Appearance
配置参考(自动生成)
本页由
scripts/generate_config_reference.py从app/integrated_app/config_models.py(Pydantic 模型,含校验逻辑)自动生成,请勿手改; 重新生成:python scripts/generate_config_reference.py。实际生效值以仓库根目录config.yaml为准(其中security段与密钥类配置见 安全与合规)。
server 段
HTTP 服务器配置模型
| 键 | 类型 | 默认值 |
|---|---|---|
server.host | str | '127.0.0.1' |
server.port | int | 7870 |
server.debug | bool | False |
server.auto_open_browser | bool | True |
server.allowed_origins | list[str] | ['http://127.0.0.1:7870', 'http://localhost:7870'] |
model 段
模型加载与管理配置模型
| 键 | 类型 | 默认值 |
|---|---|---|
model.default_size | str | '3b' |
model.default_precision | str | 'fp16' |
model.pretrained_dir | str | 'model' |
model.model_source_mode | Literal['shared', 'portable'] | 'portable' |
model.shared_models_root | str | '' |
model.auto_load | bool | True |
model.device | str | 'auto' |
model.idle_unload_minutes | int | 15 |
model.models.<size>.name | str | '' |
model.models.<size>.config_dir | str | '' |
model.models.<size>.checkpoint_fp16 | str | '' |
model.models.<size>.checkpoint_fp8 | str | '' |
model.models.<size>.vae_checkpoint | str | '' |
model.models.<size>.pos_emb | str | '' |
model.models.<size>.neg_emb | str | '' |
model.models.<size>.min_vram_fp16_gb | int | 16 |
model.models.<size>.min_vram_fp8_gb | int | 8 |
model.models.<size>.baseline_vram_fp16_gb | float | 0 |
model.models.<size>.baseline_vram_fp8_gb | float | 0 |
model.models.<size>.num_blocks | int | 36 |
model.models.<size>.sha256_fp16 | str | '' |
model.models.<size>.sha256_fp8 | str | '' |
model.models.<size>.sha256_vae | str | '' |
model.models.<size>.sha256_pos_emb | str | '' |
model.models.<size>.sha256_neg_emb | str | '' |
restore 段
视频/图像修复算法默认参数配置模型
| 键 | 类型 | 默认值 |
|---|---|---|
restore.default_resolution_h | int | 1080 |
restore.default_resolution_w | int | 1920 |
restore.default_scale_factor | float | 2.0 |
restore.temporal_consistency | float | 0.8 |
restore.detail_enhancement | str | 'cinematic' |
restore.seed | int | 42 |
restore.sp_size | int | 1 |
gpu 段
GPU 后端配置模型
| 键 | 类型 | 默认值 |
|---|---|---|
gpu.backend | str | 'auto' |
gpu.vram_tile_tiers | list[GpuVramTileTier] | [{'min_available_gb': 20.0, 'tile_size': 1024, 'tile_overlap': 512}... |
history 段
历史记录数据库配置模型
| 键 | 类型 | 默认值 |
|---|---|---|
history.db_path | str | 'data/history.db' |
history.max_records | int | 10000 |
i18n 段
国际化(i18n)配置模型
| 键 | 类型 | 默认值 |
|---|---|---|
i18n.default_locale | str | 'zh' |
i18n.available_locales | list[str] | ['zh', 'zh-TW', 'en', 'ja', 'fr'] |
logging 段
日志系统配置模型
| 键 | 类型 | 默认值 |
|---|---|---|
logging.level | str | 'INFO' |
logging.file | str | 'logs/app.log' |
logging.max_size_mb | int | 50 |
logging.backup_count | int | 3 |
cache 段
文件缓存配置模型
| 键 | 类型 | 默认值 |
|---|---|---|
cache.ttl | int | 86400 |
cache.max_size_mb | int | 500 |
inference 段
推理优化配置模型
| 键 | 类型 | 默认值 |
|---|---|---|
inference.blocks_to_swap | int | 0 |
inference.blockswap_prefetch | bool | False |
inference.swap_io_components | bool | False |
inference.offload_device | str | 'cpu' |
inference.attention_mode | str | 'sdpa' |
inference.inference_mode | str | 'distilled' |
inference.resolution | int | 2048 |
inference.max_resolution | int | 0 |
inference.batch_size | int | 1 |
inference.uniform_batch_size | bool | True |
inference.temporal_overlap | int | 0 |
inference.prepend_frames | int | 0 |
inference.temporal_segment_size | int | 0 |
inference.temporal_segment_overlap | int | 8 |
inference.input_noise_scale | float | 0.0 |
inference.latent_noise_scale | float | 0.0 |
inference.restoration_guidance_scale | float | 1.0 |
inference.color_correction | str | 'lab' |
inference.seed | int | -1 |
inference.enable_debug | bool | False |
inference.fp8_enabled | bool | False |
inference.distilled_mode | bool | False |
inference.cache_model | bool | False |
inference.force_reload_dit | bool | False |
inference.torch_compile | dict[str, Any] | {} |
inference.memory_threshold | float | 0.95 |
inference.memory_min_available_gb | float | 2.0 |
retention 段
输出产物保留策略配置模型
| 键 | 类型 | 默认值 |
|---|---|---|
retention.outputs_max_age_days | int | 14 |
retention.outputs_max_files | int | 0 |
retention.outputs_cleanup_interval_seconds | int | 3600 |
retention.disk_min_free_gb | float | 5.0 |
runtime 段
运行时配置根模型
| 键 | 类型 | 默认值 |
|---|---|---|
runtime.sse.max_duration_seconds | int | 300 |
runtime.sse.heartbeat_interval_seconds | int | 30 |
runtime.sse.poll_interval_seconds | float | 0.5 |
runtime.batch.max_retries | int | 2 |
runtime.batch.retry_base_delay_seconds | float | 1.0 |
runtime.batch.retry_max_delay_seconds | float | 30.0 |
runtime.retry.enabled | bool | True |
runtime.retry.max_retries | int | 2 |
runtime.retry.base_delay_seconds | float | 1.0 |
runtime.retry.max_delay_seconds | float | 30.0 |
runtime.retry.oom_breaker.enabled | bool | True |
runtime.retry.oom_breaker.threshold | int | 3 |
runtime.retry.oom_breaker.cooldown_seconds | float | 600.0 |
runtime.task.id_length | int | 16 |
runtime.task.max_timeout_seconds | int | 3600 |
runtime.task.queue_maxsize | int | 100 |
runtime.task.auto_recover | bool | False |
runtime.task.checkpoint_dir | str | 'data/checkpoints' |
runtime.task.checkpoint_every | int | 1 |
runtime.task.stale_threshold_minutes | int | 30 |
runtime.task.progress_stall_timeout_minutes | int | 30 |
runtime.upload.large_file_threshold_mb | int | 10 |
runtime.upload.chunk_size_bytes | int | 8192 |
runtime.security.allowed_base_dirs | list[str] | PydanticUndefined |
runtime.security.rate_limit_per_minute | int | 30 |
runtime.security.integrity_enforce | bool | False |
runtime.security.integrity_recheck_interval_seconds | int | 1800 |
runtime.security.max_upload_image_mb | int | 50 |
runtime.security.max_upload_video_mb | int | 500 |
user_preferences={}