LOW	…dreamverse/scripts/launch/launch_backend_dreamverse.sh	10	# Usage:	COMMENT
LOW	…/dreamverse/scripts/launch/launch_backend_fastvideo.sh	6	# Usage:	COMMENT
LOW	apps/dreamverse/scripts/launch/launch_frontend.sh	5	# Usage:	COMMENT

Dead Code8 hits · 16 pts

Severity	File	Line	Context
MEDIUM	collect_env.py	262	CODE
MEDIUM	collect_env.py	264	CODE
MEDIUM	collect_env.py	266	CODE
MEDIUM	collect_env.py	267	CODE
MEDIUM	collect_env.py	278	CODE
MEDIUM	fastvideo/layers/custom_op.py	60	CODE
MEDIUM	fastvideo/layers/custom_op.py	62	CODE
MEDIUM	fastvideo/layers/custom_op.py	65	CODE

Hallucination Indicators1 hit · 10 pts

Severity	File	Line	Snippet	Context
CRITICAL	…/encoders/test_minimax_h3_qwen3_vl_checkpoint_nvfp4.py	545	layer.mlp.up_proj.weight_scale.data.fill_(0xFF)	CODE

Fake / Example Data6 hits · 8 pts

Severity	File	Line	Snippet	Context
LOW	…_tests/dreamx_world/test_dreamx_world_ar_conversion.py	21	raw_tensor.write_bytes(b"placeholder")	CODE
LOW⚡	scripts/checkpoint_conversion/wan_to_diffusers.py	91	dummy_key = new_key.replace("norm_added_k", "norm_added_q")	CODE
LOW⚡	scripts/checkpoint_conversion/wan_to_diffusers.py	92	dummy_value = torch.zeros_like(v)	CODE
LOW⚡	scripts/checkpoint_conversion/wan_to_diffusers.py	93	new_state_dict[dummy_key] = dummy_value	CODE
LOW	apps/dreamverse/web/e2e/mock-backed-generation.spec.ts	38	await expect(continuation).toHaveAttribute('placeholder', /generating video/i);	CODE
LOW	…ps/dreamverse/web/e2e/preset-prompt-generation.spec.ts	54	await expect(continuation).toHaveAttribute('placeholder', /generating/i);	CODE

Synthetic Comment Markers1 hit · 8 pts

Severity	File	Line	Snippet	Context
HIGH	fastvideo/models/hf_transformer_utils.py	124	# Special handling for stop token <\|eom_id\|> generated by llama 3 tool use.	COMMENT

Slop Phrases1 hit · 2 pts

Severity	File	Line	Snippet	Context
LOW	fastvideo/models/vaes/hunyuanvae.py	168	# If you are encountering an error here, make sure to try running encoding/decoding with	COMMENT

Overly Generic Function Names1 hit · 1 pts

Severity	File	Line	Snippet	Context
LOW	…stvideo/tests/train/utils/test_distributed_strategy.py	31	def helper(self) -> int:	CODE

TODO Padding1 hit · 1 pts

Severity	File	Line	Snippet	Context
LOW	fastvideo/entrypoints/video_generator.py	326	log_stats=False, # TODO: implement	CODE

Maintenance and Governance Review

Risk Review · v3

What These Metrics Mean

Pattern-Signal History

Severity Breakdown

Directory Score Breakdown

Pattern Findings