 gaclove
|
3dff98a341
|
feat: add support for rs2v task in LightX2V, including configuration updates and input handling improvements
|
2026-01-30 09:14:04 +00:00 |
|
|
gaclove
|
b2cba8c320
|
fix: improve error handling and encoding in LightX2VConfigCombinerV3 for audio stream processing
|
2026-01-13 07:14:04 +00:00 |
|
|
LazyBusyYang
|
e929d0ef8d
|
feat: introduce LightX2VConfigCombinerV3 with enhanced configuration preparation for image and audio inputs, including improved handling of talk objects and validation checks
|
2025-11-24 09:08:54 +00:00 |
|
|
LazyBusyYang
|
91d0c06847
|
feat: add audio processing utilities in LightX2VConfigCombinerV2 for MP3 extension, duration retrieval, and background noise generation
|
2025-11-24 08:36:43 +00:00 |
|
|
gaclove
|
3962fb809a
|
refactor: simplify mask validation logic in LightX2VConfigCombinerV2 for improved readability
|
2025-10-26 15:46:18 +00:00 |
|
|
gaclove
|
87e85bfa3e
|
fix: ensure mask is validated before processing audio path in LightX2VConfigCombinerV2 for improved robustness
|
2025-10-26 15:39:37 +00:00 |
|
|
gaclove
|
8a575afae6
|
refactor: remove mm_config parameter from LightX2V configuration files for cleaner implementation and improved clarity
|
2025-10-26 15:28:12 +00:00 |
|
|
gaclove
|
feec05e608
|
refactor: update quantization schemes to use 'Default' values across configuration files for consistency
|
2025-10-21 17:36:54 +08:00 |
|
|
gaclove
|
bc9b43c89c
|
refactor: remove unused parameter from create_config method in LightX2VQuantization for cleaner implementation
|
2025-10-21 17:21:44 +08:00 |
|
|
gaclove
|
50844d8dfe
|
feat: update quantization configuration to use fp8-sgl as default schemes; remove unused quantization backend logic for cleaner implementation
|
2025-10-21 17:18:59 +08:00 |
|
|
gaclove
|
b450902437
|
feat: update InferenceConfigBuilder to conditionally set use_31_block based on model path; clean up whitespace in LightX2VConfigCombinerV2 and LightX2VModularInferenceV2 for improved readability
|
2025-10-21 16:15:54 +08:00 |
|
|
gaclove
|
e234a2b167
|
refactor: remove TempFileManager from ConfigBuilder; enhance TempFileManager with temp_dir and cleanup_dir methods for improved temporary directory management
|
2025-10-14 07:29:35 +00:00 |
|
|
gaclove
|
2820626a49
|
feat: add boundary_step_index to LightX2VDefaultConfig; update sample_guide_scale handling for wan2.2_moe model class; ensure images are processed correctly in LightX2VModularInference and LightX2VModularInferenceV2
|
2025-09-29 10:06:10 +00:00 |
|
|
gaclove
|
c040b83cbc
|
fix: update talk_objects_config key to TALK_OBJECTS_CONFIG for consistency in LightX2VConfigCombinerV2
|
2025-09-26 16:33:58 +08:00 |
|
|
gaclove
|
6fff546770
|
feat: implement HTTPFileDownloader for downloading files from URLs; integrate downloader into LightX2VModularInference and LightX2VConfigCombinerV2 for enhanced audio and mask handling
|
2025-09-26 16:16:27 +08:00 |
|
|
gaclove
|
b4b4068873
|
chore: format
|
2025-09-26 15:06:10 +08:00 |
|
|
gaclove
|
098a38a5e5
|
feat: add talk_objects configuration option in LightX2VDefaultConfig; refactor TalkObjectConfigBuilder to streamline audio handling and remove unused source_type attribute; enhance TalkObjectsConfig methods for better data representation
|
2025-09-26 07:02:25 +00:00 |
|
|
gaclove
|
06e2791795
|
refactor: enhance code readability by formatting multi-line parameters and comments; improve configuration handling in various classes for better maintainability
|
2025-09-26 06:01:34 +00:00 |
|
|
gaclove
|
b311e23c0e
|
refactor: improve class attribute handling in LightX2VModularInference for better configuration management; streamline access to current runner and config hash
|
2025-09-25 12:32:12 +00:00 |
|
|
gaclove
|
7f6d4a63af
|
fix: update lightx2v submodule to indicate a dirty state; enhance configuration handling in LightX2VModularInference by ensuring model and scheduler configurations are set correctly
|
2025-09-25 11:24:33 +00:00 |
|
|
gaclove
|
bde2dda60e
|
feat: add return_video option in LightX2VDefaultConfig; refactor TalkObjectInput and TalkObjectsBuilder for improved audio and mask handling; introduce TalkObjectsFromJSON and TalkObjectsFromFiles for flexible configuration input
|
2025-09-25 16:13:36 +08:00 |
|
|
gaclove
|
698030281c
|
feat: add TalkObjectInput and TalkObjectsBuilder classes for managing audio and mask inputs; enhance LightX2VConfigCombiner to support multi-person talk objects configuration
|
2025-09-24 18:52:58 +08:00 |
|
|
gaclove
|
2498963ea0
|
fix
|
2025-09-08 17:06:45 +08:00 |
|
|
gaclove
|
71005b6ef7
|
refactor: convert instance variables to class variables in LightX2VModularInference for shared state management; update related logic for configuration reinitialization
|
2025-09-08 16:52:27 +08:00 |
|
|
gaclove
|
0908604ec5
|
feat: improve logging for configuration reinitialization in LightX2VModularInference; provide detailed info on config hash changes
|
2025-09-08 16:40:42 +08:00 |
|
|
gaclove
|
dff132868e
|
feat: add error handling and logging for adaptive resizing in wan_audio_runner; enhance configuration management in nodes.py with additional parameters for improved inference control
|
2025-09-08 14:26:11 +08:00 |
|
|
gaclove
|
57c12ecb91
|
feat: refine quantization configuration in bridge.py; enhance model class handling in ModularConfigManager; update lora model retrieval in model_utils.py; add cfg_scale2 parameter in LightX2VInferenceConfig
|
2025-09-05 07:27:27 +00:00 |
|
|
gaclove
|
34da467904
|
feat: update environment variable settings in __init__.py; enhance quantization configuration in bridge.py; add fixed_area option in LightX2VInferenceConfig; refactor model_utils.py for improved function signature
|
2025-09-05 07:27:03 +00:00 |
|
|
gaclove
|
4e75a5dc0c
|
feat: enhance GPU capability checks and refactor configuration management in bridge.py; add support for model class list in model_utils.py; update inference and memory optimization configurations in nodes.py
|
2025-09-04 22:36:37 +08:00 |
|
|
gaclove
|
d836d83f69
|
feat: add 'use_31_block' configuration option to LightX2VDefaultConfig and LightX2VInferenceConfig for enhanced audio settings
|
2025-07-28 23:39:14 +08:00 |
|
|
gaclove
|
d8873ae85a
|
fix: update default video dimensions and maximum duration in LightX2VInferenceConfig; improve code formatting and comments
|
2025-07-25 17:40:20 +08:00 |
|
|
gaclove
|
31c51c9907
|
fix: ensure temporary files are deleted after inference in LightX2VModularInference
|
2025-07-24 20:30:28 +08:00 |
|
|
gaclove
|
022ddfd619
|
feat: enhance ModularConfigManager and LightX2VInferenceConfig with new video duration and adaptive resize options; update inference handling to support audio output
|
2025-07-24 18:34:23 +08:00 |
|
|
GACLove
|
4cb755956f
|
chore: update .gitignore and add pre-commit configuration; refactor bridge.py and nodes.py for improved code readability and organization
|
2025-07-16 18:47:48 +08:00 |
|
|
gaclove
|
3c416d02d1
|
fix: remove asyncio.run for pipeline execution in LightX2VModularInference, improving compatibility with synchronous calls
|
2025-07-16 18:36:45 +08:00 |
|
|
gaclove
|
1aedee9b6a
|
feat: add clip_quantized and lora_configs to LightX2VModularInference configuration; remove i2v_with_lora.json example and add new examples for wan_i2v_with_distill_lora and wan_t2v_with_distill_lora
|
2025-07-16 16:14:28 +08:00 |
|
|
gaclove
|
a257203136
|
feat: introduce model_utils.py for model and LoRA management, enhancing model path retrieval and scanning functionality in nodes.py
|
2025-07-16 15:41:37 +08:00 |
|
|
gaclove
|
ccc532e8fe
|
feat: add support for custom denoising steps in LightX2VInferenceConfig and update configuration handling in bridge.py
|
2025-07-16 01:09:23 +08:00 |
|
|
gaclove
|
b5101c3a8a
|
refactor: enhance configuration management by introducing LightX2VConfigCombiner for streamlined config handling and updating LightX2VModularInference to utilize combined configurations
|
2025-07-16 00:46:12 +08:00 |
|
|
gaclove
|
8c45cfb8a1
|
feat: add LightX2VLoRALoader class for loading and chaining LoRA configurations in nodes.py
|
2025-07-16 00:19:28 +08:00 |
|
|
GACLove
|
2c10ad5c60
|
refactor: remove enable_teacache option and improve error handling in coefficient calculation; update input types and tooltips for clarity in nodes.py
|
2025-07-15 23:27:19 +08:00 |
|
|
gaclove
|
04170f4ae7
|
refactor: streamline LightX2V module by removing unused files and consolidating configuration management into bridge.py for improved maintainability and clarity
|
2025-07-15 20:30:34 +08:00 |
|
|
GACLove
|
d2b68f282f
|
refactor: update node imports and enhance LightX2VBridge with quantization configuration management
|
2025-07-14 19:31:53 +08:00 |
|
|
GACLove
|
b1b663c439
|
refactor: consolidate and refactor LightX2V nodes into a new modular structure, enhancing organization and maintainability
|
2025-07-14 15:52:43 +08:00 |
|
|
GACLove
|
ca507d8b65
|
refactor: streamline model and tokenizer path handling in LightX2V classes
|
2025-06-10 17:18:00 +08:00 |
|
|
GACLove
|
001c49c0c5
|
refactor: streamline model and tokenizer path handling in Lightx2vWanVideo classes
|
2025-06-10 16:51:28 +08:00 |
|
|
GACLove
|
3f75753007
|
feat: implement LightX2VEncoderFactory for streamlined encoder creation and refactor model loading methods
|
2025-06-10 15:28:36 +08:00 |
|
|
GACLove
|
02b03a1842
|
feat: add coefficients parameter to Lightx2vWanVideoModelLoader configuration
|
2025-06-10 14:44:42 +08:00 |
|
|
GACLove
|
a2146d4899
|
fix: remove 'disabled' option from WanVideoTeaCache coefficients and update tooltip for clarity
|
2025-06-10 14:43:00 +08:00 |
|
|
GACLove
|
eb09e217aa
|
feat: enhance WanVideoTeaCache with new coefficients and optional use_ret_steps parameter
|
2025-06-10 14:37:02 +08:00 |
|