Home / ace-step / ACE-Step-1.5
AI Slop Code RIsk Monitor

ace-step/ACE-Step-1.5

The most powerful local music generation model that outperforms almost all commercial alternatives, supporting Mac, AMD, Intel, and CUDA devices.

40.0 High pattern signal View on GitHub

Maintenance and Governance Review

This report summarizes scanner evidence for ace-step/ACE-Step-1.5, a Python project with 13,414 GitHub stars. SynthScan examined 175,269 lines of code across 1166 source files, recording 4239 pattern matches distributed across 21 syntactic categories. The legacy adjusted pattern signal is 40.0 (High pattern signal); it is not an authorship estimate.

Use the findings to ask whether maintainers can explain intent, verify dependencies, simplify unnecessary scaffolding, and focus review effort. Risk Payload v1 adds ownership, supply-chain, over-engineering, and review-effort vectors when available. These are prompts for human verification, not probabilities, security findings, or measured maintenance costs.

40.0
Adjusted Signal
40.0
Raw Pattern Score
100%
Time Factor
2026-10-11
Last Push
13.4K
Stars
Python
Language
175.3K
Lines of Code
1.2K
Files
4.2K
Pattern Hits
2026-10-12
Scan Date
0.33
HC Hit Rate

Risk Review · v3

Use each vector to guide manual verification of the underlying findings.

Cognitive Ownership Risk 51.5/100 · HIGH

Can maintainers explain intent, assumptions, and trade-offs? 2216 finding(s).

Hallucinated Dependency & Supply Chain Risk 21.1/100 · LOW

Are imports declared and package/API names verified? 726 finding(s).

Agent Over-Engineering Ratio 78.8/100 · CRITICAL

Are wrappers, guards, and repeated structures necessary? 2943 finding(s).

PR Review Velocity Impact 87.4/100 · CRITICAL

Which changes need focused human review? 4239 finding(s).

What These Metrics Mean

Adjusted Signal
Legacy pattern score normalized per 1,000 scanned lines and multiplied by the repository-level time factor. It is not an estimate of AI authorship.
Raw Pattern Score
Sum of severity- and context-weighted pattern findings before applying the time factor.
Time Factor
A coarse multiplier based on repository last-push metadata. It does not identify when individual code was written.
Pattern Hits
Count of configured pattern findings. Review each location and surrounding code before acting.
HC Hit Rate
High- and critical-severity findings per scanned file. Severity is a configured detector weight, not proof of a vulnerability or authorship.
Lines / Files
Scan coverage denominators used by the legacy normalized score.

Pattern-Signal History

This chart tracks the legacy adjusted pattern score across scan runs. Changes can reflect source changes, scan coverage, or detector/scoring versions; they do not measure AI adoption or prove that a team refactored code. Compare runs only when their scoring versions are compatible. The dashed line tracks the raw pattern-hit count.

Severity Breakdown

Groups findings by configured severity weight. Severity is not a confidence estimate, security rating, or proof that a package is missing. Verify imports against the project's manifests and inspect each finding in its local design context.

CRITICAL 0HIGH 389MEDIUM 538LOW 3312

Directory Score Breakdown

This horizontal bar chart shows where configured pattern scores accumulate across directories. Treat the bars as locations to inspect, not as evidence of authorship or code quality. Use the file-level findings to decide whether a directory needs a focused ownership, dependency, complexity, or review check.

Pattern Findings

The scanner identified 4239 distinct pattern matches across 21 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 Identifiers1597 hits · 1755 pts
SeverityFileLineSnippetContext
LOW⚡profile_inference.py265def print_time_costs_breakdown(CODE
LOWprofile_inference.py1395def _print_batch_boundary_summary(results: List[Dict]):CODE
LOW⚡cli.py346def _install_prompt_edit_hook(CODE
LOWcli.py93def _get_default_checkpoint_dir() -> str:CODE
LOWcli.py224def _edit_formatted_prompt_via_file(formatted_prompt: str, instruction_path: str) -> str:CODE
LOWcli.py246def _extract_caption_lyrics_from_formatted_prompt(formatted_prompt: str) -> Tuple[Optional[str], Optional[str]]:CODE
LOWcli.py267def _extract_instruction_from_formatted_prompt(formatted_prompt: str) -> Optional[str]:CODE
LOWcli.py276def _extract_cot_metadata_from_formatted_prompt(formatted_prompt: str) -> dict:CODE
LOWcli.py497def _default_instruction_for_task(task_type: str, tracks: Optional[List[str]] = None) -> str:CODE
LOWopenrouter/stress_test.py255 def _get_instrumental_payload(self) -> dict:CODE
LOWopenrouter/openrouter_api_server.py320def _extract_prompt_and_lyrics(messages: List[ChatMessage]) -> tuple[str, str, str, List[str]]:CODE
LOWopenrouter/client_test.py131def test_natural_language_mode(base_url: str, api_key: Optional[str] = None) -> bool:CODE
LOWacestep/gpu_config_effective_free_vram_test.py89 def test_includes_pytorch_cache_when_device_free_is_zero(self):CODE
LOWacestep/gpu_config_effective_free_vram_test.py109 def test_sums_device_free_and_allocator_cache(self):CODE
LOWacestep/gpu_config_effective_free_vram_test.py123 def test_returns_zero_when_fully_allocated_no_cache(self):CODE
LOWacestep/gpu_config_effective_free_vram_test.py136 def test_returns_zero_when_cuda_unavailable(self):CODE
LOWacestep/gpu_config_effective_free_vram_test.py146 def test_debug_cap_clamps_effective_free(self):CODE
LOW⚡acestep/launcher_compat_test.py34 def test_legacy_arch_missing_requires_fix(self) -> None:CODE
LOW⚡acestep/launcher_compat_test.py42 def test_legacy_arch_present_skips_fix(self) -> None:CODE
LOW⚡acestep/launcher_compat_test.py50 def test_modern_gpu_skips_fix_even_if_arch_not_listed(self) -> None:CODE
LOWacestep/launcher_compat_test.py66 def test_determine_returns_probe_failed_on_exception(self) -> None:CODE
LOWacestep/launcher_compat_test.py77 def test_probe_exit_code_maps_to_decision(self) -> None:CODE
LOWacestep/llm_backend_compat.py7def _has_working_triton_installation() -> bool:CODE
LOWacestep/llm_backend_compat.py17def get_vllm_preflight_warning(*, device: str, platform: str | None = None) -> str | None:CODE
LOWacestep/audio_utils_uuid_test.py32 def test_different_lora_states_produce_different_uuids(self):CODE
LOWacestep/audio_utils_uuid_test.py62 def test_different_lora_scale_produces_different_uuids(self):CODE
LOWacestep/audio_utils_uuid_test.py86 def test_same_params_produce_same_uuid(self):CODE
LOWacestep/audio_utils_uuid_test.py102 def test_uuid_format_is_valid(self):CODE
LOW⚡acestep/audio_utils_uuid_test.py126 def test_lora_weights_hash_empty_when_no_lora(self):CODE
LOW⚡acestep/audio_utils_uuid_test.py131 def test_lora_weights_hash_empty_when_use_lora_false(self):CODE
LOW⚡acestep/audio_utils_uuid_test.py136 def test_different_lora_files_produce_different_hashes(self):CODE
LOWacestep/audio_utils_uuid_test.py172 def test_same_lora_file_produces_same_hash(self):CODE
LOWacestep/audio_utils_uuid_test.py192 def test_lora_weights_hash_differentiates_uuids(self):CODE
LOW⚡acestep/audio_utils_uuid_test.py224 def test_user_codes_preserved_when_lm_returns_empty(self):CODE
LOW⚡acestep/audio_utils_uuid_test.py232 def test_lm_codes_overwrite_when_non_empty(self):CODE
LOW⚡acestep/audio_utils_uuid_test.py239 def test_user_codes_preserved_when_lm_list_empty(self):CODE
LOWacestep/inference_repaint_cache_test.py21 def test_missing_sidecar_returns_none(self):CODE
LOWacestep/inference_repaint_cache_test.py26 def test_loads_relative_latent_file_from_sidecar(self):CODE
LOWacestep/inference_repaint_cache_test.py43 def test_loads_source_seed_from_sidecar(self):CODE
LOWacestep/inference_repaint_cache_test.py63 def test_resample_matching_source_seed_only_changes_collisions(self):CODE
LOWacestep/inference_repaint_cache_test.py71 def test_loads_sidecar_from_gradio_outputs_by_audio_basename(self):CODE
LOWacestep/inference_repaint_cache_test.py92 def test_gradio_output_sidecar_search_uses_literal_latest_match(self):CODE
LOWacestep/openrouter_adapter.py321def _to_generate_music_request(CODE
LOWacestep/openrouter_adapter.py416def _build_openrouter_response(CODE
LOWacestep/openrouter_adapter.py473async def _openrouter_stream_generator(CODE
LOWacestep/openrouter_adapter.py587 def _get_model_name_from_path(config_path: str) -> str:CODE
LOW⚡acestep/llm_inference_cfg_fixes_test.py228 def test_codes_generation_only_samples_valid_tokens(self):CODE
LOW⚡acestep/llm_inference_cfg_fixes_test.py238 def test_non_codes_state_samples_from_full_vocab(self):CODE
LOW⚡acestep/llm_inference_cfg_fixes_test.py248 def test_codes_generation_cfg_applied_to_valid_indices_only(self):CODE
LOW⚡acestep/llm_inference_cfg_fixes_test.py396 def test_generation_stops_when_all_sequences_finish(self):CODE
LOW⚡acestep/llm_inference_cfg_fixes_test.py431 def test_finished_sequences_are_forced_to_eos(self):CODE
LOWacestep/llm_inference_cfg_fixes_test.py44def _make_constrained_processor(state: "FSMState", vocab_size: int = 10) -> MagicMock:CODE
LOWacestep/llm_inference_cfg_fixes_test.py324 def test_cfg_nan_replaced_with_neg_inf(self):CODE
LOW⚡acestep/gpu_config.py124def get_cuda_device_capability(device_index: int = 0) -> Optional[Tuple[int, int]]:CODE
LOWacestep/gpu_config.py273def _apply_lm_backend_compatibility_overrides(config: GPUConfig) -> GPUConfig:CODE
LOWacestep/gpu_config.py940def find_best_lm_model_on_disk(CODE
LOWacestep/gpu_config.py1188def get_effective_free_vram_gb(device_index: int = 0) -> float:CODE
LOWacestep/direct_conditioning_tasks_test.py139 def test_dit_receives_source_audio_and_params_conditioning(self):CODE
LOWacestep/direct_conditioning_tasks_test.py147 def test_no_audio_codes_pre_empt_source_audio(self):CODE
LOW⚡acestep/direct_conditioning_tasks_test.py161 def test_dit_receives_source_audio_and_params_conditioning(self):CODE
1537 more matches not shown…
Decorative Section Separators499 hits · 1558 pts
SeverityFileLineSnippetContext
MEDIUM⚡docker-compose.jetson.yml1# =============================================================================COMMENT
MEDIUM⚡docker-compose.jetson.yml3# =============================================================================COMMENT
MEDIUMdocker-compose.jetson.yml31# =============================================================================COMMENT
MEDIUM⚡profile_inference.py260# =============================================================================COMMENT
MEDIUM⚡profile_inference.py262# =============================================================================COMMENT
MEDIUM⚡profile_inference.py1719# =============================================================================COMMENT
MEDIUM⚡profile_inference.py1721# =============================================================================COMMENT
MEDIUMprofile_inference.py85# =============================================================================COMMENT
MEDIUMprofile_inference.py87# =============================================================================COMMENT
MEDIUMprofile_inference.py140# =============================================================================COMMENT
MEDIUMprofile_inference.py142# =============================================================================COMMENT
MEDIUMprofile_inference.py194# =============================================================================COMMENT
MEDIUMprofile_inference.py196# =============================================================================COMMENT
MEDIUMprofile_inference.py420# =============================================================================COMMENT
MEDIUMprofile_inference.py422# =============================================================================COMMENT
MEDIUMprofile_inference.py538# =============================================================================COMMENT
MEDIUMprofile_inference.py540# =============================================================================COMMENT
MEDIUMprofile_inference.py716# =============================================================================COMMENT
MEDIUMprofile_inference.py718# =============================================================================COMMENT
MEDIUMprofile_inference.py1521# =============================================================================COMMENT
MEDIUMprofile_inference.py1523# =============================================================================COMMENT
MEDIUMprofile_inference.py1572# =============================================================================COMMENT
MEDIUMprofile_inference.py1574# =============================================================================COMMENT
MEDIUMprofile_inference.py1623# =============================================================================COMMENT
MEDIUMprofile_inference.py1625# =============================================================================COMMENT
MEDIUMprofile_inference.py1683# =============================================================================COMMENT
MEDIUMprofile_inference.py1685# =============================================================================COMMENT
MEDIUMprofile_inference.py1794# =============================================================================COMMENT
MEDIUMprofile_inference.py1796# =============================================================================COMMENT
MEDIUMprofile_inference.py2095# =============================================================================STRING
MEDIUMprofile_inference.py2097# =============================================================================STRING
MEDIUMrequirements-rocm.txt2# =================================================COMMENT
MEDIUMtrain.py31# ---------------------------------------------------------------------------COMMENT
MEDIUMtrain.py33# ---------------------------------------------------------------------------COMMENT
MEDIUM⚡train.py144# ===========================================================================COMMENT
MEDIUM⚡train.py146# ===========================================================================COMMENT
MEDIUMtrain.py265# ===========================================================================COMMENT
MEDIUMtrain.py267# ===========================================================================COMMENT
MEDIUM⚡docker-compose.yml1# =============================================================================COMMENT
MEDIUM⚡docker-compose.yml3# =============================================================================COMMENT
MEDIUM⚡docker-compose.yml15# =============================================================================COMMENT
MEDIUM⚡requirements-sidestep.txt1# ╔══════════════════════════════════════════════════════════════╗COMMENT
MEDIUM⚡requirements-sidestep.txt4# ╚══════════════════════════════════════════════════════════════╝COMMENT
MEDIUM⚡requirements-sidestep.txt9# ── Required for Side-Step CLI ────────────────────────────────COMMENT
MEDIUM⚡requirements-sidestep.txt12# ── Required for Side-Step TUI ────────────────────────────────COMMENT
MEDIUM⚡requirements-sidestep.txt20# ── Optional: Prodigy adaptive optimizer (auto-tunes LR) ─────COMMENT
MEDIUMopenrouter/stress_test.py37# =============================================================================COMMENT
MEDIUMopenrouter/stress_test.py39# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py52# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py54# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py65# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py67# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py96# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py98# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py230# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py232# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py494# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py496# =============================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py532 # =================================================================COMMENT
MEDIUMopenrouter/openrouter_api_server.py534 # =================================================================COMMENT
439 more matches not shown…
Cross-File Repetition179 hits · 895 pts
SeverityFileLineSnippetContext
HIGH…i/gradio/events/results/audio_playback_updates_test.py0load target module directly by file path for isolated testing.STRING
HIGH…/ui/gradio/interfaces/audio_player_preferences_test.py0load target module directly by file path for isolated testing.STRING
HIGHacestep/ui/gradio/interfaces/user_preferences_test.py0load target module directly by file path for isolated testing.STRING
HIGH…estep/ui/gradio/events/results/generation_info_test.py0load target module directly by file path for isolated testing.STRING
HIGHacestep/models/turbo/modeling_acestep_v15_turbo.py0general 4d attention mask generator compatible with cpu/mac/sdpa and eager mode. supports use cases: 1. causal full: is_STRING
HIGH…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py0general 4d attention mask generator compatible with cpu/mac/sdpa and eager mode. supports use cases: 1. causal full: is_STRING
HIGHacestep/models/xl_sft/modeling_acestep_v15_xl_base.py0general 4d attention mask generator compatible with cpu/mac/sdpa and eager mode. supports use cases: 1. causal full: is_STRING
HIGHacestep/models/xl_base/modeling_acestep_v15_xl_base.py0general 4d attention mask generator compatible with cpu/mac/sdpa and eager mode. supports use cases: 1. causal full: is_STRING
HIGHacestep/models/sft/modeling_acestep_v15_base.py0general 4d attention mask generator compatible with cpu/mac/sdpa and eager mode. supports use cases: 1. causal full: is_STRING
HIGHacestep/models/base/modeling_acestep_v15_base.py0general 4d attention mask generator compatible with cpu/mac/sdpa and eager mode. supports use cases: 1. causal full: is_STRING
HIGHacestep/models/turbo/modeling_acestep_v15_turbo.py0pack two sequences by concatenating and sorting them based on mask values. args: hidden1: first hidden states tensor of STRING
HIGH…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py0pack two sequences by concatenating and sorting them based on mask values. args: hidden1: first hidden states tensor of STRING
HIGHacestep/models/xl_sft/modeling_acestep_v15_xl_base.py0pack two sequences by concatenating and sorting them based on mask values. args: hidden1: first hidden states tensor of STRING
HIGHacestep/models/xl_base/modeling_acestep_v15_xl_base.py0pack two sequences by concatenating and sorting them based on mask values. args: hidden1: first hidden states tensor of STRING
HIGHacestep/models/sft/modeling_acestep_v15_base.py0pack two sequences by concatenating and sorting them based on mask values. args: hidden1: first hidden states tensor of STRING
HIGHacestep/models/base/modeling_acestep_v15_base.py0pack two sequences by concatenating and sorting them based on mask values. args: hidden1: first hidden states tensor of STRING
HIGHacestep/models/turbo/modeling_acestep_v15_turbo.py0sample timestep t and r for flow matching training. args: batch_size: batch size device: device to create tensors on dtySTRING
HIGH…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py0sample timestep t and r for flow matching training. args: batch_size: batch size device: device to create tensors on dtySTRING
HIGHacestep/models/xl_sft/modeling_acestep_v15_xl_base.py0sample timestep t and r for flow matching training. args: batch_size: batch size device: device to create tensors on dtySTRING
HIGHacestep/models/xl_base/modeling_acestep_v15_xl_base.py0sample timestep t and r for flow matching training. args: batch_size: batch size device: device to create tensors on dtySTRING
HIGHacestep/models/sft/modeling_acestep_v15_base.py0sample timestep t and r for flow matching training. args: batch_size: batch size device: device to create tensors on dtySTRING
HIGHacestep/models/base/modeling_acestep_v15_base.py0sample timestep t and r for flow matching training. args: batch_size: batch size device: device to create tensors on dtySTRING
HIGHacestep/models/turbo/modeling_acestep_v15_turbo.py0timestep embedding module for diffusion models. converts timestep values into high-dimensional embeddings using sinusoidSTRING
HIGH…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py0timestep embedding module for diffusion models. converts timestep values into high-dimensional embeddings using sinusoidSTRING
HIGHacestep/models/xl_sft/modeling_acestep_v15_xl_base.py0timestep embedding module for diffusion models. converts timestep values into high-dimensional embeddings using sinusoidSTRING
HIGHacestep/models/xl_base/modeling_acestep_v15_xl_base.py0timestep embedding module for diffusion models. converts timestep values into high-dimensional embeddings using sinusoidSTRING
HIGHacestep/models/sft/modeling_acestep_v15_base.py0timestep embedding module for diffusion models. converts timestep values into high-dimensional embeddings using sinusoidSTRING
HIGHacestep/models/base/modeling_acestep_v15_base.py0timestep embedding module for diffusion models. converts timestep values into high-dimensional embeddings using sinusoidSTRING
HIGHacestep/models/turbo/modeling_acestep_v15_turbo.py0create sinusoidal timestep embeddings. args: t: a 1-d tensor of n indices, one per batch element. these may be fractionaSTRING
HIGH…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py0create sinusoidal timestep embeddings. args: t: a 1-d tensor of n indices, one per batch element. these may be fractionaSTRING
HIGHacestep/models/xl_sft/modeling_acestep_v15_xl_base.py0create sinusoidal timestep embeddings. args: t: a 1-d tensor of n indices, one per batch element. these may be fractionaSTRING
HIGHacestep/models/xl_base/modeling_acestep_v15_xl_base.py0create sinusoidal timestep embeddings. args: t: a 1-d tensor of n indices, one per batch element. these may be fractionaSTRING
HIGHacestep/models/sft/modeling_acestep_v15_base.py0create sinusoidal timestep embeddings. args: t: a 1-d tensor of n indices, one per batch element. these may be fractionaSTRING
HIGHacestep/models/base/modeling_acestep_v15_base.py0create sinusoidal timestep embeddings. args: t: a 1-d tensor of n indices, one per batch element. these may be fractionaSTRING
HIGHacestep/models/turbo/modeling_acestep_v15_turbo.py0multi-headed attention module for acestep model. implements the attention mechanism from 'attention is all you need' papSTRING
HIGH…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py0multi-headed attention module for acestep model. implements the attention mechanism from 'attention is all you need' papSTRING
HIGHacestep/models/xl_sft/modeling_acestep_v15_xl_base.py0multi-headed attention module for acestep model. implements the attention mechanism from 'attention is all you need' papSTRING
HIGHacestep/models/xl_base/modeling_acestep_v15_xl_base.py0multi-headed attention module for acestep model. implements the attention mechanism from 'attention is all you need' papSTRING
HIGHacestep/models/sft/modeling_acestep_v15_base.py0multi-headed attention module for acestep model. implements the attention mechanism from 'attention is all you need' papSTRING
HIGHacestep/models/base/modeling_acestep_v15_base.py0multi-headed attention module for acestep model. implements the attention mechanism from 'attention is all you need' papSTRING
HIGHacestep/models/turbo/modeling_acestep_v15_turbo.py0encoder layer for acestep model. consists of self-attention and mlp (feed-forward) sub-layers with residual connections.STRING
HIGH…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py0encoder layer for acestep model. consists of self-attention and mlp (feed-forward) sub-layers with residual connections.STRING
HIGHacestep/models/xl_sft/modeling_acestep_v15_xl_base.py0encoder layer for acestep model. consists of self-attention and mlp (feed-forward) sub-layers with residual connections.STRING
HIGHacestep/models/xl_base/modeling_acestep_v15_xl_base.py0encoder layer for acestep model. consists of self-attention and mlp (feed-forward) sub-layers with residual connections.STRING
HIGHacestep/models/sft/modeling_acestep_v15_base.py0encoder layer for acestep model. consists of self-attention and mlp (feed-forward) sub-layers with residual connections.STRING
HIGHacestep/models/base/modeling_acestep_v15_base.py0encoder layer for acestep model. consists of self-attention and mlp (feed-forward) sub-layers with residual connections.STRING
HIGHacestep/models/turbo/modeling_acestep_v15_turbo.py0dit (diffusion transformer) layer for acestep model. implements a transformer layer with three main components: 1. self-STRING
HIGH…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py0dit (diffusion transformer) layer for acestep model. implements a transformer layer with three main components: 1. self-STRING
HIGHacestep/models/xl_sft/modeling_acestep_v15_xl_base.py0dit (diffusion transformer) layer for acestep model. implements a transformer layer with three main components: 1. self-STRING
HIGHacestep/models/xl_base/modeling_acestep_v15_xl_base.py0dit (diffusion transformer) layer for acestep model. implements a transformer layer with three main components: 1. self-STRING
HIGHacestep/models/sft/modeling_acestep_v15_base.py0dit (diffusion transformer) layer for acestep model. implements a transformer layer with three main components: 1. self-STRING
HIGHacestep/models/base/modeling_acestep_v15_base.py0dit (diffusion transformer) layer for acestep model. implements a transformer layer with three main components: 1. self-STRING
HIGHacestep/models/turbo/modeling_acestep_v15_turbo.py0initialize weights for different module types. todo: support separate initialization for encoders and decoders.STRING
HIGH…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py0initialize weights for different module types. todo: support separate initialization for encoders and decoders.STRING
HIGHacestep/models/xl_sft/modeling_acestep_v15_xl_base.py0initialize weights for different module types. todo: support separate initialization for encoders and decoders.STRING
HIGHacestep/models/xl_base/modeling_acestep_v15_xl_base.py0initialize weights for different module types. todo: support separate initialization for encoders and decoders.STRING
HIGHacestep/models/sft/modeling_acestep_v15_base.py0initialize weights for different module types. todo: support separate initialization for encoders and decoders.STRING
HIGHacestep/models/base/modeling_acestep_v15_base.py0initialize weights for different module types. todo: support separate initialization for encoders and decoders.STRING
HIGHacestep/models/turbo/modeling_acestep_v15_turbo.py0encoder for processing lyric text embeddings. encodes lyric text hidden states using a transformer encoder architecture STRING
HIGH…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py0encoder for processing lyric text embeddings. encodes lyric text hidden states using a transformer encoder architecture STRING
119 more matches not shown…
Empty Error Handlers140 hits · 700 pts
SeverityFileLineSnippetContext
HIGHprofile_inference.py1217except Exception:CODE
HIGHprofile_inference.py1225except Exception:CODE
HIGHprofile_inference.py1715except Exception:CODE
HIGHprofile_inference.py862except ImportError:CODE
HIGHprofile_inference.py1737except ImportError:CODE
HIGHcli.py24except ImportError:CODE
HIGHcli.py1784except (ValueError, TypeError):CODE
HIGHcli.py1795except (ValueError, TypeError):CODE
HIGHtrain.py75except ImportError:CODE
HIGHopenrouter/openrouter_api_server.py882except ValueError:CODE
HIGHopenrouter/openrouter_api_server.py364except Exception:CODE
HIGHacestep/openrouter_adapter.py165except Exception:CODE
HIGHacestep/openrouter_adapter.py175except Exception:CODE
HIGHacestep/gpu_config.py798except ValueError:CODE
HIGHacestep/gpu_config.py1028except (ValueError, TypeError):CODE
HIGHacestep/gpu_config.py1241except (ValueError, TypeError):CODE
HIGHacestep/gpu_config.py1253except Exception:CODE
HIGHacestep/llm_inference.py154except Exception:CODE
HIGHacestep/llm_inference.py274except Exception:CODE
HIGHacestep/llm_inference.py656except Exception:CODE
HIGHacestep/llm_inference.py668except ImportError:CODE
HIGHacestep/llm_inference.py1445except (ValueError, TypeError):CODE
HIGHacestep/llm_inference.py132except Exception:CODE
HIGHacestep/llm_inference.py650except Exception:CODE
HIGHacestep/llm_inference.py2294except (ValueError, TypeError):CODE
HIGHacestep/llm_inference.py2301except (ValueError, TypeError):CODE
HIGHacestep/llm_inference.py2508except ImportError:CODE
HIGHacestep/llm_inference.py2515except Exception:CODE
HIGHacestep/api_server.py177except Exception:CODE
HIGHacestep/acestep_v15_pipeline.py27except ImportError:CODE
HIGHacestep/inference.py360except (ValueError, TypeError):CODE
HIGHacestep/inference.py378except (ValueError, TypeError):CODE
HIGHacestep/inference.py1204except (ValueError, TypeError):CODE
HIGHacestep/inference.py1213except (ValueError, TypeError):CODE
HIGHacestep/inference.py1380except (ValueError, TypeError):CODE
HIGHacestep/inference.py1389except (ValueError, TypeError):CODE
HIGHacestep/inference.py1553except (ValueError, TypeError):CODE
HIGHacestep/inference.py1562except (ValueError, TypeError):CODE
HIGHacestep/inference.py679except (ValueError, TypeError):CODE
HIGHacestep/inference.py697except (ValueError, TypeError):CODE
HIGHacestep/constrained_logits_processor.py603except Exception:CODE
HIGHacestep/constrained_logits_processor.py1140except Exception:CODE
HIGHacestep/ui/gradio/api/api_routes.py162except Exception:CODE
HIGHacestep/ui/gradio/api/api_routes.py523except (ValueError, TypeError):CODE
HIGHacestep/ui/gradio/events/training/preprocess.py178except Exception:CODE
HIGHacestep/ui/gradio/events/training/dataset_ops.py103except Exception:CODE
HIGHacestep/ui/gradio/events/results/scoring.py192except (TypeError, ValueError):CODE
HIGHacestep/ui/gradio/events/results/scoring.py201except (TypeError, ValueError):CODE
HIGHacestep/ui/gradio/events/generation/metadata_loading.py214except (ValueError, TypeError):CODE
HIGHacestep/ui/gradio/events/generation/metadata_loading.py222except (ValueError, TypeError):CODE
HIGHacestep/ui/gradio/i18n/i18n.py104except KeyError:CODE
HIGHacestep/ui/gradio/interfaces/generation_defaults.py60except ImportError:CODE
HIGH…tep/third_parts/nano-vllm/nanovllm/layers/attention.py24except ImportError:CODE
HIGH…tep/third_parts/nano-vllm/nanovllm/layers/attention.py30except ImportError:CODE
HIGHacestep/third_parts/nano-vllm/nanovllm/utils/compat.py20except ImportError:CODE
HIGHacestep/third_parts/nano-vllm/nanovllm/utils/loader.py22except AttributeError:CODE
HIGHacestep/third_parts/nano-vllm/nanovllm/utils/loader.py29except AttributeError:CODE
HIGHacestep/third_parts/nano-vllm/nanovllm/utils/loader.py37except AttributeError:CODE
HIGH…/third_parts/nano-vllm/nanovllm/engine/model_runner.py306except (ValueError, TypeError):CODE
HIGHacestep/core/lora/introspection.py58except Exception:CODE
80 more matches not shown…
Unused Imports687 hits · 611 pts
SeverityFileLineSnippetContext
LOWprofile_inference.py74CODE
LOWprofile_inference.py74CODE
LOWprofile_inference.py107CODE
LOWprofile_inference.py860CODE
LOWprofile_inference.py1735CODE
LOWcli.py1063CODE
LOWtrain.py25CODE
LOWopenrouter/stress_test.py32CODE
LOWopenrouter/openrouter_api_server.py14CODE
LOWopenrouter/openrouter_api_server.py20CODE
LOWopenrouter/openrouter_api_server.py25CODE
LOWopenrouter/openrouter_api_server.py38CODE
LOWopenrouter/openrouter_api_server.py39CODE
LOWopenrouter/client_test.py17CODE
LOWacestep/gpu_config_effective_free_vram_test.py14CODE
LOWacestep/launcher_compat_test.py3CODE
LOWacestep/cli_args.py3CODE
LOWacestep/openrouter_adapter.py14CODE
LOWacestep/llm_inference_cfg_fixes_test.py13CODE
LOWacestep/openrouter_models.py7CODE
LOWacestep/llm_backend_compat_test.py3CODE
LOWacestep/llm_inference.py28CODE
LOWacestep/llm_inference.py2928CODE
LOWacestep/llm_inference.py666CODE
LOWacestep/llm_inference.py2965CODE
LOWacestep/llm_inference.py2965CODE
LOWacestep/inference_seed_fallback_test.py9CODE
LOWacestep/gpu_config_lm_memory_ratio_test.py28CODE
LOWacestep/acestep_v15_pipeline_test.py3CODE
LOWacestep/audio_utils_test.py3CODE
LOWacestep/audio_utils_test.py7CODE
LOWacestep/audio_utils.py11CODE
LOWacestep/audio_utils.py18CODE
LOWacestep/api_server.py16CODE
LOWacestep/acestep_v15_pipeline.py52CODE
LOWacestep/acestep_v15_pipeline.py52CODE
LOWacestep/acestep_v15_pipeline.py52CODE
LOWacestep/acestep_v15_pipeline.py74CODE
LOWacestep/acestep_v15_pipeline.py74CODE
LOWacestep/acestep_v15_pipeline.py74CODE
LOWacestep/cli_args_test.py3CODE
LOWacestep/inference.py15CODE
LOWacestep/inference.py24CODE
LOWacestep/inference.py24CODE
LOWacestep/inference.py24CODE
LOWacestep/inference.py24CODE
LOWacestep/inference.py24CODE
LOWacestep/model_downloader_test.py5CODE
LOWacestep/launcher_compat.py3CODE
LOWacestep/constrained_logits_processor.py24CODE
LOWacestep/constrained_logits_processor.py24CODE
LOWacestep/constrained_logits_processor.py30CODE
LOWacestep/constrained_logits_processor.py30CODE
LOWacestep/launcher_legacy_torch_fix_test.py3CODE
LOWacestep/dataset_handler.py11CODE
LOWacestep/dataset_handler.py11CODE
LOWacestep/dataset_handler.py11CODE
LOWacestep/text2music_src_audio_test.py14CODE
LOWacestep/text2music_src_audio_test.py15CODE
LOWacestep/text2music_src_audio_test.py16CODE
627 more matches not shown…
Excessive Try-Catch Wrapping421 hits · 438 pts
SeverityFileLineSnippetContext
LOW⚡profile_inference.py255 except Exception as e:CODE
LOW⚡profile_inference.py1715 except Exception:CODE
LOWprofile_inference.py920 except Exception as e:CODE
LOWprofile_inference.py1005 except Exception as e:CODE
LOWprofile_inference.py1217 except Exception:CODE
LOWprofile_inference.py1225 except Exception:CODE
MEDIUMcli.py991 print(f"Error saving configuration: {e}. Please try again.")CODE
LOW⚡cli.py335 except Exception:CODE
LOW⚡cli.py342 except Exception:CODE
LOWcli.py37 except Exception:CODE
LOWcli.py229 except Exception as e:CODE
LOWcli.py241 except Exception as e:CODE
LOWcli.py316 except Exception:CODE
LOWcli.py477 except Exception:CODE
LOWcli.py990 except Exception as e:CODE
LOWcli.py1152 except Exception as e:CODE
LOWcli.py1337 except Exception as e:CODE
LOWcli.py1648 except Exception as e:CODE
LOW⚡train.py134 except Exception as exc:CODE
LOWtrain.py188 except Exception as exc:CODE
LOWtrain.py234 except Exception as exc:CODE
LOWgenerate_examples.py130 except Exception as e:CODE
LOWopenrouter/stress_test.py233 except Exception as e:CODE
LOWopenrouter/stress_test.py305 except Exception as e:CODE
LOWopenrouter/stress_test.py726 except Exception as e:STRING
LOW⚡openrouter/openrouter_api_server.py967 except Exception as e:CODE
LOWopenrouter/openrouter_api_server.py364 except Exception:CODE
LOWopenrouter/openrouter_api_server.py585 except Exception as e:CODE
LOWopenrouter/openrouter_api_server.py743 except Exception as e:CODE
LOWopenrouter/openrouter_api_server.py785 except Exception as e:CODE
LOWopenrouter/openrouter_api_server.py825 except Exception as e:CODE
LOWopenrouter/openrouter_api_server.py1019 except Exception as e:CODE
LOWopenrouter/openrouter_api_server.py1082 except Exception as e:CODE
LOW⚡openrouter/client_test.py494 except Exception as e:CODE
LOWopenrouter/client_test.py106 except Exception as e:CODE
LOWopenrouter/client_test.py126 except Exception as e:CODE
LOWopenrouter/client_test.py183 except Exception as e:CODE
LOWopenrouter/client_test.py243 except Exception as e:CODE
LOWopenrouter/client_test.py284 except Exception as e:CODE
LOWopenrouter/client_test.py319 except Exception as e:CODE
LOWopenrouter/client_test.py411 except Exception as e:CODE
LOWopenrouter/client_test.py454 except Exception as e:CODE
LOWopenrouter/client_test.py479 except Exception as e:CODE
LOWacestep/openrouter_adapter.py165 except Exception:CODE
LOWacestep/openrouter_adapter.py175 except Exception:CODE
LOWacestep/openrouter_adapter.py674 except Exception as e:CODE
LOWacestep/openrouter_adapter.py770 except Exception as exc:CODE
LOW⚡acestep/gpu_config.py57 except Exception:CODE
LOW⚡acestep/gpu_config.py67 except Exception:CODE
LOW⚡acestep/gpu_config.py77 except Exception:CODE
LOW⚡acestep/gpu_config.py87 except Exception:CODE
LOW⚡acestep/gpu_config.py120 except Exception:CODE
LOW⚡acestep/gpu_config.py131 except Exception:CODE
LOW⚡acestep/gpu_config.py1253 except Exception:CODE
LOW⚡acestep/gpu_config.py1261 except Exception:CODE
LOW⚡acestep/gpu_config.py1265 except Exception:CODE
LOWacestep/gpu_config.py107 except Exception:CODE
LOWacestep/gpu_config.py530 except Exception as e:CODE
LOWacestep/gpu_config.py606 except Exception as e:CODE
LOWacestep/gpu_config.py623 except Exception:CODE
361 more matches not shown…
Deep Nesting225 hits · 185 pts
SeverityFileLineSnippetContext
LOWprofile_inference.py116CODE
LOWprofile_inference.py543CODE
LOWprofile_inference.py1022CODE
LOWprofile_inference.py2100CODE
LOWcli.py657CODE
LOWcli.py1000CODE
LOWgenerate_examples.py20CODE
LOWopenrouter/stress_test.py523CODE
LOWopenrouter/stress_test.py159CODE
LOWopenrouter/openrouter_api_server.py320CODE
LOWopenrouter/openrouter_api_server.py498CODE
LOWopenrouter/openrouter_api_server.py631CODE
LOWopenrouter/openrouter_api_server.py704CODE
LOWopenrouter/openrouter_api_server.py790CODE
LOWopenrouter/client_test.py131CODE
LOWopenrouter/client_test.py190CODE
LOWopenrouter/client_test.py326CODE
LOWacestep/openrouter_adapter.py236CODE
LOWacestep/openrouter_adapter.py473CODE
LOWacestep/openrouter_adapter.py575CODE
LOWacestep/openrouter_adapter.py654CODE
LOWacestep/llm_inference_cfg_fixes_test.py78CODE
LOWacestep/llm_inference_cfg_fixes_test.py122CODE
LOWacestep/gpu_config.py476CODE
LOWacestep/gpu_config.py743CODE
LOWacestep/gpu_config.py976CODE
LOWacestep/gpu_config.py1077CODE
LOWacestep/gpu_config.py1188CODE
LOWacestep/gpu_config.py1470CODE
LOWacestep/llm_inference.py91CODE
LOWacestep/llm_inference.py125CODE
LOWacestep/llm_inference.py499CODE
LOWacestep/llm_inference.py874CODE
LOWacestep/llm_inference.py985CODE
LOWacestep/llm_inference.py1145CODE
LOWacestep/llm_inference.py1282CODE
LOWacestep/llm_inference.py2216CODE
LOWacestep/llm_inference.py2362CODE
LOWacestep/llm_inference.py2618CODE
LOWacestep/llm_inference.py2792CODE
LOWacestep/llm_inference.py3036CODE
LOWacestep/llm_inference.py3397CODE
LOWacestep/llm_inference.py4099CODE
LOWacestep/llm_inference.py4157CODE
LOWacestep/llm_inference.py2852CODE
LOWacestep/model_downloader.py442CODE
LOWacestep/model_downloader.py864CODE
LOWacestep/audio_utils.py451CODE
LOWacestep/audio_utils.py511CODE
LOWacestep/audio_utils.py197CODE
LOWacestep/acestep_v15_pipeline.py144CODE
LOWacestep/inference.py514CODE
LOWacestep/constrained_logits_processor.py548CODE
LOWacestep/constrained_logits_processor.py676CODE
LOWacestep/constrained_logits_processor.py788CODE
LOWacestep/constrained_logits_processor.py851CODE
LOWacestep/constrained_logits_processor.py913CODE
LOWacestep/constrained_logits_processor.py953CODE
LOWacestep/constrained_logits_processor.py1079CODE
LOWacestep/constrained_logits_processor.py1335CODE
165 more matches not shown…
Docstring Block Structure31 hits · 155 pts
SeverityFileLineSnippetContext
HIGHacestep/cli_args.py16Parse ``--quantization`` values from CLI input. Args: value: Raw CLI value. Returns: CanonicalSTRING
HIGHacestep/model_downloader.py711Resolve a VAE variant id (or absolute path) to its on-disk directory. Args: checkpoint_dir: Root checkpointSTRING
HIGH…io/events/wiring/generation_batch_navigation_wiring.py176Register previous/next batch navigation and background pre-generation wiring. Args: context (GenerationWiriSTRING
HIGH…estep/ui/gradio/events/wiring/generation_run_wiring.py30Proxy passthrough to `res_h.generate_with_batch_management`. Args: *args (Any): Positional passthroSTRING
HIGH…/third_parts/nano-vllm/nanovllm/engine/model_runner.py30Encode IPC arguments to UTF-8 JSON bytes for shared-memory transport. Replaces ``pickle`` to avoid arbitrary code eSTRING
HIGH…/third_parts/nano-vllm/nanovllm/engine/model_runner.py84Find an available port starting from start_port. Args: start_port: The starting port number to check STRING
HIGHacestep/core/generation/handler/lyric_timestamp.py28Generate LRC timestamps by aligning decoder cross-attention to lyric tokens. Args: pred_latent (torSTRING
HIGH…estep/core/generation/handler/generate_music_decode.py26Collect decode inputs and validate raw diffusion latents. Args: outputs: ``service_generate`` outpuSTRING
HIGH…e/generation/handler/init_service_loader_components.py27Load the VAE checkpoint and return its resolved path. Args: checkpoint_dir: Root checkpoint directoSTRING
HIGH…e/generation/handler/init_service_loader_components.py69Load the text tokenizer and embedding model, then return its path. Args: checkpoint_dir: Root checkSTRING
HIGHacestep/core/generation/handler/mlx_vae_init_test.py10Load a handler mixin module for isolated tests. Args: filename: Module filename in the current test directoSTRING
HIGHacestep/core/generation/handler/mlx_vae_init_test.py88Build fake ``mlx.core`` and compile-call tracking for tests. Args: raise_compile: Whether inner ``_compile`STRING
HIGHacestep/core/generation/handler/lyric_score.py27Calculate lyric alignment scores for pure-noise and regressed-latent inputs. Args: pred_latent (torSTRING
HIGHacestep/core/generation/handler/io_audio.py14Read an audio file, with torchaudio fallback for formats unsupported by soundfile. soundfile (libsndfile) supports STRING
HIGHacestep/core/generation/handler/generate_music.py252Generate audio from text/reference inputs and return response payload. Args: captions: Text prompt STRING
HIGHacestep/training/data_module.py544Load a dataset from JSON file. Args: json_path: Path to the JSON dataset file. Returns: Tuple STRING
HIGHacestep/training/path_safety.py54Validate and normalise a user-provided path. The returned path is guaranteed to live under *base* (or the globaSTRING
HIGHacestep/training/path_safety.py97Open a file after validating its path. Convenience wrapper around ``safe_path`` + ``open``. Args: userSTRING
HIGHacestep/api/job_analysis_runtime.py19Run analysis-only branches and return response payload when handled. Args: req: Generation request object. STRING
HIGHacestep/api/job_generation_runtime.py22Run music generation and aggregate outputs across sequential MPS cover slices. Args: req: Generation requesSTRING
HIGHacestep/api/llm_generation_inputs.py44Resolve LLM flags and prepare caption/lyrics/metadata inputs for generation. Args: app_state: FastAPI app sSTRING
HIGHacestep/api/http/auth.py26Validate request auth from body ``ai_token`` or Authorization header. Args: body: Parsed request payload diSTRING
HIGHacestep/api/http/model_init_service.py66Initialize DiT and optional LM models. Args: app_state: FastAPI application state carrying handler objects STRING
HIGHacestep/api/http/release_task_request_parser.py49Parse ``/release_task`` request body into request model and temp-file list. Args: request: FastAPI request STRING
HIGHacestep/api/http/release_task_audio_paths.py15Validate user-supplied audio path and block unsafe filesystem traversal. Args: path: User-supplied path valSTRING
HIGHacestep/api/http/release_task_audio_paths.py50Persist uploaded audio file to a temporary location. Args: upload: Uploaded file wrapper from Starlette/FasSTRING
HIGHacestep/training_v2/ui/presets.py250Save a preset to the project-local presets directory. Presets are saved to ``./presets/`` (relative to the current STRING
HIGHacestep/training_v2/ui/prompt_helpers.py80Display a numbered menu and return the chosen key. Args: title: Prompt text. options: List of ``(keSTRING
HIGHacestep/training_v2/ui/prompt_helpers.py157Ask for a single value with an optional default. Args: label: Prompt text. default: Default value (STRING
HIGHacestep/training_v2/ui/flows_train.py143Interactive wizard for training (fixed or vanilla). Args: mode: Training mode ('fixed' or 'vanilla'). STRING
HIGHacestep/training_v2/cli/validation.py61Resolve target modules based on attention type selection. Args: target_modules: List of module patterns (e.STRING
Cross-Language Confusion23 hits · 140 pts
SeverityFileLineSnippetContext
HIGHacestep/llm_inference.py800 logger.error("nano-vllm is not installed. Please install it using 'cd acestep/third_parts/nano-vllm && pip iCODE
HIGHacestep/llm_inference.py801 return "❌ nano-vllm is not installed. Please install it using 'cd acestep/third_parts/nano-vllm && pip instaCODE
HIGH…step/ui/gradio/events/wiring/results_display_wiring.py30 pathString = current_audio.path || current_audio.name || current_audio.url || "";CODE
HIGH…step/ui/gradio/events/wiring/results_display_wiring.py49 const fPath = f.name || f.path || "";CODE
HIGH…step/ui/gradio/events/wiring/results_display_wiring.py65 a.href = f.url || f.data;CODE
HIGHacestep/ui/gradio/interfaces/user_preferences.py112 skip_sentinel = f"new Array({num_outputs}).fill(null)"CODE
HIGH⚡acestep/ui/gradio/interfaces/user_preferences.py130 if (!(k in prefs)) return null;CODE
HIGH⚡acestep/ui/gradio/interfaces/user_preferences.py135 if (expected && typeof v !== expected) {{CODE
HIGH⚡acestep/ui/gradio/interfaces/user_preferences.py138 return null;CODE
HIGH⚡acestep/ui/gradio/interfaces/user_preferences.py146 else return null;CODE
HIGH⚡acestep/ui/gradio/interfaces/user_preferences.py151 if (result.every(v => v === null)) return SKIP;CODE
HIGH⚡acestep/ui/gradio/interfaces/user_preferences.py159 result.push(mp3, mp3, mp3);CODE
HIGHacestep/ui/gradio/interfaces/user_preferences.py173 - First ``len(PREF_KEYS)`` elements are preference values (or null).STRING
HIGHacestep/ui/gradio/interfaces/user_preferences.py174 - Next 3 elements are mp3 visibility booleans (or null):STRING
HIGH⚡…/ui/gradio/interfaces/audio_player_preferences_test.py62 self.assertIn("if (value === null || value === undefined || value === \"\")", script)CODE
HIGH⚡…/ui/gradio/interfaces/audio_player_preferences_test.py63 self.assertIn("if (preferredVolume === null)", script)CODE
HIGH⚡acestep/ui/gradio/interfaces/user_preferences_test.py152 self.assertIn("fill(null)", js)CODE
HIGH⚡acestep/ui/gradio/interfaces/user_preferences_test.py161 self.assertIn('result.push(', js)CODE
HIGH⚡acestep/models/common/flow_edit_helpers_test.py117 out = pack_for_cfg(self.enc_hs, self.enc_am, self.ctx, self.attn, self.null, do_cfg=False)CODE
HIGH⚡acestep/models/common/flow_edit_helpers_test.py124 out = pack_for_cfg(self.enc_hs, self.enc_am, self.ctx, self.attn, self.null, do_cfg=True)CODE
HIGHacestep/models/common/flow_edit_helpers.py50 """Reduce a ``(cond, null)``-packed prediction to a guided velocity.STRING
HIGHacestep/models/common/flow_edit_helpers.py184 torch.cat([enc_hs, null], dim=0),CODE
HIGHacestep/training_v2/timestep_sampling.py117 # Per-sample mask: 0 = drop condition (replace with null), 1 = keepCOMMENT
Structural Annotation Overuse99 hits · 137 pts
SeverityFileLineSnippetContext
LOWCONTRIBUTING.md65### Step 1: Commit-Scoped Review (First Pass)COMMENT
LOW⚡CONTRIBUTING.md84### Step 2: Validate FindingsCOMMENT
LOW⚡CONTRIBUTING.md94### Step 3: Apply Minimal FixesCOMMENT
LOW⚡CONTRIBUTING.md100### Step 4: PR-Scoped Review (Second Pass)COMMENT
LOWCONTRIBUTING.md114### Step 5: Write Reviewer ResponsesCOMMENT
LOW⚡openrouter/openrouter_api_server.py963 # Step 1: Run LM sample generationCOMMENT
LOWopenrouter/openrouter_api_server.py992 # Step 2: Run audio generation with heartbeatsCOMMENT
LOW⚡acestep/llm_inference_cfg_fixes_test.py438 # Step 0: output EOS; step 1+: output token 5COMMENT
LOW⚡acestep/constrained_logits_processor.py719 # Step 1: Encode full string "keyscale: {keyscale}" (with space, as tokenizer sees it)COMMENT
LOW⚡acestep/constrained_logits_processor.py723 # Step 2: Find where context ends in full_token_idsCOMMENT
LOW⚡acestep/constrained_logits_processor.py738 # Step 3: Extract keyscale value tokens (everything after context)COMMENT
LOW⚡acestep/constrained_logits_processor.py741 # Step 4: Verify we extracted some tokens (sanity check)COMMENT
LOW⚡acestep/constrained_logits_processor.py747 # Step 5: Verify first token is a note (A-G)COMMENT
LOW⚡acestep/constrained_logits_processor.py759 # Step 6: Build prefix mappings from keyscale value tokensCOMMENT
LOW⚡acestep/ui/gradio/i18n/en.json528 "training_dataset": "## Dataset Builder Tutorial\n\n### Step 1: Load or Scan\n- **Load**: Enter path to existing datCODE
LOW⚡acestep/models/turbo/modeling_acestep_v15_turbo.py156 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡acestep/models/turbo/modeling_acestep_v15_turbo.py162 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡acestep/models/turbo/modeling_acestep_v15_turbo.py165 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡acestep/models/turbo/modeling_acestep_v15_turbo.py168 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py500 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py516 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py531 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOW⚡…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py156 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py162 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py165 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py168 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py500 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py516 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py531 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOW⚡acestep/models/xl_sft/modeling_acestep_v15_xl_base.py159 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡acestep/models/xl_sft/modeling_acestep_v15_xl_base.py165 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡acestep/models/xl_sft/modeling_acestep_v15_xl_base.py168 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡acestep/models/xl_sft/modeling_acestep_v15_xl_base.py171 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py503 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py519 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py534 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOW⚡acestep/models/xl_base/modeling_acestep_v15_xl_base.py159 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡acestep/models/xl_base/modeling_acestep_v15_xl_base.py165 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡acestep/models/xl_base/modeling_acestep_v15_xl_base.py168 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡acestep/models/xl_base/modeling_acestep_v15_xl_base.py171 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py503 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py519 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py534 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOW⚡acestep/models/sft/modeling_acestep_v15_base.py159 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡acestep/models/sft/modeling_acestep_v15_base.py165 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡acestep/models/sft/modeling_acestep_v15_base.py168 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡acestep/models/sft/modeling_acestep_v15_base.py171 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOWacestep/models/sft/modeling_acestep_v15_base.py503 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOWacestep/models/sft/modeling_acestep_v15_base.py519 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOWacestep/models/sft/modeling_acestep_v15_base.py534 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOW⚡acestep/models/mlx/dit_model.py329 # Step 1: Self-attention with AdaLNCOMMENT
LOW⚡acestep/models/mlx/dit_model.py339 # Step 2: Cross-attentionCOMMENT
LOW⚡acestep/models/mlx/dit_model.py350 # Step 3: MLP with AdaLNCOMMENT
LOW⚡acestep/models/base/modeling_acestep_v15_base.py159 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡acestep/models/base/modeling_acestep_v15_base.py165 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡acestep/models/base/modeling_acestep_v15_base.py168 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡acestep/models/base/modeling_acestep_v15_base.py171 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOWacestep/models/base/modeling_acestep_v15_base.py503 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOWacestep/models/base/modeling_acestep_v15_base.py519 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOWacestep/models/base/modeling_acestep_v15_base.py534 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
39 more matches not shown…
Redundant / Tautological Comments66 hits · 80 pts
SeverityFileLineSnippetContext
LOWinstall_uv.sh27# Check if uv is already installedCOMMENT
LOWprofile_inference.py524 # Print resultsCOMMENT
LOWstart_api_server.sh121# Check if uv is installedCOMMENT
LOWstart_api_server.sh211# Check if virtual environment existsCOMMENT
LOWstart_gradio_ui_macos_manual.sh356# Check if uv is installedCOMMENT
LOWstart_gradio_ui_macos_manual.sh406# Check if virtual environment existsCOMMENT
LOWtrain.py245 # Display resultsCOMMENT
LOWstart_gradio_ui.sh212# Check if uv is installedCOMMENT
LOWstart_gradio_ui.sh328# Check if virtual environment existsCOMMENT
LOWstart_gradio_ui_macos.sh302# Check if uv is installedCOMMENT
LOWstart_gradio_ui_macos.sh352# Check if virtual environment existsCOMMENT
LOWstart_gradio_ui_manual.sh333# Check if uv is installedCOMMENT
LOWstart_gradio_ui_manual.sh404# Check if virtual environment existsCOMMENT
LOWstart_api_server_macos.sh198# Check if uv is installedCOMMENT
LOWstart_api_server_macos.sh243# Check if virtual environment existsCOMMENT
LOWcheck_update.sh40# Check if this is a git repositoryCOMMENT
LOWcheck_update.sh137# Check if local is behind remoteCOMMENT
LOWcheck_update.sh236 # Check if any untracked files conflict with incoming changesCOMMENT
LOWopenrouter/openrouter_api_server.py644 # Check if model is initializedCOMMENT
LOWacestep/gpu_config.py1224 # Check if a per-process memory fraction has been setCOMMENT
LOWacestep/gpu_config.py1407 # Check if model size is in available modelsCOMMENT
LOWacestep/llm_inference.py1688 # Check if user provided a meaningful negative prompt (not the default)COMMENT
LOWacestep/llm_inference.py2889 # Check if this is a new field (no leading spaces and contains ':')COMMENT
LOWacestep/llm_inference.py2979 # Check if keys need "model." prefix by inspecting first keyCOMMENT
LOWacestep/llm_inference.py3287 # Check if all items are doneCOMMENT
LOWacestep/model_downloader.py646 # Check if this is a known LM modelCOMMENT
LOWacestep/model_downloader.py648 # Check if it might be a variant nameCOMMENT
LOWacestep/model_downloader.py689 # Check if this is the default turbo model (part of main)COMMENT
LOWacestep/model_downloader.py693 # Check if it's a known sub-modelCOMMENT
LOWacestep/inference.py747 # Check if LM generation failedCOMMENT
LOWacestep/inference.py918 # Check if generation failedCOMMENT
LOWacestep/inference.py1159 # Check if LLM is initializedCOMMENT
LOWacestep/inference.py1183 # Check if LLM returned empty metadata (error case)COMMENT
LOWacestep/inference.py1336 # Check if LLM is initializedCOMMENT
LOWacestep/inference.py1358 # Check if LLM returned empty metadata (error case)COMMENT
LOWacestep/inference.py1510 # Check if LLM is initializedCOMMENT
LOWacestep/inference.py1532 # Check if LLM returned empty metadata (error case)COMMENT
LOW⚡acestep/constrained_logits_processor.py751 # Check if first token starts with a note (A-G, case insensitive, with optional leading space)COMMENT
LOW⚡acestep/constrained_logits_processor.py1804 # Check if field is user-provided and we haven't started injecting yetCOMMENT
LOWacestep/constrained_logits_processor.py1260 # Check if new_prefix is a valid prefix in the active trieCOMMENT
LOWacestep/constrained_logits_processor.py1389 # Check if decoded token matches the prefix start (allowing for space prefixes)COMMENT
LOWacestep/constrained_logits_processor.py1424 # Check if this digit could lead to a valid final valueCOMMENT
LOWacestep/constrained_logits_processor.py1433 # Check if we can still reach min_valCOMMENT
LOWacestep/constrained_logits_processor.py1597 # Check if input contains </think> token sequenceCOMMENT
LOWacestep/constrained_logits_processor.py1750 # Check if we have tokens in queue for user-provided fieldCOMMENT
LOWacestep/constrained_logits_processor.py1763 # Check if we should stop at reasoning (after </think> tag)COMMENT
LOWacestep/constrained_logits_processor.py1766 # Check if the next token would complete the fixed stringCOMMENT
LOWacestep/constrained_logits_processor.py1835 # Check if field is user-provided and we haven't started injecting yetCOMMENT
LOWacestep/constrained_logits_processor.py1848 # Check if we should transition after a newline (non-indented line = new field)COMMENT
LOWacestep/constrained_logits_processor.py1900 # Check if field is user-provided and we haven't started injecting yetCOMMENT
LOWacestep/constrained_logits_processor.py1940 # Check if field is user-provided and we haven't started injecting yetCOMMENT
LOWacestep/constrained_logits_processor.py1983 # Check if field is user-provided and we haven't started injecting yetCOMMENT
LOWacestep/constrained_logits_processor.py1996 # Check if current token sequence is complete (allows newline)COMMENT
LOWacestep/constrained_logits_processor.py2019 # Check if field is user-provided and we haven't started injecting yetCOMMENT
LOWacestep/constrained_logits_processor.py2066 # Check if current token sequence is complete (allows newline)COMMENT
LOWacestep/constrained_logits_processor.py2083 # Check if field is user-provided and we haven't started injecting yetCOMMENT
LOWacestep/constrained_logits_processor.py2096 # Check if current token sequence is complete (allows newline)COMMENT
LOWacestep/constrained_logits_processor.py2204 # Check if we've completed the fixed stringCOMMENT
LOWacestep/constrained_logits_processor.py2273 # Check if we've completed a field name (detected colon)COMMENT
LOWacestep/ui/gradio/interfaces/__init__.py62 # Check if running in service mode (hide training tab)COMMENT
6 more matches not shown…
Magic Placeholder Names16 hits · 78 pts
SeverityFileLineSnippetContext
HIGHopenrouter/client_test.py9 python -m openrouter.test_client --api-key your-api-keySTRING
HIGH.claude/skills/acestep-docs/api/API.md43 "ai_token": "your-api-key",CODE
HIGH.claude/skills/acestep-docs/api/API.md53 -H 'Authorization: Bearer your-api-key' \CODE
HIGH.claude/skills/acestep-docs/api/Openrouter_API.md28Authorization: Bearer <your-api-key>CODE
HIGHdocs/ja/API.md45 "ai_token": "your-api-key",CODE
HIGHdocs/ja/API.md55 -H 'Authorization: Bearer your-api-key' \CODE
HIGHdocs/ja/Openrouter_API_DOC.md29Authorization: Bearer <your-api-key>CODE
HIGHdocs/zh/API.md45 "ai_token": "your-api-key",CODE
HIGHdocs/zh/API.md55 -H 'Authorization: Bearer your-api-key' \CODE
HIGHdocs/zh/Openrouter_API_DOC.md29Authorization: Bearer <your-api-key>CODE
HIGHdocs/ko/API.md45 "ai_token": "your-api-key",CODE
HIGHdocs/ko/API.md55 -H 'Authorization: Bearer your-api-key' \CODE
HIGHdocs/ko/Openrouter_API_DOC.md29Authorization: Bearer <your-api-key>CODE
HIGHdocs/en/API.md46 "ai_token": "your-api-key",CODE
HIGHdocs/en/API.md56 -H 'Authorization: Bearer your-api-key' \CODE
HIGHdocs/en/Openrouter_API_DOC.md29Authorization: Bearer <your-api-key>CODE
Verbosity Indicators56 hits · 69 pts
SeverityFileLineSnippetContext
LOW⚡openrouter/openrouter_api_server.py963 # Step 1: Run LM sample generationCOMMENT
LOWopenrouter/openrouter_api_server.py992 # Step 2: Run audio generation with heartbeatsCOMMENT
LOW⚡acestep/llm_inference_cfg_fixes_test.py438 # Step 0: output EOS; step 1+: output token 5COMMENT
LOW⚡acestep/constrained_logits_processor.py719 # Step 1: Encode full string "keyscale: {keyscale}" (with space, as tokenizer sees it)COMMENT
LOW⚡acestep/constrained_logits_processor.py723 # Step 2: Find where context ends in full_token_idsCOMMENT
LOW⚡acestep/constrained_logits_processor.py738 # Step 3: Extract keyscale value tokens (everything after context)COMMENT
LOW⚡acestep/constrained_logits_processor.py741 # Step 4: Verify we extracted some tokens (sanity check)COMMENT
LOW⚡acestep/constrained_logits_processor.py747 # Step 5: Verify first token is a note (A-G)COMMENT
LOW⚡acestep/constrained_logits_processor.py759 # Step 6: Build prefix mappings from keyscale value tokensCOMMENT
LOW⚡acestep/ui/gradio/i18n/en.json528 "training_dataset": "## Dataset Builder Tutorial\n\n### Step 1: Load or Scan\n- **Load**: Enter path to existing datCODE
LOW⚡acestep/models/turbo/modeling_acestep_v15_turbo.py156 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡acestep/models/turbo/modeling_acestep_v15_turbo.py162 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡acestep/models/turbo/modeling_acestep_v15_turbo.py165 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡acestep/models/turbo/modeling_acestep_v15_turbo.py168 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py500 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py516 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py531 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOW⚡…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py156 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py162 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py165 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py168 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py500 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py516 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py531 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOW⚡acestep/models/xl_sft/modeling_acestep_v15_xl_base.py159 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡acestep/models/xl_sft/modeling_acestep_v15_xl_base.py165 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡acestep/models/xl_sft/modeling_acestep_v15_xl_base.py168 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡acestep/models/xl_sft/modeling_acestep_v15_xl_base.py171 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py503 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py519 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py534 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOW⚡acestep/models/xl_base/modeling_acestep_v15_xl_base.py159 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡acestep/models/xl_base/modeling_acestep_v15_xl_base.py165 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡acestep/models/xl_base/modeling_acestep_v15_xl_base.py168 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡acestep/models/xl_base/modeling_acestep_v15_xl_base.py171 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py503 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py519 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py534 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOW⚡acestep/models/sft/modeling_acestep_v15_base.py159 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡acestep/models/sft/modeling_acestep_v15_base.py165 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡acestep/models/sft/modeling_acestep_v15_base.py168 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡acestep/models/sft/modeling_acestep_v15_base.py171 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOWacestep/models/sft/modeling_acestep_v15_base.py503 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOWacestep/models/sft/modeling_acestep_v15_base.py519 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOWacestep/models/sft/modeling_acestep_v15_base.py534 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOW⚡acestep/models/mlx/dit_model.py329 # Step 1: Self-attention with AdaLNCOMMENT
LOW⚡acestep/models/mlx/dit_model.py339 # Step 2: Cross-attentionCOMMENT
LOW⚡acestep/models/mlx/dit_model.py350 # Step 3: MLP with AdaLNCOMMENT
LOW⚡acestep/models/base/modeling_acestep_v15_base.py159 # Step 1: Concatenate hidden states and masks along sequence dimensionCOMMENT
LOW⚡acestep/models/base/modeling_acestep_v15_base.py165 # Step 2: Sort indices so that mask values of 1 come before 0COMMENT
LOW⚡acestep/models/base/modeling_acestep_v15_base.py168 # Step 3: Reorder hidden states using sorted indicesCOMMENT
LOW⚡acestep/models/base/modeling_acestep_v15_base.py171 # Step 4: Create new mask based on valid sequence lengthsCOMMENT
LOWacestep/models/base/modeling_acestep_v15_base.py503 # Step 1: Self-attention with adaptive layer norm (AdaLN)COMMENT
LOWacestep/models/base/modeling_acestep_v15_base.py519 # Step 2: Cross-attention (if enabled) for conditioning on encoder outputsCOMMENT
LOWacestep/models/base/modeling_acestep_v15_base.py534 # Step 3: Feed-forward (MLP) with adaptive layer normCOMMENT
LOWacestep/training_v2/ui/flows_train.py179 # Step 0: config depthCOMMENT
Modern Structural Boilerplate58 hits · 62 pts
SeverityFileLineSnippetContext
LOWacestep/ui/gradio/__init__.py8__all__ = ["create_gradio_interface"]CODE
LOWacestep/ui/gradio/events/results_handlers.py75__all__ = [CODE
LOWacestep/ui/gradio/events/wiring/__init__.py32__all__ = [CODE
LOWacestep/ui/gradio/events/results/batch_management.py22__all__ = [CODE
LOWacestep/ui/gradio/events/generation/llm_actions.py15__all__ = [CODE
LOWacestep/core/scoring/__init__.py13__all__ = [CODE
LOWacestep/core/lora/__init__.py8__all__ = ["collect_adapter_names", "build_lora_registry", "apply_scale_to_adapter", "LoraService"]CODE
LOWacestep/core/generation/handler/__init__.py41__all__ = [CODE
LOWacestep/training/__init__.py55__all__ = [CODE
LOWacestep/training/dataset_builder.py9__all__ = [CODE
LOWacestep/training/path_safety.py38def set_safe_root(root: str) -> None:CODE
LOWacestep/training/dataset_builder_modules/__init__.py4__all__ = [CODE
LOWacestep/models/turbo/configuration_acestep_v15.py5__all__ = ["AceStepConfig"]CODE
LOWacestep/models/xl_turbo/configuration_acestep_v15.py5__all__ = ["AceStepConfig"]CODE
LOWacestep/models/xl_sft/configuration_acestep_v15.py5__all__ = ["AceStepConfig"]CODE
LOWacestep/models/xl_sft/apg_guidance.py15__all__ = [CODE
LOWacestep/models/common/configuration_acestep_v15.py265__all__ = ["AceStepConfig"]CODE
LOWacestep/models/common/dcw_correction.py44__all__ = [CODE
LOWacestep/models/xl_base/configuration_acestep_v15.py5__all__ = ["AceStepConfig"]CODE
LOWacestep/models/xl_base/apg_guidance.py15__all__ = [CODE
LOWacestep/models/sft/configuration_acestep_v15.py5__all__ = ["AceStepConfig"]CODE
LOWacestep/models/sft/apg_guidance.py15__all__ = [CODE
LOWacestep/models/mlx/dit_convert.py8logger = logging.getLogger(__name__)CODE
LOWacestep/models/mlx/__init__.py7logger = logging.getLogger(__name__)CODE
LOWacestep/models/mlx/vae_convert.py15logger = logging.getLogger(__name__)CODE
LOW⚡acestep/models/mlx/vae_model.py17logger = logging.getLogger(__name__)CODE
LOWacestep/models/mlx/dcw_correction_mlx.py28logger = logging.getLogger(__name__)CODE
LOWacestep/models/base/configuration_acestep_v15.py5__all__ = ["AceStepConfig"]CODE
LOWacestep/models/base/apg_guidance.py15__all__ = [CODE
LOWacestep/api/job_llm_preparation.py6__all__ = [CODE
LOW…ep/api/train_api_dataset_scan_load_routes_http_test.py81 def set_all_instrumental(self, value: bool) -> None:CODE
LOW…ep/api/train_api_dataset_scan_load_routes_http_test.py86 def set_custom_tag(self, tag: str, position: str) -> None:CODE
LOWacestep/api/http/auth.py12def set_api_key(key: Optional[str]) -> None:CODE
LOWacestep/api/jobs/store.py148 def update_progress(self, job_id: str, progress: float, stage: Optional[str] = None) -> None:CODE
LOWacestep/api/jobs/store.py194 def update_status_text(self, job_id: str, text: str) -> None:CODE
LOWacestep/api/jobs/store.py200 def update_progress_text(self, job_id: str, text: str) -> None:CODE
LOWacestep/training_v2/preprocess_discovery.py18logger = logging.getLogger(__name__)CODE
LOW⚡acestep/training_v2/trainer_helpers.py29logger = logging.getLogger(__name__)CODE
LOW⚡acestep/training_v2/preprocess.py39logger = logging.getLogger(__name__)CODE
LOW⚡acestep/training_v2/trainer_basic_loop.py26logger = logging.getLogger(__name__)CODE
LOWacestep/training_v2/trainer_fixed.py55logger = logging.getLogger(__name__)CODE
LOWacestep/training_v2/model_discovery.py18logger = logging.getLogger(__name__)CODE
LOWacestep/training_v2/estimate.py23logger = logging.getLogger(__name__)CODE
LOW⚡acestep/training_v2/settings.py23logger = logging.getLogger(__name__)CODE
LOW⚡acestep/training_v2/gpu_utils.py18logger = logging.getLogger(__name__)CODE
LOWacestep/training_v2/model_loader.py22logger = logging.getLogger(__name__)CODE
LOWacestep/training_v2/tensorboard_utils.py20logger = logging.getLogger(__name__)CODE
LOW⚡acestep/training_v2/fixed_lora_module.py67logger = logging.getLogger(__name__)CODE
LOWacestep/training_v2/trainer_vanilla.py58logger = logging.getLogger(__name__)CODE
LOW⚡acestep/training_v2/optim.py36logger = logging.getLogger(__name__)CODE
LOWacestep/training_v2/ui/presets.py29logger = logging.getLogger(__name__)CODE
LOWacestep/training_v2/ui/__init__.py42def set_plain_mode(value: bool) -> None:CODE
LOW⚡acestep/training_v2/ui/flows_setup.py27logger = logging.getLogger(__name__)CODE
LOWacestep/training_v2/ui/flows.py13__all__ = ["wizard_train", "wizard_preprocess", "wizard_estimate"]CODE
LOWacestep/training_v2/cli/config_builder.py20logger = logging.getLogger(__name__)CODE
LOWacestep/text_tasks/passphrase_store_test.py56 def set_password(service: str, username: str, passphrase: str) -> None:CODE
LOWacestep/text_tasks/secure_secret_store_test.py71 def set_password(service: str, username: str, secret: str) -> None:CODE
LOWacestep/text_tasks/secure_secret_store_test.py155 def set_password(*_args) -> None:CODE
Over-Commented Block66 hits · 52 pts
SeverityFileLineSnippetContext
LOWinstall_uv.sh1#!/usr/bin/env bashCOMMENT
LOWdocker-compose.jetson.yml1# =============================================================================COMMENT
LOWdocker-compose.jetson.yml21#COMMENT
LOWdocker-compose.jetson.yml61 # acestep-v15-base — Full base model (Extract/Lego/Complete)COMMENT
LOWstart_api_server.sh21# Preferred download source: auto (default), huggingface, or modelscopeCOMMENT
LOWrequirements-rocm.txt1# ACE-Step 1.5 - AMD ROCm 7.2 Windows DependenciesCOMMENT
LOWstart_gradio_ui_rocm.sh1#!/usr/bin/env bashCOMMENT
LOWstart_api_server_rocm.sh1#!/usr/bin/env bashCOMMENT
LOWstart_api_server_rocm.sh41# API key for authentication (optional)COMMENT
LOWstart_gradio_ui.sh101# LLM (Language Model) initialization settingsCOMMENT
LOWdocker-compose.yml1# =============================================================================COMMENT
LOWstart_gradio_ui_rocm_manual.sh1#!/usr/bin/env bashCOMMENT
LOWstart_gradio_ui_manual.sh221COMMENT
LOWstart_api_server_macos.sh1#!/usr/bin/env bashCOMMENT
LOWstart_api_server_macos.sh21PORT=8001COMMENT
LOWrequirements-sidestep.txt1# ╔══════════════════════════════════════════════════════════════╗COMMENT
LOWrequirements-sidestep.txt21# Uncomment to enable the Prodigy optimizer.COMMENT
LOWacestep/constants.py61# Valid time signatures - common musical meter patternsCOMMENT
LOWacestep/constants.py81# - cover: Style transfer using reference audioCOMMENT
LOWacestep/constants.py141COMMENT
LOWacestep/inference.py581 need_audio_codes = not user_provided_audio_codesCOMMENT
LOWacestep/inference.py621 use_random_seed_for_dit = FalseCOMMENT
LOWacestep/debug_utils.py21COMMENT
LOWacestep/third_parts/nano-vllm/nanovllm/utils/context.py21# `_CONTEXT` variable. In concurrent serving scenarios (API server withCOMMENT
LOW…/core/generation/handler/service_generate_flow_edit.py81 src_text_hs, src_lyric_hs = embed_source(handler, src_text_ids, src_lyric_ids)COMMENT
LOWacestep/models/__init__.py1# ACE-Step model definitionsCOMMENT
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py1# Copyright 2025 The ACESTEO Team. All rights reserved.COMMENT
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py101 # Causal sliding: only attend to past window stepsCOMMENT
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py801 # 如果有 padding mask (attention_mask [B, L])，直接传给它即可。COMMENT
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py1# Copyright 2025 The ACESTEO Team. All rights reserved.COMMENT
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py101 # Causal sliding: only attend to past window stepsCOMMENT
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py801 # 如果有 padding mask (attention_mask [B, L])，直接传给它即可。COMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py1# Copyright 2025 The ACESTEO Team. All rights reserved.COMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py101 # (B) Handle sliding windowCOMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py641COMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py801 # 场景 A: Flash Attention 模式COMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py1121 # 如果有 padding mask (attention_mask [B, L])，直接传给它即可。COMMENT
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py1401 # 判断是否使用 Flash Attention 2COMMENT
LOWacestep/models/common/configuration_acestep_v15.py1# coding=utf-8COMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py1# Copyright 2025 The ACESTEO Team. All rights reserved.COMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py101 # (B) Handle sliding windowCOMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py641COMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py801 # 场景 A: Flash Attention 模式COMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py1121 # 如果有 padding mask (attention_mask [B, L])，直接传给它即可。COMMENT
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py1401 # 判断是否使用 Flash Attention 2COMMENT
LOWacestep/models/sft/modeling_acestep_v15_base.py1# Copyright 2025 The ACESTEO Team. All rights reserved.COMMENT
LOWacestep/models/sft/modeling_acestep_v15_base.py101 # (B) Handle sliding windowCOMMENT
LOWacestep/models/sft/modeling_acestep_v15_base.py641COMMENT
LOWacestep/models/sft/modeling_acestep_v15_base.py801 # 场景 A: Flash Attention 模式COMMENT
LOWacestep/models/mlx/vae_model.py41 # weights (exp overflows float16 at alpha > ~11). This is only a problemCOMMENT
LOWacestep/models/mlx/dit_generate.py1# MLX diffusion generation loop for AceStep DiT decoder.COMMENT
LOWacestep/models/base/modeling_acestep_v15_base.py1# Copyright 2025 The ACESTEO Team. All rights reserved.COMMENT
LOWacestep/models/base/modeling_acestep_v15_base.py101 # (B) Handle sliding windowCOMMENT
LOWacestep/models/base/modeling_acestep_v15_base.py641COMMENT
LOWacestep/models/base/modeling_acestep_v15_base.py801 # 场景 A: Flash Attention 模式COMMENT
LOW…-transcription/scripts/acestep-lyrics-transcription.sh1#!/bin/bashCOMMENT
LOW.claude/skills/acestep/scripts/acestep.sh1#!/bin/bashCOMMENT
LOW…/skills/acestep-thumbnail/scripts/acestep-thumbnail.sh1#!/bin/bashCOMMENT
LOW.claude/skills/acestep-simplemv/scripts/render-mv.sh1#!/bin/bashCOMMENT
LOW.claude/skills/acestep-simplemv/scripts/render.sh1#!/bin/bashCOMMENT
6 more matches not shown…
AI Structural Patterns49 hits · 33 pts
SeverityFileLineSnippetContext
LOWopenrouter/openrouter_api_server.py1124CODE
LOWacestep/llm_inference.py874CODE
LOWacestep/llm_inference.py1145CODE
LOWacestep/llm_inference.py1282CODE
LOWacestep/llm_inference.py2042CODE
LOWacestep/llm_inference.py3960CODE
LOWacestep/inference.py1290CODE
LOWacestep/ui/gradio/events/results/batch_queue.py14CODE
LOWacestep/ui/gradio/events/results/generation_progress.py42CODE
LOWacestep/third_parts/nano-vllm/nanovllm/models/qwen3.py88CODE
LOWacestep/core/generation/handler/service_generate.py21CODE
LOW…step/core/generation/handler/generate_music_request.py222CODE
LOWacestep/core/generation/handler/diffusion.py18CODE
LOW…step/core/generation/handler/generate_music_execute.py18CODE
LOW…p/core/generation/handler/init_service_orchestrator.py48CODE
LOW…ep/core/generation/handler/service_generate_execute.py72CODE
LOWacestep/core/generation/handler/conditioning_batch.py21CODE
LOWacestep/core/generation/handler/generate_music.py195CODE
LOW…tep/core/generation/handler/conditioning_masks_test.py25CODE
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py478CODE
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py1307CODE
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py1701CODE
LOWacestep/models/turbo/modeling_acestep_v15_turbo.py1852CODE
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py478CODE
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py1323CODE
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py1713CODE
LOW…estep/models/xl_turbo/modeling_acestep_v15_xl_turbo.py1864CODE
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py481CODE
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py1326CODE
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py1716CODE
LOWacestep/models/xl_sft/modeling_acestep_v15_xl_base.py1848CODE
LOWacestep/models/common/configuration_acestep_v15.py148CODE
LOWacestep/models/common/flow_edit_pipeline.py76CODE
LOWacestep/models/common/flow_edit.py54CODE
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py481CODE
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py1326CODE
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py1716CODE
LOWacestep/models/xl_base/modeling_acestep_v15_xl_base.py1848CODE
LOWacestep/models/sft/modeling_acestep_v15_base.py481CODE
LOWacestep/models/sft/modeling_acestep_v15_base.py1310CODE
LOWacestep/models/sft/modeling_acestep_v15_base.py1704CODE
LOWacestep/models/sft/modeling_acestep_v15_base.py1836CODE
LOWacestep/models/mlx/dit_model.py432CODE
LOWacestep/models/mlx/dit_generate.py168CODE
LOWacestep/models/base/modeling_acestep_v15_base.py481CODE
LOWacestep/models/base/modeling_acestep_v15_base.py1310CODE
LOWacestep/models/base/modeling_acestep_v15_base.py1704CODE
LOWacestep/models/base/modeling_acestep_v15_base.py1836CODE
LOWacestep/training_v2/ui/help_formatter.py94CODE
Self-Referential Comments9 hits · 30 pts
SeverityFileLineSnippetContext
MEDIUM⚡openrouter/client_test.py22# This file is an executable API client script, not a pytest test module.COMMENT
MEDIUMacestep/audio_utils.py83 # Create a copy to avoid modifying original in-placeCOMMENT
MEDIUMacestep/inference.py965 # Create a copy of params dict for this audioCOMMENT
MEDIUMacestep/ui/gradio/events/training/preprocess_test.py34 # Create a dummy .pt fileCOMMENT
MEDIUMacestep/ui/gradio/events/results/generation_info.py16# This file is in acestep/ui/gradio/events/results/, need 6 levels up to project rootCOMMENT
MEDIUMacestep/ui/gradio/events/generation/metadata_loading.py160 # This file is in acestep/ui/gradio/events/generation/COMMENT
MEDIUMacestep/ui/gradio/events/generation/service_init.py123 # This file is in acestep/ui/gradio/events/generation/COMMENT
MEDIUM⚡acestep/training/data_module.py26 # Create a dummy class for type hintsCOMMENT
MEDIUMacestep/training/data_module_test.py66 # Create a manifest with one good and one bad pathCOMMENT
Example Usage Blocks8 hits · 14 pts
SeverityFileLineSnippetContext
LOWinstall_uv.sh5# Usage:COMMENT
LOW⚡docker-compose.jetson.yml9# Usage:COMMENT
LOW⚡docker-compose.yml5# Usage:COMMENT
LOW…-transcription/scripts/acestep-lyrics-transcription.sh7# Usage:COMMENT
LOW.claude/skills/acestep/scripts/acestep.sh7# Usage:COMMENT
LOW…/skills/acestep-thumbnail/scripts/acestep-thumbnail.sh4# Usage:COMMENT
LOW.claude/skills/acestep-simplemv/scripts/render-mv.sh4# Usage:COMMENT
LOW.claude/skills/acestep-simplemv/scripts/render.sh4# Usage:COMMENT
AI Slop Vocabulary5 hits · 11 pts
SeverityFileLineSnippetContext
LOW⚡acestep/constrained_logits_processor.py1793 # Avoid infinite recursion: if we're still in a fixed_strings state, just return scoresCOMMENT
LOW⚡acestep/constrained_logits_processor.py1795 # This shouldn't happen, but if it does, just return scores to avoid recursionCOMMENT
MEDIUM⚡acestep/ui/gradio/i18n/en.json518 "generation_repaint": "## Repaint Mode Tutorial\n\n**Best for:** Fixing or regenerating specific sections of generatCODE
LOW…ep/third_parts/nano-vllm/nanovllm/engine/llm_engine.py65 # For now, just use the same prompt (user should provide unconditional_prompt)COMMENT
MEDIUMacestep/core/scoring/lm_score.py339 # Calculate recall using the robust logicCOMMENT
Modern AI Meta-Vocabulary3 hits · 10 pts
SeverityFileLineSnippetContext
MEDIUMacestep/ui/gradio/i18n/pt.json513 "generation_edit": "## Edit (Overlay Flow-Edit)\n\n**Ideal para:** transformar um áudio existente em direção a uma nCODE
MEDIUM⚡acestep/ui/gradio/i18n/en.json526 "generation_edit": "## Edit (Flow-Edit Overlay)\n\n**Best for:** morphing an existing audio toward new lyrics or a nCODE
MEDIUM⚡acestep/training_v2/trainer_fixed.py68# FixedLoRATrainer -- orchestrationCOMMENT
Slop Phrases2 hits · 5 pts
SeverityFileLineSnippetContext
LOW⚡acestep/llm_inference.py2907 # Don't forget to save the last fieldCOMMENT
MEDIUM.github/workflows/codeql.yml54 # If you are analyzing a compiled language, you can modify the 'build-mode' for that language to customize howCOMMENT