Repository Analysis

skypilot-org/skypilot

Run, manage, and scale AI workloads on any AI infrastructure. Use one system to access & manage all AI compute (Kubernetes, Slurm, 20+ clouds, on-prem).

26.2 Moderate AI signal View on GitHub

Analysis Overview

This report presents the forensic synthetic code analysis of skypilot-org/skypilot, a Python project with 10,298 GitHub stars. SynthScan v2.0 examined 571,032 lines of code across 1980 source files, recording 11801 pattern matches distributed across 25 syntactic categories. The overall adjusted score of 26.2 places this repository in the Moderate AI signal band.

The scanner applied 160+ deterministic lexical heuristics, multi-line block detectors, abstract syntax tree depth profilers, and a cross-file Jaccard similarity matrix to construct a statistically normalised synthetic code estimate. All matches are individually weighted by severity coefficient and contextual multiplier before summation, and the resulting headline score is temporally discounted to account for the repository's development history relative to the commercial emergence of large language model coding tooling (November 2022 onward).

26.2
Adjusted Score
26.2
Raw Score
100%
Time Factor
2026-07-14
Last Push
10.3K
Stars
Python
Language
571.0K
Lines of Code
2.0K
Files
11.8K
Pattern Hits
2026-07-14
Scan Date
0.33
HC Hit Rate

What These Metrics Mean

Adjusted Score
Primary synthetic code indicator. Raw score normalised per 1,000 lines of code and multiplied by the temporal discount factor. This is the definitive comparative metric — use it to rank repositories by AI authorship density.
Raw Score
The unmodified sum of all severity-weighted, context-multiplied pattern match scores before temporal discounting. Reflects the absolute signal strength independent of when the repository was last active.
Time Factor
The temporal discount multiplier (0–100%) applied to the raw score. Repositories last updated before ChatGPT's launch (Nov 2022) receive a 5% factor. Full signal is only assigned to repositories active in the post-adoption era (Jan 2024+).
Pattern Hits
Total count of individual pattern matches across all files and categories. A high hit count with a low score may indicate a very large codebase with isolated AI snippets; a low count with a high score indicates dense, concentrated AI signatures.
HC Hit Rate
High+Critical pattern hits per file, averaged across the repository. This orthogonal signal catches repositories where a few files are densely packed with high-severity AI tells — a strong indicator even when the normalised score appears moderate due to codebase size.
Lines of Code / Files
Total lines and files analysed. The scanner examines 94 file extensions. These denominators are used to normalise the score, enabling fair comparison between repositories of vastly different sizes.

Score History

This chart maps the temporal evolution of the adjusted synthetic code score across successive scan runs. An upward trajectory indicates ongoing incorporation of AI-generated code or expanding LLM-assisted scaffolding; a stable or declining trajectory may reflect active human refactoring, code removal, or the adoption of stricter authorship policies. The dashed secondary line (right axis) independently tracks total raw pattern hit count, which can diverge from the normalised score when codebase size changes significantly between scans.

Severity Breakdown

Classifies detected patterns by their diagnostic confidence and structural impact. CRITICAL patterns (coefficient 10) represent definitive synthetic signatures — hallucinated imports, explicit LLM attribution metadata — virtually never produced by human authors. HIGH (5) indicates strong structural tells such as cross-file repetition or cross-linguistic idioms. MEDIUM (2) covers recognisable conversational padding and AI-specific vocabulary. LOW (1) captures subtle indicators like tautological comments and generic boilerplate that require density to carry independent signal.

CRITICAL 2HIGH 647MEDIUM 749LOW 10403

Directory Score Breakdown

This horizontal bar chart decomposes the repository's raw synthetic code score by top-level directory, allowing you to pinpoint precisely which modules or components carry the highest AI authorship density. Directories with disproportionately high scores relative to their size warrant targeted manual review: concentrated AI signatures often trace back to mass-generated configuration layers, auto-ported test suites, LLM-scaffolded boilerplate classes, or entire subsystems authored under heavy copilot assistance. Use this view to prioritise your human code-review effort.

Pattern Findings

The scanner identified 11801 distinct pattern matches across 25 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a statistically significant AI authorship indicator. 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 diagnostic confidence level (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — comment-context matches receive a ×1.5 weight because LLMs systematically over-annotate. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier as dense clusters constitute far stronger evidence of synthetic authorship than isolated hits.

Hyper-Verbose Identifiers6511 hits · 5588 pts
SeverityFileLineSnippetContext
LOWllm/rl-post-training-jobgroup/code/trainer.py99 def save_static_policy_sync_assets(self):CODE
LOWllm/rl-post-training-jobgroup/code/trainer.py211 def sample_from_replay_buffer(self,CODE
LOWllm/vicuna-llama-2/scripts/flash_attn_patch.py102def _prepare_decoder_attention_mask(self, attention_mask, input_shape,CODE
LOWllm/vicuna-llama-2/scripts/flash_attn_patch.py108def replace_llama_attn_with_flash_attn():CODE
LOWllm/vicuna-llama-2/scripts/xformers_patch.py34def replace_llama_attn_with_xformers_attn():CODE
LOWllm/vicuna-llama-2/scripts/train.py87def safe_save_model_for_hf_trainer(trainer: transformers.Trainer,CODE
LOWllm/vicuna-llama-2/scripts/train.py244def make_supervised_data_module(tokenizer: transformers.PreTrainedTokenizer,CODE
LOWllm/vicuna-llama-2/scripts/train.py276def cleanup_incomplete_checkpoints(output_dir):CODE
LOWllm/batch_inference/scripts/monitor_progress.py356 def get_throughput_chart_data(self) -> Dict:CODE
LOWllm/batch_inference/scripts/monitor_progress.py62 def update_throughput_history(self, worker_id: str, metrics: Dict):CODE
LOWllm/batch_inference/scripts/monitor_progress.py598 def get_token_throughput_chart_data(self) -> Dict:CODE
LOWllm/rl-code-execution-sandbox/code/trainer.py103 def save_static_policy_sync_assets(self):CODE
LOWllm/rl-code-execution-sandbox/code/trainer.py225 def sample_from_replay_buffer(self,CODE
LOWsky/task.py129def _check_docker_login_config(task_envs: Dict[str, str],CODE
LOWsky/task.py187def _with_docker_login_config(CODE
LOWsky/task.py226def _with_docker_username_for_runpod(CODE
LOWsky/task.py247def get_plaintext_envs_and_secrets(CODE
LOWsky/task.py556 def expand_and_validate_file_mounts(self):CODE
LOWsky/task.py603 def _validate_mount_dest_is_absolute(self, path: str, location: str):CODE
LOWsky/task.py616 def expand_and_validate_workdir(self):CODE
LOWsky/task.py1045 def resolve_and_validate_volumes(self) -> None:CODE
LOWsky/task.py1329 def get_estimated_inputs_size_gigabytes(self) -> Optional[float]:CODE
LOWsky/task.py1354 def get_estimated_outputs_size_gigabytes(self) -> Optional[float]:CODE
LOWsky/task.py1358 def _ensure_consistent_priority(CODE
LOWsky/task.py1865 def get_local_to_remote_file_mounts(self) -> Optional[Dict[str, str]]:CODE
LOWsky/task.py1886 def get_cloud_to_remote_file_mounts(self) -> Optional[Dict[str, str]]:CODE
LOWsky/task.py1929 def update_envs_and_secrets_from_workdir(self) -> 'Task':CODE
LOWsky/task.py2118 def get_required_cloud_features(CODE
LOWsky/task.py2202def _factor_out_common_resource_fields(configs: List[Dict[str, Union[str,CODE
LOWsky/sky_logging.py250def generate_tmp_logging_file_path(file_name: str) -> str:CODE
LOWsky/execution.py62def _maybe_clone_disk_from_cluster(clone_disk_from: Optional[str],CODE
LOWsky/execution.py159def _compute_set_autostop_args_for_hooks_only_relaunch(CODE
LOWsky/check.py99def _get_workspace_allowed_clouds(workspace: str) -> List[str]:CODE
LOWsky/check.py122def _get_workspace_cloud_capabilities(CODE
LOWsky/check.py181 def _execute_check_logic_for_workspace(CODE
LOWsky/check.py191 def check_one_cloud_one_capability(CODE
LOWsky/check.py501def get_cached_enabled_clouds_or_refresh(CODE
LOWsky/check.py548def get_cloud_credential_file_mounts(CODE
LOWsky/global_user_state.py365def _sqlite_supports_returning() -> bool:CODE
LOWsky/global_user_state.py658def set_user_preferred_workspace(user_id: str,CODE
LOWsky/global_user_state.py1078def get_terminal_or_last_status_change_event(CODE
LOWsky/global_user_state.py1100def _get_last_or_terminal_cluster_event_multiple(CODE
LOWsky/global_user_state.py1138def get_last_cluster_event_of_type_multiple(CODE
LOWsky/global_user_state.py1177def get_last_status_change_times(CODE
LOWsky/global_user_state.py1224def cleanup_cluster_events_with_retention(retention_hours: float,CODE
LOWsky/global_user_state.py1238async def cluster_event_retention_daemon():CODE
LOWsky/global_user_state.py1386def get_cluster_events_by_name(CODE
LOWsky/global_user_state.py1429def _get_user_hash_or_current_user(user_hash: Optional[str]) -> str:CODE
LOWsky/global_user_state.py1531def get_handle_from_cluster_name(CODE
LOWsky/global_user_state.py1544def get_handles_from_cluster_names(CODE
LOWsky/global_user_state.py1568def get_cluster_name_to_handle_map(CODE
LOWsky/global_user_state.py1587async def get_status_from_cluster_name_async(CODE
LOWsky/global_user_state.py1604def get_status_from_cluster_name(CODE
LOWsky/global_user_state.py1659def set_cluster_autostop_value(cluster_name: str, idle_minutes: int,CODE
LOWsky/global_user_state.py1697def get_cluster_provision_log_path(cluster_name: str) -> Optional[str]:CODE
LOWsky/global_user_state.py1708def get_all_cluster_provision_log_paths() -> List[str]:CODE
LOWsky/global_user_state.py1717def get_cluster_history_provision_log_path(cluster_name: str) -> Optional[str]:CODE
LOWsky/global_user_state.py1771def get_cluster_storage_mounts_metadata(CODE
LOWsky/global_user_state.py1783def set_cluster_storage_mounts_metadata(CODE
LOWsky/global_user_state.py1799def get_cluster_skylet_ssh_tunnel_metadata(CODE
6451 more matches not shown…
Cross-Language Confusion384 hits · 2017 pts
SeverityFileLineSnippetContext
HIGHllm/batch_inference/scripts/monitor_progress.py974 if (context.parsed.y !== null) {{CODE
HIGHllm/batch_inference/scripts/monitor_progress.py1034 if (context.parsed.y !== null) {{CODE
HIGHllm/batch_inference/scripts/monitor_progress.py1297 if (context.parsed.y !== null) {{CODE
HIGHllm/batch_inference/scripts/monitor_progress.py1359 if (event.key === 'F5' || (event.ctrlKey && event.key === 'r')) {{CODE
HIGHsky/task.py732 # ``secrets:`` (null) or array form have no inline values.COMMENT
HIGHsky/cloud_stores.py130 '2> /dev/null || true; 'CODE
HIGHsky/global_user_state.py884 # set owner to server default (null)COMMENT
HIGHsky/global_user_state.py886 # set storage_mounts_metadata to server default (null)COMMENT
HIGHsky/provision/docker_utils.py46 '(sudo apt-get update && sudo apt-get install -y unzip)) && 'CODE
HIGHsky/provision/docker_utils.py49 'unzip -q /tmp/awscliv2.zip -d /tmp && sudo /tmp/aws/install 'CODE
HIGHsky/provision/docker_utils.py444 'sudo systemctl stop jupyter > /dev/null 2>&1 || true;'CODE
HIGHsky/provision/docker_utils.py445 'sudo systemctl disable jupyter > /dev/null 2>&1 || true;'CODE
HIGHsky/provision/docker_utils.py446 'sudo systemctl stop jupyterhub > /dev/null 2>&1 || true;'CODE
HIGHsky/provision/docker_utils.py447 'sudo systemctl disable jupyterhub > /dev/null 2>&1 || true;',CODE
HIGHsky/provision/docker_utils.py348 '--quiet || true')CODE
HIGHsky/provision/docker_utils.py382 '{ which jq || sudo apt update && sudo apt install -y jq; } && 'CODE
HIGHsky/provision/docker_utils.py382 '{ which jq || sudo apt update && sudo apt install -y jq; } && 'CODE
HIGHsky/provision/docker_utils.py424 'flock -x -w 120 200 || exit 1; 'CODE
HIGHsky/provision/docker_utils.py480 f'echo "Port {port}" | sudo tee -a /etc/ssh/sshd_config > /dev/null;'CODE
HIGHsky/provision/docker_utils.py579 shm_output = self._run('cat /proc/meminfo || true')CODE
HIGHsky/provision/instance_setup.py43 'which prlimit && for id in $(pgrep -f raylet/raylet); 'CODE
HIGHsky/provision/instance_setup.py44 'do sudo prlimit --nofile=1048576:1048576 --pid=$id || true; done;')CODE
HIGHsky/provision/instance_setup.py135 f'2> /dev/null || echo {constants.SKY_REMOTE_RAY_PORT});'CODE
HIGHsky/provision/slurm/instance.py583 f'touch {container_init_done_dir}/$SLURM_PROCID && sleep infinity')STRING
HIGHsky/provision/slurm/instance.py606 f' if ! kill -0 $CONTAINER_PID 2>/dev/null; then\n'STRING
HIGHsky/provision/slurm/instance.py650 kill $(cat "{skypilot_runtime_dir}/.sky/skylet_pid") 2>/dev/null || trueCODE
HIGHsky/provision/slurm/instance.py657 srun --nodes={num_nodes} --ntasks-per-node=1 enroot remove -f {shlex.quote(_enroot_container_name_global_scope(clustCODE
HIGHsky/provision/paperspace/utils.py128 'if ! command -v docker &> /dev/null; then \n'CODE
HIGHsky/provision/paperspace/utils.py135 '$(. /etc/os-release && echo "$VERSION_CODENAME") stable" | \\\n'CODE
HIGHsky/provision/kubernetes/instance.py992 ' timeout 2 sudo -l >/dev/null 2>&1 && echo succeed || 'CODE
HIGHsky/provision/kubernetes/instance.py1016 'while ! grep -q "Fetched" /tmp/apt-update.log 2>/dev/null; do 'CODE
HIGHsky/provision/kubernetes/instance.py1018 ' cat /tmp/apt-update.log || true; 'CODE
HIGHsky/provision/kubernetes/utils.py3650 debug_cmd = 'kubectl describe ingress && kubectl describe ingressclass'CODE
HIGHsky/skylet/constants.py232 'sudo which nvidia-smi && echo "Checking Nvidia ECC Mode" && 'CODE
HIGHsky/skylet/constants.py239 '{ sudo reboot || echo "Failed to reboot. ECC mode may not be disabled"; } 'CODE
HIGHsky/skylet/constants.py93 f'cat {SKY_RAY_PATH_FILE} 2> /dev/null || which ray)')CODE
HIGHsky/skylet/constants.py99SKY_SLURM_UNSET_PYTHONPATH = ('$(which env 2>/dev/null || echo /usr/bin/env) 'CODE
HIGHsky/skylet/constants.py133 'unset VIRTUAL_ENV && unset VIRTUAL_ENV_PROMPT')CODE
HIGHsky/skylet/constants.py282 '{ conda init && source ~/.bashrc; };'CODE
HIGHsky/skylet/constants.py446 '(systemctl reload sshd || service ssh reload); 'CODE
HIGHsky/skylet/job_lib.py491 # status does not need to be set if the end_at is not null, sinceCOMMENT
HIGHsky/skylet/job_lib.py396 'INSERT INTO jobs VALUES (null, ?, ?, ?, ?, ?, ?, null, ?, 0, null, ?, null)', # pylint: disable=line-too-lCODE
HIGHsky/skylet/job_lib.py401 'INSERT INTO jobs VALUES (null, ?, ?, ?, ?, ?, ?, null, ?, 0, null, ?)', # pylint: disable=line-too-longCODE
HIGHsky/skylet/log_lib.py371 . $(conda info --base 2> /dev/null)/etc/profile.d/conda.sh > /dev/null 2>&1 || trueCODE
HIGHsky/skylet/log_lib.py371 . $(conda info --base 2> /dev/null)/etc/profile.d/conda.sh > /dev/null 2>&1 || trueCODE
HIGHsky/skylet/providers/command_runner.py290 'sudo systemctl stop jupyter > /dev/null 2>&1 || true;'CODE
HIGHsky/skylet/providers/command_runner.py291 'sudo systemctl disable jupyter > /dev/null 2>&1 || true;'CODE
HIGHsky/skylet/providers/command_runner.py292 'sudo systemctl stop jupyterhub > /dev/null 2>&1 || true;'CODE
HIGHsky/skylet/providers/command_runner.py293 'sudo systemctl disable jupyterhub > /dev/null 2>&1 || true;',CODE
HIGHsky/skylet/ray_patches/__init__.py47 which patch >/dev/null 2>&1 || sudo yum install -y patch || trueCODE
HIGHsky/backends/docker_utils.py23RUN apt-get update && apt-get -y install sudoCODE
HIGHsky/backends/docker_utils.py31 '/dev/null || true 'CODE
HIGHsky/backends/docker_utils.py56 setup_command: commands to run for setup. eg. "pip install numpy && aptSTRING
HIGHsky/backends/backend_utils.py618 # it present but null). This happens when restarting a clusterCOMMENT
HIGHsky/backends/task_codegen.py164 tac $file | grep "vfs cache: cleaned:" -m 1 | grep "in use 0, to upload 0, uploading 0" -q || exitcoCODE
HIGHsky/backends/task_codegen.py168 CACHE_STATUS=$(tac $file | grep "vfs cache: cleaned:" -m 1 | sed 's/.*vfs cache: cleaned: //' 2>CODE
HIGHsky/backends/task_codegen.py170 UPLOADING_FILES=$(tac $file | head -30 | grep -E "queuing for upload" | head -2 | sed 's/.*INFO CODE
HIGHsky/backends/task_codegen.py178 LAST_LINE=$(tac $file | grep -v "^$" | head -1 | sed 's/.*INFO : //' | sed 's/.*ERROR : //'CODE
HIGHsky/backends/local_docker_backend.py62 container with a sleep task (tail -f /dev/null). (Task run commands areSTRING
HIGHsky/backends/playground/demo_dockerutils.py12 setup_command = 'apt-get update && apt-get install htop && pip install 'CODE
324 more matches not shown…
Decorative Section Separators322 hits · 1016 pts
SeverityFileLineSnippetContext
MEDIUMllm/vicuna-llama-2/scripts/train.py17# ==============================================================================COMMENT
MEDIUMsky/core.py59# ======================COMMENT
MEDIUMsky/core.py61# ======================COMMENT
MEDIUMsky/core.py1201# ==================COMMENT
MEDIUMsky/core.py1203# ==================COMMENT
MEDIUMsky/core.py1621# ======================COMMENT
MEDIUMsky/core.py1623# ======================COMMENT
MEDIUMsky/core.py1663# ===================COMMENT
MEDIUMsky/core.py1665# ===================COMMENT
MEDIUMsky/core.py1939# =================COMMENT
MEDIUMsky/core.py1941# =================COMMENT
MEDIUMsky/provision/seeweb/instance.py505# =============================================================================COMMENT
MEDIUMsky/provision/seeweb/instance.py507# =============================================================================COMMENT
MEDIUMsky/backends/backend_utils.py3370# =====================================COMMENT
MEDIUMsky/catalog/common.py921# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/catalog/common.py923# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/server/requests/payloads.py1066# =============================================================================COMMENT
MEDIUMsky/server/requests/payloads.py1068# =============================================================================COMMENT
MEDIUMsky/utils/controller_utils.py1034 # ================================================================COMMENT
MEDIUMsky/utils/controller_utils.py1036 # ================================================================COMMENT
MEDIUMsky/workspaces/core.py66# =========================COMMENT
MEDIUMsky/workspaces/core.py68# =========================COMMENT
MEDIUMsky/workspaces/core.py613# =========================COMMENT
MEDIUMsky/workspaces/core.py615# =========================COMMENT
MEDIUMsky/workspaces/core.py1097# ===========================COMMENT
MEDIUMsky/workspaces/core.py1099# ===========================COMMENT
MEDIUMsky/dashboard/src/lib/analytics.js17// ── Core provider-backed tracking ───────────────────────────────────────────COMMENT
MEDIUMsky/dashboard/src/lib/analytics.js27// ── Domain helpers (thin wrappers over trackEvent) ──────────────────────────COMMENT
MEDIUMsky/recipes/db.py162# =============================================================================COMMENT
MEDIUMsky/recipes/db.py164# =============================================================================COMMENT
MEDIUMsky/recipes/db.py242# =============================================================================COMMENT
MEDIUMsky/recipes/db.py244# =============================================================================COMMENT
MEDIUMsky/batch/worker.py27# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py29# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py49# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py51# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py62# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py64# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py155# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py157# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py190# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py192# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py282# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py284# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py306# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/worker.py308# ---------------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py228 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py230 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py253 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py255 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py358 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py360 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py456 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py458 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py623 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py625 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py739 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py741 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py825 # ------------------------------------------------------------------COMMENT
MEDIUMsky/batch/coordinator.py827 # ------------------------------------------------------------------COMMENT
262 more matches not shown…
Excessive Try-Catch Wrapping925 hits · 920 pts
SeverityFileLineSnippetContext
LOWsetup.py90 except Exception as e: # pylint: disable=broad-exceptCODE
LOWsetup.py109 except Exception as e: # pylint: disable=broad-exceptCODE
LOWsetup.py122 except Exception as e: # pylint: disable=broad-exceptCODE
MEDIUMsetup.py117def revert_commit_hash():CODE
LOWllm/train-eval-jobgroup/code/evaluator.py176 except Exception as e:CODE
MEDIUMllm/train-eval-jobgroup/code/evaluator.py177 print(f"Error evaluating {checkpoint_path}: {e}")CODE
LOWllm/rl-post-training-jobgroup/code/trainer.py131 except Exception as e:CODE
LOWllm/rl-post-training-jobgroup/code/trainer.py167 except Exception as e:CODE
MEDIUMllm/rl-post-training-jobgroup/code/trainer.py168 print(f"Error generating response: {e}")CODE
LOWllm/rl-post-training-jobgroup/code/trainer.py208 except Exception as e:CODE
LOWllm/rl-post-training-jobgroup/code/trainer.py228 except Exception as e:CODE
LOWllm/rl-post-training-jobgroup/code/trainer.py242 except Exception as e:CODE
LOWllm/rl-post-training-jobgroup/code/trainer.py339 except Exception as e:CODE
LOWllm/rl-post-training-jobgroup/code/trainer.py365 except Exception as e:CODE
LOWllm/rl-post-training-jobgroup/code/data_server.py78 except Exception as e:CODE
MEDIUMllm/rl-post-training-jobgroup/code/data_server.py79 print(f"Error loading dataset: {e}")CODE
LOWllm/rag/scripts/serve_rag.py100 except Exception as e:CODE
LOWllm/rag/scripts/serve_rag.py217 except Exception as e:CODE
LOWllm/rag/scripts/serve_rag.py243 except Exception as e:CODE
LOWllm/rag/scripts/serve_rag.py303 except Exception as e:CODE
LOWllm/rag/scripts/serve_rag.py343 except Exception as e:CODE
LOWllm/rag/scripts/compute_embeddings.py223 except Exception as e:CODE
LOWllm/rag/scripts/build_rag.py58 except Exception as e:CODE
LOWllm/rag/scripts/build_rag.py142 except Exception as e:CODE
MEDIUMllm/batch_inference/scripts/monitor_progress.py354 print(f"Error updating metrics: {e}")CODE
MEDIUMllm/batch_inference/scripts/monitor_progress.py230 print(f"Error reading history for worker {worker_id}: {e}")CODE
MEDIUMllm/batch_inference/scripts/monitor_progress.py271 print(f"Error reading metrics from {file}: {e}")CODE
LOWllm/batch_inference/scripts/monitor_progress.py353 except Exception as e:CODE
LOWllm/batch_inference/scripts/monitor_progress.py229 except Exception as e:CODE
LOWllm/batch_inference/scripts/monitor_progress.py270 except Exception as e:CODE
LOWllm/batch_inference/scripts/base_vector_processor.py152 except Exception as e:CODE
LOWllm/batch_inference/scripts/base_vector_processor.py161 except Exception as e:CODE
LOWllm/batch_inference/scripts/base_vector_processor.py198 except Exception as e:CODE
LOWllm/batch_inference/scripts/base_vector_processor.py221 except Exception as e:CODE
LOWllm/batch_inference/scripts/base_vector_processor.py245 except Exception as e:CODE
LOWllm/batch_inference/scripts/base_vector_processor.py248 except Exception as e:CODE
LOWllm/batch_inference/scripts/base_vector_processor.py329 except Exception as e:CODE
LOWllm/batch_inference/scripts/text_vector_processor.py136 except Exception as e:CODE
LOWllm/batch_inference/scripts/text_vector_processor.py220 except Exception as e:CODE
LOWllm/batch_inference/scripts/text_vector_processor.py248 except Exception as e:CODE
LOWllm/batch_inference/scripts/text_vector_processor.py470 except Exception as e:CODE
LOWllm/batch_inference/scripts/text_vector_processor.py505 except Exception as e:CODE
LOW…l-code-execution-sandbox/code/sandbox_reward_server.py153 except Exception as e: # pylint: disable=broad-exceptCODE
LOW…l-code-execution-sandbox/code/sandbox_reward_server.py212 except Exception as e: # pylint: disable=broad-exceptCODE
LOW…l-code-execution-sandbox/code/sandbox_reward_server.py219 except Exception as e2: # pylint: disable=broad-exceptCODE
LOW…l-code-execution-sandbox/code/sandbox_reward_server.py228 except Exception as e: # pylint: disable=broad-exceptCODE
LOWllm/rl-code-execution-sandbox/code/trainer.py135 except Exception: # pylint: disable=broad-exceptCODE
LOWllm/rl-code-execution-sandbox/code/trainer.py171 except Exception as e: # pylint: disable=broad-exceptCODE
MEDIUMllm/rl-code-execution-sandbox/code/trainer.py172 print(f"Error generating response: {e}")CODE
LOWllm/rl-code-execution-sandbox/code/trainer.py222 except Exception as e: # pylint: disable=broad-exceptCODE
LOWllm/rl-code-execution-sandbox/code/trainer.py242 except Exception as e: # pylint: disable=broad-exceptCODE
LOWllm/rl-code-execution-sandbox/code/trainer.py256 except Exception: # pylint: disable=broad-exceptCODE
LOWllm/rl-code-execution-sandbox/code/trainer.py353 except Exception as e: # pylint: disable=broad-exceptCODE
LOWllm/rl-code-execution-sandbox/code/trainer.py379 except Exception as e: # pylint: disable=broad-exceptCODE
LOWllm/rl-code-execution-sandbox/code/data_server.py173 except Exception as e: # pylint: disable=broad-exceptCODE
MEDIUMllm/rl-code-execution-sandbox/code/data_server.py174 print(f"Error loading dataset: {e}")CODE
LOWllm/distributed-rl-eval/src/eval_head.py186 except Exception as e:CODE
LOWllm/distributed-rl-eval/src/eval_head.py209 except Exception as e:CODE
MEDIUMllm/distributed-rl-eval/src/eval_head.py210 # print(f"Error getting SkyServe services: {e}")COMMENT
LOWllm/distributed-rl-eval/src/eval_head.py254 except Exception as e:CODE
865 more matches not shown…
Self-Referential Comments276 hits · 818 pts
SeverityFileLineSnippetContext
MEDIUMllm/rag/scripts/serve_rag.py126 # Create a set to track unique contentCOMMENT
MEDIUMllm/rag/scripts/serve_rag.py402 # Initialize global variablesCOMMENT
MEDIUMllm/rag/scripts/build_rag.py91 # Create a temporary directory for building the database. TheCOMMENT
MEDIUMllm/batch_inference/batch_compute_vectors.py84 # Create a unique worker IDCOMMENT
MEDIUMllm/batch_inference/scripts/monitor_progress.py462 # Create a dataset for each worker sessionCOMMENT
MEDIUMllm/batch_inference/scripts/monitor_progress.py485 # Create the dataset for this worker sessionCOMMENT
MEDIUMllm/batch_inference/scripts/monitor_progress.py705 # Create a bin key by rounding the timestamp to the nearest binCOMMENT
MEDIUMllm/batch_inference/scripts/monitor_progress.py735 # Create a single data point at the current time with a value of 0COMMENT
MEDIUMllm/batch_inference/scripts/text_vector_processor.py540 # Create the output directory if it doesn't existCOMMENT
MEDIUMsky/global_user_state.py1105 # Create a priority expression: TERMINAL (0) before STATUS_CHANGE (1)COMMENT
MEDIUMsky/optimizer.py129 # This function is effectful: mutates every node in 'dag' by settingCOMMENT
MEDIUMsky/optimizer.py566 # Define the decision variables.COMMENT
MEDIUMsky/optimizer.py923 # The following code is to generate the tableCOMMENT
MEDIUMsky/optimizer.py1083 # Create a temporary single-task DAG for optimizationCOMMENT
MEDIUMsky/provision/vsphere/instance.py301 # Create the customization specCOMMENT
MEDIUMsky/provision/vsphere/instance.py305 # Create a custom script to inject the ssh public key into the instanceCOMMENT
MEDIUMsky/provision/vsphere/vsphere_utils.py226 # Create the Field if not exsit, then set value.COMMENT
MEDIUMsky/provision/vsphere/common/vapiconnect.py63 # Create the stub for the session service and login by creating a session.COMMENT
MEDIUMsky/provision/vsphere/common/cls_api_helper.py58 # Create a local content library backed the VC datastoreCOMMENT
MEDIUMsky/provision/vsphere/common/cls_api_helper.py71 # Create a new library item in the content library for uploading theCOMMENT
MEDIUMsky/provision/vsphere/common/cls_api_helper.py151 # Create a library itemCOMMENT
MEDIUMsky/provision/vsphere/common/cls_api_helper.py158 # Create a new upload session for uploading the filesCOMMENT
MEDIUMsky/provision/vsphere/common/vim_utils.py22 # Create a set for the names for faster lookupsCOMMENT
MEDIUMsky/provision/oci/query_utils.py299 # Create the skypilot_vcn and related resourcesCOMMENT
MEDIUMsky/provision/oci/query_utils.py341 # Create a public subnet.COMMENT
MEDIUMsky/provision/gcp/instance_utils.py1131 # Create a new MIG with size 0 and resize it later for triggeringCOMMENT
MEDIUMsky/provision/gcp/config.py738 # Create a SkyPilot VPC network if it doesn't existCOMMENT
MEDIUMsky/provision/gcp/volume_utils.py17 # Define the default mapping from disk tiers to disk types.COMMENT
MEDIUMsky/provision/gcp/mig_utils.py71 # Create the regional instance template requestCOMMENT
MEDIUMsky/provision/runpod/api/commands.py1"""This module provides functions to generate GraphQL mutations for deployingSTRING
MEDIUMsky/provision/runpod/api/commands.py58 """This module provides functions to generate GraphQL mutations forSTRING
MEDIUMsky/provision/runpod/api/pods.py1"""This module provides functions to generate GraphQL mutations for deployingSTRING
MEDIUMsky/provision/primeintellect/utils.py124 # Create a more informative error messageCOMMENT
MEDIUMsky/provision/lambda_cloud/lambda_utils.py327 # Create a data structure that matches the API schemaCOMMENT
MEDIUMsky/provision/aws/config.py161 # Create the placement groupCOMMENT
MEDIUMsky/provision/kubernetes/instance.py571 # Create a set of pod names we're waiting forCOMMENT
MEDIUMsky/provision/kubernetes/instance.py729 # Create a set of pod names we're waiting forCOMMENT
MEDIUMsky/provision/kubernetes/config.py514 # Create the role in the skypilot-system namespace if it does not exist.COMMENT
MEDIUMsky/provision/kubernetes/config.py526 # Create the role binding in the skypilot-system namespace, and haveCOMMENT
MEDIUMsky/skylet/constants.py288 # Create a separate python environment for SkyPilot dependencies.COMMENT
MEDIUMsky/skylet/providers/ibm/vpc_provider.py778 # Define the function parametersCOMMENT
MEDIUMsky/skylet/providers/ibm/node_provider.py732 # Create a VPC when creating a head nodeCOMMENT
MEDIUMsky/backends/docker_utils.py175 # Create an empty dirCOMMENT
MEDIUMsky/backends/backend_utils.py2111 # Create a session with retry logicCOMMENT
MEDIUMsky/backends/local_docker_backend.py98 # Define the Docker-in-Docker mountCOMMENT
MEDIUMsky/catalog/common.py472 # The following code is redundant with the code in resources.py::_set_cpus()COMMENT
MEDIUMsky/catalog/common.py496 # The following code is partially redundant with the code inCOMMENT
MEDIUMsky/catalog/hyperbolic_catalog.py11# Initialize cloud variable at module levelCOMMENT
MEDIUMsky/catalog/data_fetchers/fetch_vsphere.py312 # Create an empty DataFrame with the desired columnsCOMMENT
MEDIUMsky/catalog/data_fetchers/fetch_seeweb.py273 # Create a row for each regionCOMMENT
MEDIUMsky/ssh_node_pools/deploy/deploy.py956 # Create a Kubernetes Service for dcgm-exporter with PrometheusSTRING
MEDIUMsky/ssh_node_pools/deploy/deploy.py672 # Create the directory for the kubeconfig file if it doesn't existSTRING
MEDIUMsky/ssh_node_pools/deploy/tunnel/ssh-tunnel.sh197 # Create a temporary file for the JSON output to avoid shell escaping issuesCOMMENT
MEDIUMsky/server/server.py1209 # Create the session with the tokenCOMMENT
MEDIUMsky/server/server.py2102 # Create a temporary zip fileCOMMENT
MEDIUMsky/server/requests/process.py73# Define the worker function outside of the class to avoid pickling selfCOMMENT
MEDIUMsky/server/requests/queues/mp_queue.py25 # Defining a local function instead of a lambda functionCOMMENT
MEDIUMsky/utils/directory_utils.py5# This file is in '<project_root>/sky/utils/directory_utils.py'COMMENT
MEDIUMsky/utils/context.py416 # Create a copy of the current contextvars Context so that setting theCOMMENT
MEDIUMsky/utils/context.py442 # Create a copy of the current contextvars Context so that setting theCOMMENT
216 more matches not shown…
Unused Imports886 hits · 795 pts
SeverityFileLineSnippetContext
LOWllm/vicuna-llama-2/scripts/flash_attn_patch.py2CODE
LOWllm/vicuna-llama-2/scripts/flash_attn_patch.py10CODE
LOWllm/rag/scripts/serve_rag.py8CODE
LOWllm/rag/scripts/serve_rag.py10CODE
LOWllm/rag/scripts/serve_rag.py10CODE
LOWllm/rag/scripts/serve_rag.py20CODE
LOWllm/rag/scripts/compute_embeddings.py8CODE
LOWllm/rag/scripts/build_rag.py7CODE
LOWllm/rag/scripts/build_rag.py19CODE
LOWllm/gpt-oss-finetuning/sft/train.py10CODE
LOWllm/batch_inference/batch_compute_vectors.py13CODE
LOWllm/batch_inference/batch_compute_vectors.py14CODE
LOWllm/batch_inference/batch_compute_vectors.py15CODE
LOWllm/batch_inference/scripts/base_vector_processor.py8CODE
LOWllm/batch_inference/scripts/base_vector_processor.py10CODE
LOWllm/batch_inference/scripts/base_vector_processor.py12CODE
LOWllm/batch_inference/scripts/base_vector_processor.py14CODE
LOWllm/batch_inference/scripts/text_vector_processor.py5CODE
LOWllm/batch_inference/scripts/text_vector_processor.py15CODE
LOWllm/distributed-rl-eval/src/eval_head.py31CODE
LOWllm/distributed-rl-eval/src/game_server.py9CODE
LOWllm/distributed-rl-eval/src/game_server.py10CODE
LOWsky/execution.py42CODE
LOWsky/execution.py43CODE
LOWsky/global_user_state.py44CODE
LOWsky/global_user_state.py45CODE
LOWsky/global_user_state.py47CODE
LOWsky/__init__.py93CODE
LOWsky/__init__.py94CODE
LOWsky/__init__.py96CODE
LOWsky/__init__.py97CODE
LOWsky/__init__.py98CODE
LOWsky/__init__.py99CODE
LOWsky/__init__.py100CODE
LOWsky/__init__.py101CODE
LOWsky/__init__.py102CODE
LOWsky/__init__.py103CODE
LOWsky/__init__.py104CODE
LOWsky/__init__.py105CODE
LOWsky/__init__.py106CODE
LOWsky/__init__.py107CODE
LOWsky/__init__.py108CODE
LOWsky/__init__.py109CODE
LOWsky/__init__.py110CODE
LOWsky/__init__.py111CODE
LOWsky/__init__.py112CODE
LOWsky/__init__.py113CODE
LOWsky/__init__.py114CODE
LOWsky/__init__.py115CODE
LOWsky/__init__.py117CODE
LOWsky/__init__.py118CODE
LOWsky/__init__.py119CODE
LOWsky/__init__.py120CODE
LOWsky/__init__.py121CODE
LOWsky/__init__.py122CODE
LOWsky/__init__.py123CODE
LOWsky/__init__.py124CODE
LOWsky/__init__.py125CODE
LOWsky/__init__.py126CODE
LOWsky/__init__.py127CODE
826 more matches not shown…
Cross-File Repetition132 hits · 660 pts
SeverityFileLineSnippetContext
HIGHllm/rag/batch_compute_embeddings.py0calculate the range of indices this job should process. args: start_idx: global start index end_idx: global end index joSTRING
HIGHllm/batch_inference/batch_compute_vectors.py0calculate the range of indices this job should process. args: start_idx: global start index end_idx: global end index joSTRING
HIGHexamples/vector_database/batch_compute_vectors.py0calculate the range of indices this job should process. args: start_idx: global start index end_idx: global end index joSTRING
HIGHsky/models.py0set state from pickled state, for backward compatibility.STRING
HIGHsky/resources.py0set state from pickled state, for backward compatibility.STRING
HIGHsky/serve/service_spec.py0set state from pickled state, for backward compatibility.STRING
HIGHsky/serve/replica_managers.py0set state from pickled state, for backward compatibility.STRING
HIGHsky/core.py0tears down the kubernetes cluster started by local_up.STRING
HIGHsky/server/server.py0tears down the kubernetes cluster started by local_up.STRING
HIGHsky/client/sdk.py0tears down the kubernetes cluster started by local_up.STRING
HIGHsky/skylet/job_lib.py0convert this python enum value to protobuf enum value.STRING
HIGHsky/skylet/autostop_lib.py0convert this python enum value to protobuf enum value.STRING
HIGHsky/jobs/state.py0convert this python enum value to protobuf enum value.STRING
HIGHsky/skylet/log_lib.py0process the stream of a process in threads, blocking.STRING
HIGH…s/testdata/slurm_codegen/slurm_single_node_with_gpu.py0process the stream of a process in threads, blocking.STRING
HIGH…testdata/slurm_codegen/slurm_codegen_with_container.py0process the stream of a process in threads, blocking.STRING
HIGH…/backends/testdata/ray_codegen/single_node_with_gpu.py0process the stream of a process in threads, blocking.STRING
HIGH…sky/backends/testdata/ray_codegen/multi_node_2nodes.py0process the stream of a process in threads, blocking.STRING
HIGHsky/skylet/log_lib.py0runs a command and logs its output to a file. args: cmd: the command to run. log_path: the path to the log file. stream_STRING
HIGH…s/testdata/slurm_codegen/slurm_single_node_with_gpu.py0runs a command and logs its output to a file. args: cmd: the command to run. log_path: the path to the log file. stream_STRING
HIGH…testdata/slurm_codegen/slurm_codegen_with_container.py0runs a command and logs its output to a file. args: cmd: the command to run. log_path: the path to the log file. stream_STRING
HIGH…/backends/testdata/ray_codegen/single_node_with_gpu.py0runs a command and logs its output to a file. args: cmd: the command to run. log_path: the path to the log file. stream_STRING
HIGH…sky/backends/testdata/ray_codegen/multi_node_2nodes.py0runs a command and logs its output to a file. args: cmd: the command to run. log_path: the path to the log file. stream_STRING
HIGHsky/skylet/log_lib.py0\ #!/bin/bash source ~/.bashrc set -a . $(conda info --base 2> /dev/null)/etc/profile.d/conda.sh > /dev/null 2>&1 || truSTRING
HIGH…s/testdata/slurm_codegen/slurm_single_node_with_gpu.py0\ #!/bin/bash source ~/.bashrc set -a . $(conda info --base 2> /dev/null)/etc/profile.d/conda.sh > /dev/null 2>&1 || truSTRING
HIGH…testdata/slurm_codegen/slurm_codegen_with_container.py0\ #!/bin/bash source ~/.bashrc set -a . $(conda info --base 2> /dev/null)/etc/profile.d/conda.sh > /dev/null 2>&1 || truSTRING
HIGH…/backends/testdata/ray_codegen/single_node_with_gpu.py0\ #!/bin/bash source ~/.bashrc set -a . $(conda info --base 2> /dev/null)/etc/profile.d/conda.sh > /dev/null 2>&1 || truSTRING
HIGH…sky/backends/testdata/ray_codegen/multi_node_2nodes.py0\ #!/bin/bash source ~/.bashrc set -a . $(conda info --base 2> /dev/null)/etc/profile.d/conda.sh > /dev/null 2>&1 || truSTRING
HIGHsky/catalog/verda_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/yotta_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/vsphere_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/vast_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/scp_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/lambda_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/nebius_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/seeweb_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/paperspace_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/runpod_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/oci_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/mithril_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/shadeform_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/fluidstack_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/do_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/primeintellect_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/cudo_catalog.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/clouds/primeintellect.py0returns the cost, or the cheapest cost among all zones for spot.STRING
HIGHsky/catalog/verda_catalog.py0returns a list of instance types that have the given accelerator.STRING
HIGHsky/catalog/yotta_catalog.py0returns a list of instance types that have the given accelerator.STRING
HIGHsky/catalog/paperspace_catalog.py0returns a list of instance types that have the given accelerator.STRING
HIGHsky/catalog/runpod_catalog.py0returns a list of instance types that have the given accelerator.STRING
HIGHsky/catalog/shadeform_catalog.py0returns a list of instance types that have the given accelerator.STRING
HIGHsky/catalog/do_catalog.py0returns a list of instance types that have the given accelerator.STRING
HIGHsky/catalog/primeintellect_catalog.py0returns a list of instance types that have the given accelerator.STRING
HIGHsky/catalog/vsphere_catalog.py0returns a list of instance types satisfying the required count of accelerators with sorted prices and a list of candidatSTRING
HIGHsky/catalog/fluidstack_catalog.py0returns a list of instance types satisfying the required count of accelerators with sorted prices and a list of candidatSTRING
HIGHsky/catalog/cudo_catalog.py0returns a list of instance types satisfying the required count of accelerators with sorted prices and a list of candidatSTRING
HIGHsky/catalog/scp_catalog.py0filter the instance types based on resource requirements. returns a list of instance types satisfying the required countSTRING
HIGHsky/catalog/lambda_catalog.py0filter the instance types based on resource requirements. returns a list of instance types satisfying the required countSTRING
HIGHsky/catalog/nebius_catalog.py0filter the instance types based on resource requirements. returns a list of instance types satisfying the required countSTRING
HIGHsky/catalog/azure_catalog.py0filter the instance types based on resource requirements. returns a list of instance types satisfying the required countSTRING
72 more matches not shown…
Docstring Block Structure124 hits · 620 pts
SeverityFileLineSnippetContext
HIGHsky/task.py131Validates a valid docker login config in task_envs and task_secrets. Docker login variables must be specified togetSTRING
HIGHsky/task.py1217Updates environment variables for use inside the setup/run commands. Args: envs: (optional) either a STRING
HIGHsky/task.py1265Updates secret env vars for use inside the setup/run commands. Args: secrets: (optional) either a lisSTRING
HIGHsky/task.py1495Updates the file mounts for this task. Different from set_file_mounts(), this function updates into the STRING
HIGHsky/task.py1533Sets the storage mounts for this task. Storage mounts are a dictionary: ``{mount_path: sky.Storage object}``, STRING
HIGHsky/task.py1598Updates the storage mounts for this task. Different from set_storage_mounts(), this function updates into the STRING
HIGHsky/cloud_stores.py217Returns whether 'url' of the AZ Container is a directory. In cloud object stores, a "directory" refers to a regSTRING
HIGHsky/execution.py716Launches a cluster or task. The task's setup and run commands are executed under the task's workdir (when speciSTRING
HIGHsky/execution.py935Executes a task on an existing cluster. This function performs two actions: (1) workdir syncing, if the task hSTRING
HIGHsky/global_user_state.py1934Resolve cluster_hash from either cluster_hash or cluster_name. Validates that exactly one of cluster_hash or clusteSTRING
HIGHsky/core.py72Finds the best execution plan for the given DAG. Args: dag: the DAG to optimize. minimize: whether STRING
HIGHsky/core.py378Gets the endpoint for a given cluster and port number (endpoint). Args: cluster: The name of the cluster. STRING
HIGHsky/core.py1489Tails per-event lifecycle-hook logs of a cluster. Args: cluster_name: name of the cluster. event: oSTRING
HIGHsky/core.py1535Downloads the logs of jobs. Args: cluster_name: (str) name of the cluster. job_ids: (List[str]) jobSTRING
HIGHsky/core.py1577Get the status of jobs. Args: cluster_name: (str) name of the cluster. job_ids: (List[str]) job idsSTRING
HIGHsky/core.py1833Gets Slurm real-time GPU availability grouped by partition. This function calls the Slurm backend to fetch GPU infoSTRING
HIGHsky/core.py1974Create a debug dump for troubleshooting. Args: request_ids: List of request IDs or prefixes to include in tSTRING
HIGHsky/admin_policy.py215Validates and mutates the user request and returns mutated request. Args: user_request: The user reSTRING
HIGHsky/admin_policy.py240Validates and mutates the user request and returns mutated request. Args: user_request: The user reSTRING
HIGHsky/optimizer.py1043Optimize a JobGroup DAG. This method optimizes all tasks in a JobGroup to run on the same infrastructurSTRING
HIGHsky/resources.py198Initialize a Resources object. All fields are optional. ``Resources.is_launchable`` decides whether thSTRING
HIGHsky/provision/mithril/utils.py408Wait for a bid to create instances and for them to have SSH destinations. Args: bid_name: The name of the bSTRING
HIGHsky/provision/mithril/utils.py688Update a spot bid to pause or unpause it. Args: bid_id: The FID of the bid to update. paused: True STRING
HIGHsky/provision/slurm/utils.py866Resolve a canonical GPU name to the raw GRES type on a Slurm cluster. Queries live node metadata and applies fuzzy STRING
HIGHsky/provision/primeintellect/utils.py193Create a pod/instance via Prime Intellect API. Args: name: User-visible name of the pod. STRING
HIGHsky/provision/aws/config.py723Find or create a security group in the specified VPC. Args: ec2: The initialized EC2 client object. STRING
HIGHsky/provision/kubernetes/utils.py2491Returns the label key and value for the given GPU/TPU type. Args: acc_type: The GPU/TPU type required by thSTRING
HIGHsky/metrics/utils.py371Starts a port forward to a service in a Kubernetes cluster. Args: context: Kubernetes context name nSTRING
HIGHsky/metrics/utils.py506Sends a metrics request to a Prometheus endpoint via port forwarding. Args: context: Kubernetes context nameSTRING
HIGHsky/metrics/utils.py646Get GPU metrics for a single Kubernetes context. Args: context: Kubernetes context name stats: OptioSTRING
HIGHsky/metrics/utils.py690Get Sky Endpoint serving-engine metrics for a single K8s context. Mirrors get_metrics_for_context() but federates tSTRING
HIGHsky/backends/backend_utils.py2419Check if the task is compatible to clone disk from the source cluster. Args: cluster_name: The name of the STRING
HIGHsky/backends/backend_utils.py3215Refresh the cluster, and return the possibly updated record. The function will update the cached cluster status in STRING
HIGHsky/backends/backend_utils.py3522Check if the jobs/serve controller is up. The controller is accessible when it is in UP or INIT state, and the ssh STRING
HIGHsky/backends/backend_utils.py4335Gets the endpoint for a given cluster and port number (endpoint). Args: cluster: The name of the cluster. STRING
HIGHsky/backends/cloud_vm_ray_backend.py5955Runs 'cmd' on the cluster's head node. It will try to fetch the head node IP if it is not cached. ArgsSTRING
HIGHsky/server/server.py3439Serves the Next.js dashboard application. Args: request: The incoming HTTP request (used to attach the CSP STRING
HIGHsky/server/common.py396Make an authenticated async HTTP request to the API server using aiohttp. Automatically handles service account tokSTRING
HIGHsky/server/common.py873Check if the API server is healthy. Args: endpoint (Optional[str]): The endpoint of the API server. STRING
HIGHsky/server/versions.py230Decorator to enforce a minimum remote API version for an SDK function. New SDK method must be decorated with this dSTRING
HIGHsky/utils/git.py98Parse git URL into components. Supports various git URL formats: - HTTPS: https://github.com/user/repo.STRING
HIGHsky/utils/context_utils.py187Decorator to make a synchronous function cancellable via context. Guards the function execution by checking contextSTRING
HIGHsky/utils/locks.py76Acquire the lock. Args: blocking: If True, block until lock is acquired or timeout. STRING
HIGHsky/utils/interactive_utils.py31Receive file descriptor via Unix socket using SCM_RIGHTS. Args: sock: Connected Unix socket. Returns: STRING
HIGHsky/utils/dag_utils.py440Load a JobGroup from a multi-document YAML file. JobGroup YAML format: --- name: my-job-group STRING
HIGHsky/utils/infra_utils.py39Parse the infra string into cloud, region, and zone components. The format of the infra string is `cloud`, `cloSTRING
HIGHsky/utils/subprocess_utils.py226Run a function in parallel on a list of arguments. Args: func: The function to run in parallel argsSTRING
HIGHsky/utils/common_utils.py570Redact sensitive values from --secret arguments. Args: argv: Command line arguments Returns: MSTRING
HIGHsky/utils/resources_utils.py112Validate and normalize a local_disk string to canonical form. Accepts various input formats and normalizes to 'modeSTRING
HIGHsky/utils/resources_utils.py206Check if launched local disk satisfies the requested requirement. Used by Resources.less_demanding_than() to validaSTRING
HIGHsky/utils/resources_utils.py579Parse a time string using the given unit multipliers. Args: time: Time string with optional unit suffix. STRING
HIGHsky/utils/resources_utils.py611Convert a time string to minutes. Args: time: Time string with optional unit suffix (e.g., '30m', '2h', '1dSTRING
HIGHsky/utils/resources_utils.py627Convert a time string to seconds. Args: time: Time string with optional unit suffix (e.g., '30s', '5m', '1hSTRING
HIGHsky/workspaces/core.py467Updates a specific workspace configuration. Args: workspace_name: The name of the workspace to update. STRING
HIGHsky/workspaces/core.py525Creates a new workspace configuration. Args: workspace_name: The name of the workspace to create. cSTRING
HIGHsky/workspaces/core.py573Deletes a workspace configuration. Args: workspace_name: The name of the workspace to delete. Returns:STRING
HIGHsky/workspaces/core.py630Updates the entire SkyPilot configuration. Args: config: The new configuration to save. Returns: STRING
HIGHsky/workspaces/core.py1163Resolves the effective workspace for a user when none was set. Precedence (a future admin-assignment tier can splicSTRING
HIGHsky/clouds/slurm.py884Validate region (cluster) and zone (partition). Args: region: Slurm cluster name. zone:STRING
HIGHsky/clouds/kubernetes.py1431Returns the user identity for a specific Kubernetes context. Args: context: The name of the KuberneSTRING
64 more matches not shown…
Redundant / Tautological Comments362 hits · 547 pts
SeverityFileLineSnippetContext
LOWllm/train-eval-jobgroup/code/evaluator.py181 # Check if training is complete (look for training_complete marker)COMMENT
LOWllm/rl-post-training-jobgroup/code/reward_server.py111 # Check if answers matchCOMMENT
LOWllm/rag/scripts/serve_rag.py313 # Check if the request exists and is ready for LLM processingCOMMENT
LOWllm/kimi-k2/kimi-k2.sky.yaml38 # Set VLLM_HOST_IP to the IP of the current node based on rankCOMMENT
LOWsky/task.py154 # Check if any docker variables existCOMMENT
LOWsky/task.py158 # Check if variables are split across envs and secretsCOMMENT
LOWsky/task.py167 # Check if all variables are present in the chosen locationCOMMENT
LOWsky/task.py857 # Check if it is str pathCOMMENT
LOWsky/task.py1068 # Check if this is an ephemeral volume config or external volumeCOMMENT
LOWsky/execution.py833 # Check if cluster exists and we are doing fast provisioningCOMMENT
LOWsky/execution.py1000 # Check if cluster is autostopping - reject exec on autostopping clustersCOMMENT
LOWsky/global_user_state.py2263 # Check if we need to fetch the current user's name,COMMENT
LOWsky/core.py1171 # Check if autostop/autodown is required and supportedCOMMENT
LOWsky/optimizer.py1440 # Check if there exists launchable resourcesCOMMENT
LOWsky/resources.py1516 # Check if the single key matches the region or is None (any region)COMMENT
LOWsky/resources.py1562 # Check if any of the specified images are not docker imagesCOMMENT
LOWsky/resources.py2238 # Check if we have any free resourcesCOMMENT
LOWsky/provision/docker_utils.py260 # Set returncode to 1 if failed to connect to dockerCOMMENT
LOWsky/provision/provisioner.py632 # Check if head node Ray is aliveCOMMENT
LOWsky/provision/provisioner.py662 # Check if head node Ray is aliveCOMMENT
LOWsky/provision/vsphere/instance.py153 # Check if the user want to create a cpu or a gpu instance by checking ifCOMMENT
LOWsky/provision/vsphere/instance.py159 # Check if the user want to create a cpu instanceCOMMENT
LOWsky/provision/mithril/instance.py97 # Check if there's a paused bid that needs to be resumedCOMMENT
LOWsky/provision/mithril/instance.py147 # Check if we have enough instancesCOMMENT
LOWsky/provision/mithril/utils.py266 # Check if we should retry based on status codeCOMMENT
LOWsky/provision/mithril/utils.py327 # Check if this key already existsCOMMENT
LOWsky/provision/slurm/instance.py376 # Check if job already existsCOMMENT
LOWsky/provision/slurm/instance.py969 # Check if we are running inside a Slurm cluster (only happens withCOMMENT
LOWsky/provision/slurm/utils.py622 # Check if gpu_partition_map redirects this GPU type to useCOMMENT
LOWsky/provision/slurm/utils.py670 # Check if CPU and memory requirements are met on at least oneCOMMENT
LOWsky/provision/hyperbolic/instance.py349 # Check if any instances are in ONLINE stateCOMMENT
LOWsky/provision/hyperbolic/instance.py354 # Check if any instances are in a failed stateCOMMENT
LOWsky/provision/hyperbolic/instance.py365 # Check if any instances are in TERMINATED stateCOMMENT
LOWsky/provision/hyperbolic/instance.py375 # Check if any instances are in TERMINATED stateCOMMENT
LOWsky/provision/hyperbolic/instance.py380 # Check if any instances are in a failed stateCOMMENT
LOWsky/provision/hyperbolic/instance.py391 # Check if any instances are in ONLINE stateCOMMENT
LOWsky/provision/cudo/cudo_wrapper.py120 # Check if this machine type matches our requirementsCOMMENT
LOWsky/provision/verda/instance.py167 # Check if it's a resource unavailability errorCOMMENT
LOWsky/provision/verda/instance.py321 # Check if all terminated instances are goneCOMMENT
LOWsky/provision/gcp/instance.py327 # Check if the number of running instances is the same as the requested.COMMENT
LOWsky/provision/gcp/instance.py500 # Check if the instance is actually stopped.COMMENT
LOWsky/provision/gcp/config.py180 # Check if we have any TPUs defined, and if so,COMMENT
LOWsky/provision/gcp/config.py533 # Check if the required ports are a subset of the effective portsCOMMENT
LOWsky/provision/gcp/config.py721 # Check if VPC for subnet has sufficient firewall rules.COMMENT
LOWsky/provision/gcp/config.py786 # Check if VPC existsCOMMENT
LOWsky/provision/gcp/config.py796 # Check if subnet existsCOMMENT
LOWsky/provision/gcp/config.py849 # Check if VPC existsCOMMENT
LOWsky/provision/gcp/constants.py148 # Check if filesystem is already formatted (ext4)COMMENT
LOWsky/provision/gcp/constants.py167 # Check if already mountedCOMMENT
LOWsky/provision/primeintellect/instance.py329 # Check if all terminated instances are goneCOMMENT
LOWsky/provision/primeintellect/utils.py66 # Check if it's a resource unavailability errorCOMMENT
LOWsky/provision/primeintellect/utils.py307 # Check if the public key is already addedCOMMENT
LOWsky/provision/lambda_cloud/lambda_utils.py310 # Check if this rule already exists to avoid duplicatesCOMMENT
LOWsky/provision/kubernetes/instance.py661 # Check if cluster is autoscaling and update spinner message.COMMENT
LOWsky/provision/kubernetes/instance.py771 # Check if pod is terminated/preempted/failed (unchanged).COMMENT
LOWsky/provision/kubernetes/instance.py1212 # Check if the error is due to the AppArmor annotation and retry.COMMENT
LOWsky/provision/kubernetes/instance.py1673 # Check if any PVCs with access mode ReadWriteOnce or ReadWriteOncePodCOMMENT
LOWsky/provision/kubernetes/instance.py2274 # Check if node is cordoned (unschedulable)COMMENT
LOWsky/provision/kubernetes/instance.py3072 # Check if the pods are running or pendingCOMMENT
LOWsky/provision/kubernetes/volume.py718 # Check if PV matches storage classCOMMENT
302 more matches not shown…
Deep Nesting704 hits · 532 pts
SeverityFileLineSnippetContext
LOWllm/train-eval-jobgroup/code/evaluator.py95CODE
LOWllm/rl-post-training-jobgroup/code/trainer.py107CODE
LOWllm/rl-post-training-jobgroup/code/trainer.py347CODE
LOWllm/rag/scripts/build_rag.py63CODE
LOWllm/batch_inference/scripts/monitor_progress.py62CODE
LOWllm/batch_inference/scripts/monitor_progress.py196CODE
LOWllm/batch_inference/scripts/monitor_progress.py233CODE
LOWllm/batch_inference/scripts/monitor_progress.py532CODE
LOWllm/batch_inference/scripts/monitor_progress.py598CODE
LOWllm/batch_inference/scripts/monitor_progress.py813CODE
LOWllm/batch_inference/scripts/base_vector_processor.py203CODE
LOWllm/batch_inference/scripts/text_vector_processor.py140CODE
LOWllm/rl-code-execution-sandbox/code/trainer.py111CODE
LOWllm/rl-code-execution-sandbox/code/trainer.py361CODE
LOWllm/distributed-rl-eval/src/eval_head.py123CODE
LOWllm/distributed-rl-eval/src/eval_head.py289CODE
LOWsky/task.py2202CODE
LOWsky/task.py556CODE
LOWsky/task.py649CODE
LOWsky/task.py1045CODE
LOWsky/task.py1214CODE
LOWsky/task.py1262CODE
LOWsky/task.py1332CODE
LOWsky/task.py1668CODE
LOWsky/task.py1929CODE
LOWsky/task.py1980CODE
LOWsky/sky_logging.py166CODE
LOWsky/cloud_stores.py216CODE
LOWsky/execution.py112CODE
LOWsky/execution.py202CODE
LOWsky/execution.py343CODE
LOWsky/execution.py688CODE
LOWsky/check.py148CODE
LOWsky/check.py597CODE
LOWsky/check.py683CODE
LOWsky/check.py181CODE
LOWsky/check.py215CODE
LOWsky/global_user_state.py397CODE
LOWsky/global_user_state.py678CODE
LOWsky/global_user_state.py987CODE
LOWsky/global_user_state.py2093CODE
LOWsky/global_user_state.py2655CODE
LOWsky/skypilot_config.py621CODE
LOWsky/skypilot_config.py1072CODE
LOWsky/skypilot_config.py1266CODE
LOWsky/core.py560CODE
LOWsky/core.py1667CODE
LOWsky/authentication.py102CODE
LOWsky/authentication.py249CODE
LOWsky/optimizer.py1548CODE
LOWsky/optimizer.py1617CODE
LOWsky/optimizer.py1671CODE
LOWsky/optimizer.py239CODE
LOWsky/optimizer.py429CODE
LOWsky/optimizer.py738CODE
LOWsky/optimizer.py999CODE
LOWsky/optimizer.py1094CODE
LOWsky/optimizer.py1262CODE
LOWsky/optimizer.py1312CODE
LOWsky/optimizer.py1381CODE
644 more matches not shown…
Over-Commented Block423 hits · 350 pts
SeverityFileLineSnippetContext
LOWrequirements-dev.txt1# formattingCOMMENT
LOWrequirements-dev.txt41# For mocking AWSCOMMENT
LOWrequirements-dev.txt61grpcio-tools==1.63.0COMMENT
LOWformat.sh1#!/usr/bin/env bashCOMMENT
LOWllm/train-eval-jobgroup/train-eval-jobgroup.yaml1# Parallel Training and Evaluation with Shared VolumeCOMMENT
LOWllm/train-eval-jobgroup/train-eval-ckpts-volume.yaml1# Volume definition for train-eval-jobgroup shared checkpoint storageCOMMENT
LOW…/rl-post-training-jobgroup/rlhf-math-jobgroup-cpu.yaml1# RLHF Math Training with Job Groups - CPU Test VersionCOMMENT
LOWllm/rl-post-training-jobgroup/rlhf-math-jobgroup.yaml1# RLHF Math Training with Job GroupsCOMMENT
LOWllm/rl-post-training-jobgroup/rlhf-math-jobgroup.yaml21# cache-aware load balancing across all SGLang instances for optimal KV cache reuse.COMMENT
LOWllm/rl-post-training-jobgroup/rlhf-policy-volume.yaml1# Volume definition for rl-post-training-jobgroup shared policy weights.COMMENT
LOWllm/vicuna-llama-2/scripts/hardcoded_questions.py1# This code is based on lmsys-org/fastchat. Below is the original copyright:COMMENT
LOWllm/vicuna-llama-2/scripts/train_xformers.py1# This code is based on lmsys-org/fastchat. Below is the original copyright:COMMENT
LOWllm/vicuna-llama-2/scripts/xformers_patch.py1# This code is based on lmsys-org/fastchat. Below is the original copyright:COMMENT
LOWllm/vicuna-llama-2/scripts/train.py1# This code is based on tatsu-lab/stanford_alpaca. Below is the original copyright:COMMENT
LOWllm/vicuna-llama-2/scripts/train.py21# Licensed under the Apache License, Version 2.0 (the "License");COMMENT
LOWllm/llama-3_2/llama3_2-vision-11b.yaml1# Serving Meta Llama 3.2 on your own infra.COMMENT
LOWllm/llama-3_2/llama3_2-vision-11b.yaml21# },COMMENT
LOWllm/llama-3_2/llama3_2-vision-11b.yaml41# curl -L $ENDPOINT/v1/modelsCOMMENT
LOWllm/llama-3_2/llama3_2.yaml1# Serving Meta Llama 3.2 on your own infra.COMMENT
LOWllm/llama-3_2/llama3_2.yaml21# },COMMENT
LOWllm/llama-3_2/llama3_2.yaml41# curl -L $ENDPOINT/v1/modelsCOMMENT
LOWllm/llama-3/llama3.yaml1# Serving Meta Llama-3 on your own infra.COMMENT
LOWllm/llama-3/llama3.yaml21# },COMMENT
LOWllm/llama-3/llama3.yaml41# curl -L $ENDPOINT/v1/modelsCOMMENT
LOWllm/llama-3/gui.yaml1# Starts a GUI server that connects to the Llama-3 OpenAI API server.COMMENT
LOWllm/qwen/gui.yaml1# Starts a GUI server that connects to the Qwen OpenAI API server.COMMENT
LOWllm/llama-3_1-finetuning/configs/8B-lora.yaml1# Config for multi-device LoRA finetuning in lora_finetune_distributed.pyCOMMENT
LOWllm/verl/multinode.yaml1# Multi-node distributed training with Verl (Volcano Engine Reinforcement Learning) framework.COMMENT
LOWllm/verl/multinode.yaml161 trainer.resume_mode=auto 2>&1 | tee verl_training.logCOMMENT
LOW…rch-tooling/verl-search-interaction-google-search.yaml1# Search Tool Interaction Inference (Google Search backend)COMMENT
LOW…search-tooling/verl-search-interaction-rl-trainer.yaml1# Search Tool Interaction Training with VERL (RL Trainer)COMMENT
LOWllm/verl/search-tooling/verl-search-interaction.yaml1# Search Tool Interaction Training with VERLCOMMENT
LOWllm/verl/search-tooling/verl-search-interaction.yaml141 # conda initCOMMENT
LOW…verl/search-tooling/verl-search-interaction-infer.yaml1# Search Tool Interaction InferenceCOMMENT
LOW…/search-tooling/verl-search-interaction-retrieval.yaml1# Search Tool Retrieval ServiceCOMMENT
LOWllm/dbrx/dbrx.yaml1# Serving Databricks DBRX on your own infra.COMMENT
LOWllm/dbrx/dbrx.yaml21# "role": "user",COMMENT
LOWllm/rl-code-execution-sandbox/rl-code-jobgroup.yaml1# RL Code-Execution Training with Job Groups (sandbox-backed reward)COMMENT
LOWllm/rl-code-execution-sandbox/rl-code-jobgroup.yaml21#COMMENT
LOWllm/rl-code-execution-sandbox/rl-code-jobgroup-cpu.yaml1# RL Code-Execution Training with Job Groups - CPU Test VersionCOMMENT
LOWllm/rl-code-execution-sandbox/rl-code-volume.yaml1# Volume definition for rl-code-execution-sandbox shared policy weights.COMMENT
LOWllm/llama-4-finetuning/llama-4-maverick.yaml21# /checkpoints directory to write checkpoints, which writes to local disk firstCOMMENT
LOWllm/llama-4-finetuning/llama-4-scout-sft.yaml21# /checkpoints directory to write checkpoints, which writes to local disk firstCOMMENT
LOWllm/llama-4-finetuning/llama-4-maverick-lora.yaml21COMMENT
LOWllm/llama-4-finetuning/llama-4-maverick-sft.yaml21# /checkpoints directory to write checkpoints, which writes to local disk firstCOMMENT
LOWllm/llama-4-finetuning/configs/llama4_lora_sft.yaml41ddp_timeout: 180000000COMMENT
LOW…4-finetuning/configs/llama4_maverick_full_sft_cpu.yaml41bf16: trueCOMMENT
LOWllm/ollama/ollama.yaml1# Run LLMs on CPUs with OllamaCOMMENT
LOWllm/fairseq2/sft.sky.yaml1# Fairseq2: Instruction Fine-tuning LLMs on SkyPilotCOMMENT
LOWllm/fairseq2/multinode.sky.yaml1# Fairseq2: Multi-node Fine-tuning on SkyPilotCOMMENT
LOWllm/falcon/train.py1# Adapted from https://gist.github.com/pacman100/1731b41f7a90a87b457e8c5415ff1c14COMMENT
LOWllm/llama-3_1/llama-3_1.yaml1# Serving Meta Llama-3.1 on your own infra.COMMENT
LOWllm/llama-3_1/llama-3_1.yaml21# "role": "system",COMMENT
LOWllm/llama-3_1/llama-3_1.yaml41# ENDPOINT=$(sky serve status --endpoint llama31)COMMENT
LOWsky/task.py721 # Unlike envs, a null-valued secret in dict form (e.g. ``API_KEY:``COMMENT
LOWsky/task.py921COMMENT
LOWsky/task.py1621 return self.set_storage_mounts(task_storage_mounts)COMMENT
LOWsky/cloud_stores.py181class AzureBlobCloudStorage(CloudStorage):COMMENT
LOWsky/execution.py401 if controller_utils.high_availability_specified(cluster_name):COMMENT
LOWsky/execution.py841 # concurrent calls from queueing up many sequential reprovisionCOMMENT
363 more matches not shown…
Structural Annotation Overuse148 hits · 243 pts
SeverityFileLineSnippetContext
LOWllm/gpt-oss/README.md15## Step 0: Setup infrastructureCOMMENT
LOWllm/gpt-oss/README.md32## Step 1: Run gpt-oss modelsCOMMENT
LOWllm/gpt-oss/README.md48## Step 2: Get resultsCOMMENT
LOWllm/gpt-oss/README.md103## Step 3: Scale with SkyServeCOMMENT
LOWllm/gemma3/README.md22## Step 0: SetupCOMMENT
LOWllm/gemma3/README.md41## Step 1: Run it with SkyPilotCOMMENT
LOWllm/gemma3/README.md53## Step 2: Get ResultsCOMMENT
LOWllm/rag/README.md27## Step 0: Set Up The EnvironmentCOMMENT
LOWllm/rag/README.md39## Step 1: Compute Embeddings from Legal DocumentsCOMMENT
LOWllm/rag/README.md109## Step 2: Build RAG with Vector DatabaseCOMMENT
LOWllm/rag/README.md123## Step 3: Serve the RAGCOMMENT
LOWllm/rag/scripts/templates/index.html296 // Step 1: Get documents firstCOMMENT
LOWllm/rag/scripts/templates/index.html340 // Step 2: Start the LLM reasoning process in the backgroundCOMMENT
LOWllm/gpt-oss-finetuning/README.md15## Step 0: Setup infrastructureCOMMENT
LOWllm/gpt-oss-finetuning/README.md44## Step 1: Run gpt-oss modelsCOMMENT
LOWllm/gpt-oss-finetuning/README.md122## Step 2: Monitor and get resultsCOMMENT
LOWllm/falcon_h1/Readme.md20## Step 0: Install SkyPilotCOMMENT
LOWllm/falcon_h1/Readme.md34## Step 1: Run it with SkyPilotCOMMENT
LOWllm/falcon_h1/Readme.md62## Step 2: Get ResultsCOMMENT
LOWllm/deepseek-r1-distilled/README.md19## Step 0: Bring any infraCOMMENT
LOWllm/deepseek-r1-distilled/README.md43## Step 1: Run it with SkyPilotCOMMENT
LOWllm/deepseek-r1-distilled/README.md72## Step 2: Get ResultsCOMMENT
LOWllm/deepseek-janus/README.md13## Step 0: Bring any infraCOMMENT
LOWllm/deepseek-janus/README.md29## Step 1: Run it with SkyPilotCOMMENT
LOWllm/deepseek-janus/README.md74## Step 2: Access the deployed serverCOMMENT
LOWllm/llama-3_1/README.md38### Step 0: Bring your infraCOMMENT
LOWllm/llama-3_1/README.md69### Step 1: Get a GPU dev node (pod or VM)COMMENT
LOWllm/llama-3_1/README.md91### Step 2: Inside the dev node, test servingCOMMENT
LOWllm/llama-3_1/README.md152### Step 3: Package and deploy using SkyPilotCOMMENT
LOWsky/optimizer.py1105 # Step 1: Get feasible resources for each taskCOMMENT
LOWsky/optimizer.py1130 # Step 2: Find common cloud+region combinationsCOMMENT
LOWsky/optimizer.py1141 # Step 3: Select best infra based on minimize targetCOMMENT
LOWsky/optimizer.py1163 # Step 4: Assign resources for each task on the selected infraCOMMENT
LOWsky/optimizer.py1196 # Step 5: Print optimizer table for job groupsCOMMENT
LOWsky/provision/nebius/utils.py383 # Step 1: kick off async deletion of all rules.COMMENT
LOWsky/provision/nebius/utils.py400 # Step 2: delete the SG with retry-on-dependency.COMMENT
LOWsky/catalog/data_fetchers/fetch_seeweb.py71 # Step 1: Try to get from config fileCOMMENT
LOWsky/catalog/data_fetchers/fetch_seeweb.py82 # Step 2: Try environment variableCOMMENT
LOWsky/ssh_node_pools/deploy/deploy.py478 # Step 1: Install k3s on the head nodeCOMMENT
LOWsky/ssh_node_pools/deploy/deploy.py654 # Step 3: Configure local kubectl to connect to the clusterSTRING
LOWsky/ssh_node_pools/deploy/deploy.py965 # Step 1: Get the selector labels from the dcgm-exporter DaemonSetSTRING
LOWsky/ssh_node_pools/deploy/deploy.py602 # Step 2: Install k3s on worker nodes and join them to the master nodeSTRING
LOWsky/ssh_node_pools/deploy/deploy.py992 # Step 2: Create the Service with the discovered selectorSTRING
LOWsky/utils/git.py264 # Step 2: Try with token if providedCOMMENT
LOWsky/utils/git.py283 # Step 3: Try SSH access with available keysCOMMENT
LOWsky/utils/git.py385 # Step 1: Check provided SSH key path firstCOMMENT
LOWsky/utils/git.py416 # Step 2: Check SSH config for host-specific configurationCOMMENT
LOWsky/utils/git.py427 # Step 3: Search for default SSH keysCOMMENT
LOWsky/utils/git.py237 # Step 1: Try public access first (most common case)COMMENT
LOWsky/utils/command_runner.py882 # Step 1: Transfer the script to remote machine using rsyncCOMMENT
LOWsky/utils/command_runner.py893 # Step 2: Execute the script on remote machineCOMMENT
LOWsky/utils/controller_utils.py1102 # Step 1: Translate the workdir to SkyPilot storage.COMMENT
LOWsky/utils/controller_utils.py1136 # Step 2: Translate the local file mounts with folder in src to SkyPilotCOMMENT
LOWsky/utils/controller_utils.py1164 # Step 3: Translate local file mounts with file in src to SkyPilot storage.COMMENT
LOWsky/utils/controller_utils.py1212 # Step 4: Upload storage from sourcesCOMMENT
LOWsky/utils/controller_utils.py1254 # Step 5: Add the file download into the file mounts, such asCOMMENT
LOWsky/utils/controller_utils.py1274 # Step 6: Replace the source field that is local path in all storage_mountsCOMMENT
LOWsky/utils/controller_utils.py1293 # Step 7: Convert all `MOUNT` mode storages which don't specify a sourceCOMMENT
LOWsky/dashboard/src/components/users.jsx1485 // Step 1: Load users first and show them immediatelyCOMMENT
LOWsky/dashboard/src/components/users.jsx1505 // Step 2: Load clusters and jobs in background and update countsCOMMENT
88 more matches not shown…
Example Usage Blocks131 hits · 194 pts
SeverityFileLineSnippetContext
LOWformat.sh4# Usage:COMMENT
LOWllm/openrlhf/openrlhf_sft.yaml2# Usage:COMMENT
LOWllm/openrlhf/openrlhf_rm_training.yaml2# Usage:COMMENT
LOWllm/openrlhf/openrlhf_dpo.yaml2# Usage:COMMENT
LOWllm/gemma/serve.yaml2# Usage:COMMENT
LOWllm/localgpt/localgpt.yaml2# Usage:COMMENT
LOWllm/train-eval-jobgroup/train-eval-jobgroup.yaml17# Usage:COMMENT
LOWllm/mixtral/serve.yaml2# Usage:COMMENT
LOW…/rl-post-training-jobgroup/rlhf-math-jobgroup-cpu.yaml10# Usage:COMMENT
LOWllm/rl-post-training-jobgroup/rlhf-math-jobgroup.yaml24# Usage:COMMENT
LOWllm/codellama/gui.yaml4# Usage:COMMENT
LOWllm/codellama/endpoint.yaml2# Usage:COMMENT
LOWllm/codellama/tabby.yaml4# Usage:COMMENT
LOWllm/kimi-k2-thinking/kimi-k2-thinking.sky.yaml4# Usage:COMMENT
LOW…-k2-thinking/kimi-k2-thinking-high-throughput.sky.yaml4# Usage:COMMENT
LOWllm/llama-3_2/llama3_2-vision-11b.yaml3# Usage:COMMENT
LOWllm/llama-3_2/llama3_2.yaml3# Usage:COMMENT
LOWllm/llama-3/llama3.yaml3# Usage:COMMENT
LOWllm/llama-3/gui.yaml6# Usage:COMMENT
LOWllm/qwen/gui.yaml5# Usage:COMMENT
LOWllm/llama-3_1-finetuning/serve.yaml3# Usage:COMMENT
LOWllm/llama-3_1-finetuning/lora.yaml3# Usage:COMMENT
LOWllm/tgi/serve.yaml3# Usage:COMMENT
LOWllm/verl/verl-grpo.yaml1# Usage:COMMENT
LOWllm/verl/multinode.yaml11# Usage:COMMENT
LOWllm/verl/verl-ppo.yaml1# Usage:COMMENT
LOW…rch-tooling/verl-search-interaction-google-search.yaml7# Usage:COMMENT
LOW…search-tooling/verl-search-interaction-rl-trainer.yaml10# Usage:COMMENT
LOWllm/verl/search-tooling/verl-search-interaction.yaml8# Usage:COMMENT
LOW…verl/search-tooling/verl-search-interaction-infer.yaml7# Usage:COMMENT
LOW…/search-tooling/verl-search-interaction-retrieval.yaml6# Usage:COMMENT
LOWllm/dbrx/dbrx.yaml3# Usage:COMMENT
LOWllm/rl-code-execution-sandbox/rl-code-jobgroup.yaml28# Usage:COMMENT
LOWllm/rl-code-execution-sandbox/rl-code-jobgroup-cpu.yaml17# Usage:COMMENT
LOWllm/llama-4-finetuning/llama-4-maverick.yaml3# Usage:COMMENT
LOWllm/llama-4-finetuning/llama-4-scout-sft.yaml3# Usage:COMMENT
LOWllm/llama-4-finetuning/llama-4-maverick-lora.yaml3# Usage:COMMENT
LOWllm/llama-4-finetuning/llama-4-maverick-sft.yaml3# Usage:COMMENT
LOWllm/nemorl/nemorl.sky.yaml5# Usage:COMMENT
LOWllm/ollama/ollama.yaml3# Usage:COMMENT
LOWllm/fairseq2/sft.sky.yaml6# Usage:COMMENT
LOWllm/fairseq2/multinode.sky.yaml6# Usage:COMMENT
LOWllm/kimi-k2/kimi-k2.sky.yaml4# Usage:COMMENT
LOWllm/axolotl/axolotl.yaml1# Usage:COMMENT
LOWllm/axolotl/axolotl-spot.yaml1# Usage:COMMENT
LOWllm/axolotl/axolotl-docker.yaml1# Usage:COMMENT
LOWllm/llama-3_1/llama-3_1.yaml3# Usage:COMMENT
LOWsky/skylet/constants.py19# Example usage:COMMENT
LOWsky/utils/kubernetes/generate_kubeconfig.sh17# Usage:COMMENT
LOWtests/test_smoke.py3# Example usage:COMMENT
LOWtests/test_yamls/test_serve_autoscaler.yaml3# Usage:COMMENT
LOWtests/smoke_tests/test_basic.py3# Example usage:COMMENT
LOWtests/smoke_tests/test_cluster_job.py3# Example usage:COMMENT
LOWtests/smoke_tests/test_managed_job.py3# Example usage:COMMENT
LOWtests/smoke_tests/test_images.py3# Example usage:COMMENT
LOWtests/smoke_tests/test_sky_serve.py3# Example usage:COMMENT
LOWtests/smoke_tests/test_mount_and_storage.py3# Example usage:COMMENT
LOWtests/smoke_tests/smoke_tests_utils.py972# Example usage:COMMENT
LOWtests/smoke_tests/test_quick_tests_core.py7# Example usage:COMMENT
LOWtests/smoke_tests/test_region_and_zone.py3# Example usage:COMMENT
71 more matches not shown…
Verbosity Indicators87 hits · 146 pts
SeverityFileLineSnippetContext
LOWllm/rag/scripts/templates/index.html296 // Step 1: Get documents firstCOMMENT
LOWllm/rag/scripts/templates/index.html340 // Step 2: Start the LLM reasoning process in the backgroundCOMMENT
LOWsky/cloud_stores.py240 # If there are more, we need to check if it is a directory or a file.COMMENT
LOWsky/optimizer.py1105 # Step 1: Get feasible resources for each taskCOMMENT
LOWsky/optimizer.py1130 # Step 2: Find common cloud+region combinationsCOMMENT
LOWsky/optimizer.py1141 # Step 3: Select best infra based on minimize targetCOMMENT
LOWsky/optimizer.py1163 # Step 4: Assign resources for each task on the selected infraCOMMENT
LOWsky/optimizer.py1196 # Step 5: Print optimizer table for job groupsCOMMENT
LOWsky/provision/nebius/utils.py383 # Step 1: kick off async deletion of all rules.COMMENT
LOWsky/provision/nebius/utils.py400 # Step 2: delete the SG with retry-on-dependency.COMMENT
LOWsky/catalog/data_fetchers/fetch_seeweb.py71 # Step 1: Try to get from config fileCOMMENT
LOWsky/catalog/data_fetchers/fetch_seeweb.py82 # Step 2: Try environment variableCOMMENT
LOWsky/ssh_node_pools/deploy/deploy.py478 # Step 1: Install k3s on the head nodeCOMMENT
LOWsky/ssh_node_pools/deploy/deploy.py654 # Step 3: Configure local kubectl to connect to the clusterSTRING
LOWsky/ssh_node_pools/deploy/deploy.py965 # Step 1: Get the selector labels from the dcgm-exporter DaemonSetSTRING
LOWsky/ssh_node_pools/deploy/deploy.py602 # Step 2: Install k3s on worker nodes and join them to the master nodeSTRING
LOWsky/ssh_node_pools/deploy/deploy.py992 # Step 2: Create the Service with the discovered selectorSTRING
LOWsky/utils/git.py264 # Step 2: Try with token if providedCOMMENT
LOWsky/utils/git.py283 # Step 3: Try SSH access with available keysCOMMENT
LOWsky/utils/git.py385 # Step 1: Check provided SSH key path firstCOMMENT
LOWsky/utils/git.py416 # Step 2: Check SSH config for host-specific configurationCOMMENT
LOWsky/utils/git.py427 # Step 3: Search for default SSH keysCOMMENT
LOWsky/utils/git.py237 # Step 1: Try public access first (most common case)COMMENT
LOWsky/utils/command_runner.py882 # Step 1: Transfer the script to remote machine using rsyncCOMMENT
LOWsky/utils/command_runner.py893 # Step 2: Execute the script on remote machineCOMMENT
LOWsky/utils/controller_utils.py1102 # Step 1: Translate the workdir to SkyPilot storage.COMMENT
LOWsky/utils/controller_utils.py1136 # Step 2: Translate the local file mounts with folder in src to SkyPilotCOMMENT
LOWsky/utils/controller_utils.py1164 # Step 3: Translate local file mounts with file in src to SkyPilot storage.COMMENT
LOWsky/utils/controller_utils.py1212 # Step 4: Upload storage from sourcesCOMMENT
LOWsky/utils/controller_utils.py1254 # Step 5: Add the file download into the file mounts, such asCOMMENT
LOWsky/utils/controller_utils.py1274 # Step 6: Replace the source field that is local path in all storage_mountsCOMMENT
LOWsky/utils/controller_utils.py1293 # Step 7: Convert all `MOUNT` mode storages which don't specify a sourceCOMMENT
LOWsky/dashboard/src/components/users.jsx1485 // Step 1: Load users first and show them immediatelyCOMMENT
LOWsky/dashboard/src/components/users.jsx1505 // Step 2: Load clusters and jobs in background and update countsCOMMENT
LOWsky/dashboard/src/components/users.jsx2874 // Step 1: Fetch service account tokens (using cache)COMMENT
LOWsky/dashboard/src/components/users.jsx2878 // Step 2: Fetch clusters and jobs data in parallelCOMMENT
LOWsky/dashboard/src/components/users.jsx2882 // Step 3: Calculate counts for each service accountCOMMENT
LOWsky/dashboard/src/components/infra.jsx1439 // Step 1: Call sshDownNodePool to get request_id for streamingCOMMENT
LOWsky/dashboard/src/components/infra.jsx1460 // Step 2: After streaming completes, call the parent's delete handlerCOMMENT
LOWsky/dashboard/src/data/connectors/clusters.jsx338 // Step 1: schedule server-side download; result is a mapping job_id -> folder path on API serverCOMMENT
LOWsky/dashboard/src/data/connectors/clusters.jsx353 // Step 2: request the zip and trigger browser downloadCOMMENT
LOWsky/dashboard/src/data/connectors/jobs.jsx865 // Step 1: dispatch the request and grab its server-side ID.COMMENT
LOWsky/dashboard/src/data/connectors/jobs.jsx906 // Step 2: long-poll /api/get, retrying on edge-timeout responses.COMMENT
LOWsky/dashboard/src/data/connectors/infra.jsx234 // Step 1: Get all accessible workspaces for the user (use cache for performance)COMMENT
LOWsky/dashboard/src/data/connectors/infra.jsx264 // Step 2: Fetch expanded clouds for all workspaces in a single batch callCOMMENT
LOWsky/dashboard/src/data/connectors/infra.jsx313 // Step 3: Get detailed GPU information for all contextsCOMMENT
LOWsky/dashboard/src/data/connectors/infra.jsx377 // Step 1: Get all accessible workspaces for the user (use cache for performance)COMMENT
LOWsky/dashboard/src/data/connectors/infra.jsx389 // Step 2: Fetch expanded clouds for all workspaces in a single batch callCOMMENT
LOWsky/dashboard/src/data/connectors/workspaces.jsx5 // Step 1: Call the /workspaces endpoint to schedule the taskCOMMENT
LOWsky/dashboard/src/data/connectors/workspaces.jsx13 // Step 2: Get the request_id from the response headers or bodyCOMMENT
LOWsky/dashboard/src/data/connectors/workspaces.jsx59 // Step 3: Poll the /api/get endpoint with the request_id.COMMENT
LOWsky/dashboard/src/data/connectors/workspaces.jsx97 // Step 4: Check task status and return the actual workspace dataCOMMENT
LOWsky/dashboard/src/data/connectors/workspaces.jsx195 // Step 1: Call the /enabled_clouds endpoint to schedule the taskCOMMENT
LOWsky/dashboard/src/data/connectors/workspaces.jsx216 // Step 2: Get the request_id from the response headers or bodyCOMMENT
LOWsky/dashboard/src/data/connectors/workspaces.jsx256 // Step 3: Poll the /api/get endpoint with the request_id.COMMENT
LOWtests/smoke_tests/test_mount_and_storage.py606 # For Azure, we need to check if the output is empty list, as it returns []COMMENT
LOWtests/unit_tests/test_sky/test_cli_infra_override.py164 # Step 1: Parse infra string through CLI handlerCOMMENT
LOWtests/unit_tests/test_sky/test_cli_infra_override.py168 # Step 2: Process through override params parserCOMMENT
LOWtests/unit_tests/test_sky/test_cli_infra_override.py244 # Step 1: CLI processing should convert gcp/* to proper parametersCOMMENT
LOWtests/unit_tests/test_sky/test_cli_infra_override.py253 # Step 2: Override params should convert '*' to None for proper overrideCOMMENT
27 more matches not shown…
Modern Structural Boilerplate125 hits · 121 pts
SeverityFileLineSnippetContext
LOWllm/rag/scripts/serve_rag.py25logger = logging.getLogger(__name__)CODE
LOWllm/rag/scripts/compute_embeddings.py23logger = logging.getLogger(__name__)CODE
LOWllm/rag/scripts/build_rag.py24logger = logging.getLogger(__name__)CODE
LOWsky/global_user_state.py1758def set_cluster_info(cluster_name: str, metadata: Dict[str, Any]) -> None:CODE
LOWsky/global_user_state.py2740def set_allowed_clouds(allowed_clouds: List[str], workspace: str) -> None:CODE
LOWsky/global_user_state.py3025def update_volume_config(name: str, config: models.VolumeConfig) -> None:CODE
LOWsky/global_user_state.py3223def update_service_account_token_last_used(token_id: str) -> None:CODE
LOWsky/global_user_state.py3380def set_cluster_yaml(cluster_name: str, yaml_str: str) -> None:CODE
LOWsky/global_user_state.py3474def set_system_config(config_key: str, config_value: str) -> None:CODE
LOWsky/skypilot_config.py254def _set_loaded_config_path_serialized(path: Optional[str]) -> None:CODE
LOWsky/skypilot_config.py262def _set_config_overridden(config_overridden: bool) -> None:CODE
LOWsky/skypilot_config.py1187 def update_to_none_if_set(keys: Tuple[str, ...]) -> None:CODE
LOWsky/skypilot_config.py231def _set_loaded_config(config: config_utils.Config) -> None:CODE
LOWsky/skypilot_config.py1266def update_api_server_config_no_lock(config: config_utils.Config) -> None:CODE
LOWsky/dag.py99 def set_execution(self, execution: 'DagExecution') -> None:CODE
LOWsky/__init__.py53def _set_http_proxy_env_vars() -> None:CODE
LOWsky/__init__.py177__all__ = [CODE
LOWsky/resources.py832 def set_requires_fuse(self, value: bool) -> None:CODE
LOWsky/resources.py1093 def _set_priority(self, priority: Optional[int]) -> None:CODE
LOWsky/resources.py1108 def _set_priority_class(self, priority_class: Optional[str]) -> None:CODE
LOWsky/resources.py1201 def _set_local_disk(self, local_disk: Optional[str]) -> None:CODE
LOWsky/provision/slurm/config.py6logger = logging.getLogger(__name__)CODE
LOWsky/provision/paperspace/utils.py126 def set_sky_key_script(self, public_key: str) -> None:CODE
LOWsky/provision/gcp/config.py18logger = logging.getLogger(__name__)CODE
LOWsky/provision/scp/instance.py17logger = logging.getLogger(__name__)CODE
LOWsky/provision/kubernetes/config.py13logger = logging.getLogger(__name__)CODE
LOWsky/skylet/configs.py62def set_config(key: str, value: Union[bytes, str]) -> None:CODE
LOWsky/skylet/job_lib.py416def set_log_dir_no_lock(job_id: int, log_dir: str) -> None:CODE
LOWsky/skylet/job_lib.py448def update_job_metadata(job_id: int, updates: Dict[str, Any]) -> None:CODE
LOWsky/skylet/job_lib.py484def _set_status_no_lock(job_id: int, status: JobStatus) -> None:CODE
LOWsky/skylet/job_lib.py509def set_status(job_id: int, status: JobStatus) -> None:CODE
LOWsky/skylet/job_lib.py517def set_exit_codes(job_id: int, exit_codes: List[int]) -> None:CODE
LOWsky/skylet/job_lib.py552def set_job_started(job_id: int) -> None:CODE
LOWsky/skylet/job_lib.py998def update_status() -> None:STRING
LOWsky/skylet/autostop_lib.py232def set_autostopping_started() -> None:CODE
LOWsky/skylet/autostop_lib.py334def set_hooks(hooks: Optional[List[Dict[str, Any]]]) -> None:CODE
LOWsky/skylet/autostop_lib.py367def set_last_active_time_to_now() -> None:CODE
LOWsky/skylet/providers/ibm/node_provider.py455 def set_node_tags(self, node_id, tags) -> None:CODE
LOWsky/backends/__init__.py11__all__ = [CODE
LOWsky/backends/cloud_vm_ray_backend.py2058 def update_ssh_ports(self, max_attempts: int = 1) -> None:CODE
LOWsky/backends/cloud_vm_ray_backend.py2382 def _set_skylet_ssh_tunnel(self, tunnel: Optional[SSHTunnelInfo]) -> None:CODE
LOWsky/backends/cloud_vm_ray_backend.py3960 def _setup_node(node_id: int) -> None:CODE
LOWsky/backends/playground/demo_dockerutils.py8logger = logging.getLogger(__name__)CODE
LOWsky/catalog/__init__.py386__all__ = [CODE
LOWsky/catalog/oci_catalog.py30logger = logging.getLogger(__name__)CODE
LOWsky/catalog/data_fetchers/fetch_vsphere.py18logger = logging.getLogger(__name__)CODE
LOWsky/catalog/data_fetchers/fetch_nebius.py20logger = logging.getLogger(__name__)CODE
LOWsky/catalog/data_fetchers/fetch_mithril.py20logger = logging.getLogger(__name__)CODE
LOWsky/ssh_node_pools/core.py46 def update_pools(self, pools_config: Dict[str, Any]) -> None:CODE
LOWsky/ssh_node_pools/core.py119def update_pools(pools_config: Dict[str, Any]) -> None:CODE
LOWsky/ssh_node_pools/deploy/__init__.py4__all__ = ['run']CODE
LOWsky/server/versions.py60def set_remote_api_version(api_version: int) -> None:CODE
LOWsky/server/versions.py68def set_remote_version(version: str) -> None:CODE
LOWsky/server/clean_env.py36def set_clean_server_env(env: Dict[str, str]) -> None:CODE
LOWsky/server/state.py17def set_block_requests(shutting_down: bool) -> None:CODE
LOWsky/server/requests/log_provider.py127def set_log_provider(lp: LogProvider) -> None:CODE
LOWsky/server/requests/storage.py213def set_request_backend(backend: RequestBackend) -> None:CODE
LOWsky/server/requests/process.py13logger = logging.getLogger(__name__)CODE
LOWsky/server/requests/requests.py1128def set_exception_stacktrace(e: BaseException) -> None:CODE
LOWsky/server/requests/requests.py1134def set_request_failed(request_id: str, e: BaseException) -> None:CODE
65 more matches not shown…
AI Structural Patterns134 hits · 112 pts
SeverityFileLineSnippetContext
LOWllm/verl/code/preprocess_rstar_coder.py62CODE
LOWllm/batch_inference/scripts/text_vector_processor.py30CODE
LOWllm/distributed-rl-eval/src/eval_head.py688CODE
LOWsky/task.py321CODE
LOWsky/execution.py202CODE
LOWsky/execution.py688CODE
LOWsky/resources.py161CODE
LOWsky/provision/mithril/utils.py715CODE
LOWsky/provision/paperspace/utils.py218CODE
LOWsky/provision/gcp/config.py1000CODE
LOWsky/provision/gcp/config.py1016CODE
LOWsky/provision/gcp/config.py1084CODE
LOWsky/provision/gcp/config.py1129CODE
LOWsky/provision/gcp/config.py1145CODE
LOWsky/provision/gcp/config.py1164CODE
LOWsky/provision/runpod/api/commands.py28CODE
LOWsky/provision/runpod/api/pods.py23CODE
LOWsky/provision/primeintellect/utils.py290CODE
LOWsky/skylet/log_lib.py158CODE
LOWsky/backends/cloud_vm_ray_backend.py5938CODE
LOWsky/catalog/verda_catalog.py47CODE
LOWsky/catalog/yotta_catalog.py45CODE
LOWsky/catalog/vsphere_catalog.py71CODE
LOWsky/catalog/vast_catalog.py63CODE
LOWsky/catalog/vast_catalog.py87CODE
LOWsky/catalog/scp_catalog.py54CODE
LOWsky/catalog/lambda_catalog.py59CODE
LOWsky/catalog/nebius_catalog.py54CODE
LOWsky/catalog/azure_catalog.py117CODE
LOWsky/catalog/seeweb_catalog.py37CODE
LOWsky/catalog/seeweb_catalog.py48CODE
LOWsky/catalog/seeweb_catalog.py62CODE
LOWsky/catalog/seeweb_catalog.py69CODE
LOWsky/catalog/seeweb_catalog.py85CODE
LOWsky/catalog/seeweb_catalog.py72CODE
LOWsky/catalog/seeweb_catalog.py117CODE
LOWsky/catalog/seeweb_catalog.py148CODE
LOWsky/catalog/seeweb_catalog.py153CODE
LOWsky/catalog/seeweb_catalog.py188CODE
LOWsky/catalog/seeweb_catalog.py209CODE
LOWsky/catalog/paperspace_catalog.py51CODE
LOWsky/catalog/aws_catalog.py255CODE
LOWsky/catalog/__init__.py57CODE
LOWsky/catalog/__init__.py217CODE
LOWsky/catalog/__init__.py266CODE
LOWsky/catalog/gcp_catalog.py282CODE
LOWsky/catalog/slurm_catalog.py28CODE
LOWsky/catalog/runpod_catalog.py46CODE
LOWsky/catalog/oci_catalog.py104CODE
LOWsky/catalog/ibm_catalog.py100CODE
LOWsky/catalog/mithril_catalog.py71CODE
LOWsky/catalog/hyperbolic_catalog.py68CODE
LOWsky/catalog/shadeform_catalog.py98CODE
LOWsky/catalog/fluidstack_catalog.py53CODE
LOWsky/catalog/do_catalog.py51CODE
LOWsky/catalog/primeintellect_catalog.py42CODE
LOWsky/catalog/cudo_catalog.py52CODE
LOWsky/usage/usage_lib.py679CODE
LOWsky/utils/command_runner.py201CODE
LOWsky/utils/command_runner.py606CODE
74 more matches not shown…
AI Slop Vocabulary44 hits · 89 pts
SeverityFileLineSnippetContext
MEDIUMllm/falcon/train.py35# by using the RewardTrainer from trl, we will leverage PEFT library to finetuneCOMMENT
MEDIUMsky/task.py656 # More robust handling for 'envs': explicitly convert keys and values toCOMMENT
MEDIUMsky/task.py669 # More robust handling for 'secrets': explicitly convert keys and valuesCOMMENT
MEDIUMsky/cloud_stores.py188 # does not leverage AzCopy's efficient multi-threaded capabilities, leadingCOMMENT
MEDIUMsky/provision/gcp/constants.py417 # robust.STRING
LOWsky/provision/kubernetes/utils.py5101 # If we have already converted the same kubeconfig before, just return.COMMENT
MEDIUMsky/provision/kubernetes/debug.py469 # error). Per-item isolation would be more robust but noisier; revisit if aCOMMENT
MEDIUMsky/skylet/constants.py70# Python 3.11 for a more robust setup.COMMENT
MEDIUMsky/catalog/data_fetchers/fetch_gcp.py179# TODO(woosuk): Make this more robust.STRING
LOWsky/ssh_node_pools/deploy/deploy.py711 # Initially just set to the effective master IPSTRING
MEDIUMsky/utils/message_utils.py12 """Encode a payload to make it more robust for parsing.STRING
MEDIUMsky/utils/git.py375 """Get SSH key path and content using comprehensive strategy.STRING
MEDIUMsky/utils/git.py218 """Validate the repository access with comprehensive authenticationSTRING
MEDIUMsky/utils/resources_utils.py464 # to the zones in the same region in order to leverage the region-levelCOMMENT
LOWsky/utils/controller_utils.py506 # use a broad except and just return None.COMMENT
LOWsky/utils/db/db_utils.py601 # so we should not put it in the cache. Instead, just return.COMMENT
MEDIUMsky/dashboard/src/components/infra.jsx3186 // TODO: Add back when SSH Node Pool add operation is more robustCOMMENT
MEDIUMsky/adaptors/slurm.py481 # TODO(kevin): Use json output for more robust parsing.COMMENT
LOWsky/recipes/db.py412 # No updates requested, just return current stateCOMMENT
MEDIUMsky/serve/replica_managers.py844 # It should be robust enough for `execution.launch` to handle casesCOMMENT
MEDIUMsky/jobs/job_group_networking.py683 # TODO(zhwu): The current handling is not robust against the case whereSTRING
LOWsky/jobs/utils.py524 # handle.cluster_yaml is None, we will just return an empty dictCOMMENT
LOWsky/jobs/utils.py2214 # just return.COMMENT
LOWsky/jobs/utils.py2286 # aren't following, just return.COMMENT
MEDIUMsky/data/mounting_utils.py35# Creates a fusermount3 soft link on older (<22) Ubuntu systems to utilizeCOMMENT
MEDIUMsky/client/sdk.py109# The overloads are not comprehensive - e.g. get_result Literal[False] could beCOMMENT
MEDIUMtests/smoke_tests/test_managed_job.py805 # 2. A short prefix is robust against provisioning backends thatCOMMENT
MEDIUMtests/smoke_tests/smoke_tests_utils.py299# than a fixed column index. This keeps the wait robust regardless of whichCOMMENT
MEDIUMtests/smoke_tests/test_region_and_zone.py274 # TODO(zpoint): this is a temporary fix. We should make it more robust.STRING
MEDIUMtests/smoke_tests/test_lifecycle_hooks.py902 # is robust to that variation. ``kubectl exec`` against a missingCOMMENT
MEDIUMtests/unit_tests/test_serve_utils.py344 # by message content so the assertion stays robust to otherCOMMENT
MEDIUM…nit_tests/test_sky/skylet/test_managed_jobs_service.py78 """Seed the database with test jobs in various states for comprehensive testing."""STRING
MEDIUM…t_tests/test_sky/backends/test_cloud_vm_ray_backend.py264 # Create a comprehensive taskCOMMENT
LOW…s/unit_tests/test_sky/server/requests/test_requests.py558 # Don't loop - just return after one iterationCOMMENT
LOW…tils/kubernetes/test_skypilot_config_configmap_sync.py75 # Should not raise any exceptions, just return earlyCOMMENT
LOWtests/unit_tests/test_sky/volumes/test_k8s_volume.py2038 # Should not raise, just return empty dictCOMMENT
MEDIUMtests/unit_tests/test_sky/jobs/test_jobs_state.py50 """Seed the database with test jobs in various states for comprehensive testing."""STRING
MEDIUMtests/unit_tests/kubernetes/test_kubernetes_utils.py995 """Test pod with comprehensive Kubernetes features."""STRING
MEDIUMtests/load_tests/db_scale_tests/run_scale_test.py199 """Run the comprehensive scale tests manually."""STRING
LOW…/load_tests/db_scale_tests/inject_postgres_clusters.py73 # If unpickling fails, just return the original blobCOMMENT
MEDIUMtests/load_tests/db_scale_tests/scale_test_utils.py420 # callers (including the bash test harness) previously maskedCOMMENT
MEDIUMcharts/skypilot/values.yaml503 # For GKE, use backend service-based external passthrough Network Load Balancer as per best practices.COMMENT
MEDIUM…training_network_storage_benchmarks/e2e/e2e_network.py108 """Log comprehensive training time statistics."""STRING
MEDIUM.github/workflows/release-build.yml140 # Use Python for robust version comparisonCOMMENT
Slop Phrases32 hits · 65 pts
SeverityFileLineSnippetContext
MEDIUMllm/codellama/tabby.yaml9# After the Tabby server is started, you can add the endpoint (URL:port) to VSCodeCOMMENT
MEDIUMllm/vicuna/dummy.json45 "value": "You can call me Vicuna, and I was trained by Large Model Systems Organization (LMSYS) researchers as aCODE
MEDIUMllm/vicuna/dummy.json58 "value": "As a language model, I go by the name Vicuna and was trained by researchers from Large Model Systems OCODE
MEDIUMllm/vicuna/dummy.json110 "value": "You can call me Vicuna, and I was trained by Large Model Systems Organization (LMSYS) researchers as aCODE
MEDIUMllm/vicuna/dummy.json123 "value": "As a language model, I go by the name Vicuna and was trained by researchers from Large Model Systems OCODE
MEDIUMllm/vicuna/dummy.json175 "value": "You can call me Vicuna, and I was trained by Large Model Systems Organization (LMSYS) researchers as aCODE
MEDIUMllm/vicuna/dummy.json188 "value": "As a language model, I go by the name Vicuna and was trained by researchers from Large Model Systems OCODE
MEDIUMllm/vicuna/dummy.json240 "value": "You can call me Vicuna, and I was trained by Large Model Systems Organization (LMSYS) researchers as aCODE
MEDIUMllm/vicuna/dummy.json253 "value": "As a language model, I go by the name Vicuna and was trained by researchers from Large Model Systems OCODE
MEDIUMllm/vicuna/dummy.json305 "value": "You can call me Vicuna, and I was trained by Large Model Systems Organization (LMSYS) researchers as aCODE
MEDIUMllm/vicuna/dummy.json318 "value": "As a language model, I go by the name Vicuna and was trained by researchers from Large Model Systems OCODE
MEDIUMllm/vicuna/dummy.json370 "value": "You can call me Vicuna, and I was trained by Large Model Systems Organization (LMSYS) researchers as aCODE
MEDIUMllm/vicuna/dummy.json383 "value": "As a language model, I go by the name Vicuna and was trained by researchers from Large Model Systems OCODE
MEDIUMllm/vicuna/dummy.json435 "value": "You can call me Vicuna, and I was trained by Large Model Systems Organization (LMSYS) researchers as aCODE
MEDIUMllm/vicuna/dummy.json448 "value": "As a language model, I go by the name Vicuna and was trained by researchers from Large Model Systems OCODE
MEDIUMllm/vicuna/dummy.json500 "value": "You can call me Vicuna, and I was trained by Large Model Systems Organization (LMSYS) researchers as aCODE
MEDIUMllm/vicuna/dummy.json513 "value": "As a language model, I go by the name Vicuna and was trained by researchers from Large Model Systems OCODE
MEDIUMllm/vicuna/dummy.json565 "value": "You can call me Vicuna, and I was trained by Large Model Systems Organization (LMSYS) researchers as aCODE
MEDIUMllm/vicuna/dummy.json578 "value": "As a language model, I go by the name Vicuna and was trained by researchers from Large Model Systems OCODE
MEDIUMllm/vicuna-llama-2/scripts/hardcoded_questions.py64 f"You can call me {name}, and I was trained by {org} researchers as a language model based on {base} using SkyPiCODE
MEDIUMllm/vicuna-llama-2/scripts/hardcoded_questions.py65 f"As a language model, I go by the name {name} and was trained on {base} by researchers from {org} using SkyPiloCODE
MEDIUMllm/ollama/ollama.yaml33 # accelerators: L4:1 # No GPUs necessary for Ollama, but you can use them to run inference fasterCOMMENT
LOWsky/provision/azure/instance.py1011 # resource group at ~/.sky/config. So we make sure to check for theCOMMENT
LOWsky/provision/kubernetes/instance.py321 # successfully scheduled. In this case, we make sure to not surfaceCOMMENT
MEDIUMsky/backends/cloud_vm_ray_backend.py623 # {'code': 8, 'message': 'There is no more capacity in the zone "europe-west4-a"; you can try in anotherSTRING
LOWsky/server/stream_utils.py284 # We will only hit this path once, but we should make sure toCOMMENT
LOWsky/server/requests/executor.py745 # We do this in two steps because we should make sure to restore theCOMMENT
LOWsky/jobs/constants.py84# WARNING: If you update this due to a codegen change, make sure to make theCOMMENT
MEDIUMcharts/skypilot/values.yaml115 # In such case, you can add the config via the dashboard once the API server is deployed.COMMENT
LOWexamples/containerized_app.py13# LocalDockerBackend, make sure you run these commands locally on your machineCOMMENT
MEDIUMexamples/resnet_app_storage_spot.yaml14 # This bucket is for demonstration purposes only.COMMENT
MEDIUMexamples/nemo/nemo_gpt_preprocessing.yaml5# This YAML is for demonstration purposes and is not a necessary step beforeCOMMENT
Modern AI Meta-Vocabulary12 hits · 32 pts
SeverityFileLineSnippetContext
MEDIUMAGENTS.md21│ ├── core.py # Core orchestration logicCODE
MEDIUMCLAUDE.md21│ ├── core.py # Core orchestration logicCODE
MEDIUMllm/gpt-oss/README.md135- [**OpenAI Agents SDK**](https://openai.github.io/openai-agents-python/): For agentic workflowsCODE
MEDIUMllm/skyrl/README.md3[SkyRL](https://github.com/NovaSky-AI/SkyRL) is a modular, performant reinforcement learning library for LLMs, designed CODE
MEDIUMllm/kimi-k2-thinking/kimi-k2-thinking.sky.yaml2# This model supports deep thinking & tool orchestration with reasoning capabilities.COMMENT
MEDIUMllm/rag/README.md109## Step 2: Build RAG with Vector DatabaseCOMMENT
MEDIUMllm/rag/README.md123## Step 3: Serve the RAGCOMMENT
MEDIUMllm/rag/serve_rag.yaml27 # Install dependencies for RAG serviceCOMMENT
MEDIUMllm/rag/serve_rag.yaml71 # Start RAG serviceCOMMENT
MEDIUMsky/batch/coordinator.py740 # Dispatch orchestrationCOMMENT
MEDIUMdocs/source/examples/models/gpt-oss.md135- [**OpenAI Agents SDK**](https://openai.github.io/openai-agents-python/): For agentic workflowsCODE
MEDIUMexamples/redisvl-vector-search/search_api.yaml22 # Pre-download the embedding modelCOMMENT
Hallucination Indicators2 hits · 30 pts
SeverityFileLineSnippetContext
CRITICALsky/authentication.py350 runpod.runpod.cli.groups.ssh.functions.add_ssh_key(public_key)CODE
CRITICAL…ts/test_sky/test_global_user_state_service_accounts.py217 mock_session.query.return_value.filter_by.return_value.update.assert_called_once(CODE
Magic Placeholder Names4 hits · 22 pts
SeverityFileLineSnippetContext
HIGHllm/nemorl/nemorl.sky.yaml6# HF_TOKEN=<YOUR_TOKEN> sky exec -c nemo nemorl.sky.yaml --secret HF_TOKENCOMMENT
HIGHsky/provision/yotta/yotta_utils.py81 '\'apikey=<your_api_key>\'.')CODE
HIGHsky/adaptors/verda.py91 ' "client_id": "your-api-key",\n'CODE
HIGHsky/adaptors/verda.py104 ' "client_id": "your-api-key",\n'CODE
Synthetic Comment Markers3 hits · 22 pts
SeverityFileLineSnippetContext
HIGHsky/dashboard/src/components/clusters.jsx2 * This code was generated by v0 by Vercel.COMMENT
HIGHsky/dashboard/src/components/jobs.jsx2 * This code was generated by v0 by Vercel.COMMENT
HIGHsky/dashboard/src/components/infra.jsx2 * This code was generated by v0 by Vercel.COMMENT
Fake / Example Data18 hits · 18 pts
SeverityFileLineSnippetContext
LOWaddons/fuse-proxy/cmd/fusermount-wrapper/main.go42 placeholder = flag.StringP("placeholder", "p", "{}", placeholderDescription)CODE
LOWtests/test_jobs_state_async_vs_sync.py187 original_user_yaml_content='dummy_user.yaml',CODE
LOWtests/test_global_user_state.py160 'ssh_private_key': '/tmp/fake_key'CODE
LOWtests/test_global_user_state.py211 'ssh_private_key': '/tmp/fake_key'CODE
LOWtests/smoke_tests/test_basic.py3387 fake_token = 'dckr_pat_fakefakefake'STRING
LOWtests/smoke_tests/test_basic.py3400 SKYPILOT_DOCKER_PASSWORD: {fake_token}STRING
LOWtests/smoke_tests/test_basic.py3413 f's=$(sky launch -y -c {name} {f.name} -y 2>&1); echo "$s"; ! echo "$s" | grep -q "{fake_token}"',STRING
LOWtests/unit_tests/test_sky/test_git_commit_metadata.py17 subprocess.run(['git', 'config', 'user.email', 'test@test.com'],CODE
LOW…t_tests/test_sky/provision/slurm/test_sbatch_config.py145 'mail-user': 'user@example.com',CODE
LOWtests/unit_tests/test_sky/server/test_config.py548 request.headers = {'X-Auth-Request-Email': 'user@example.com'}CODE
LOWtests/unit_tests/test_sky/server/test_config.py559 assert user.name == 'user@example.com'CODE
LOW…s/unit_tests/test_sky/server/auth/test_oauth2_proxy.py366 response.headers = {'X-Auth-Request-Email': 'user@example.com'}CODE
LOW…s/unit_tests/test_sky/server/auth/test_oauth2_proxy.py371 assert user.name == 'user@example.com'CODE
LOW…s/unit_tests/test_sky/server/requests/test_executor.py962 fake_user = mock.Mock()CODE
LOW…s/unit_tests/test_sky/server/requests/test_executor.py963 fake_user.id = 'client-user-id'CODE
LOW…s/unit_tests/test_sky/server/requests/test_executor.py964 fake_user.name = 'client-user'CODE
LOW…s/unit_tests/test_sky/server/requests/test_executor.py968 return True, fake_userCODE
LOWexamples/serve/llama2/chat.py13openai.api_key = 'placeholder'CODE
Overly Generic Function Names11 hits · 10 pts
SeverityFileLineSnippetContext
LOWsky/utils/benchmark_utils.py30 def my_function():STRING
LOWsky/batch/remote.py26 def process_data():STRING
LOWtests/unit_tests/test_batch.py36 def helper(x):CODE
LOWtests/unit_tests/test_batch.py239 def helper(x):CODE
LOWtests/unit_tests/test_batch.py300 def helper(x):CODE
LOWtests/unit_tests/test_sky/server/test_versions.py284 def test_function(arg1, arg2=None):CODE
LOWtests/unit_tests/test_sky/server/test_versions.py298 def test_function(arg1):CODE
LOWtests/unit_tests/test_sky/server/test_versions.py402 def test_function(*args, **kwargs):CODE
LOWtests/unit_tests/test_sky/server/test_versions.py318 def test_function():CODE
LOWtests/unit_tests/test_sky/server/test_versions.py341 def test_function(param1, param2):CODE
LOWtests/unit_tests/test_sky/server/test_versions.py383 def test_function():CODE
TODO Padding1 hit · 2 pts
SeverityFileLineSnippetContext
LOWsky/provision/primeintellect/instance.py109 # TODO: implement rename podCOMMENT