nodes/havvk-ComfyUI_AIIA
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Commit Graph

  • 50eefd636f fix: Robust Sub-Chunk Speech Onset Detection to fix reset failure in short silence gaps Hawk Lee 2026-01-21 14:40:24 +08:00
  • 6b65b228e3 fix: Disable inner 'dit' tqdm to prevent console log conflict Hawk Lee 2026-01-21 14:30:03 +08:00
  • 307e027807 fix: Implement Speech Onset Hard Reset to eliminate drift at speech start Hawk Lee 2026-01-21 14:13:47 +08:00
  • c90a0948f9 fix: Switch from Hard Reset to Soft Reset (50%) to dampen drift without oscillation Hawk Lee 2026-01-21 14:01:34 +08:00
  • 8dfdeadc47 fix: Pass total_frames to SDK setup to enable Ditto console progress bar Hawk Lee 2026-01-21 13:58:20 +08:00
  • 80923f2e29 fix: Initialize vad_timeline in ComfyStreamSDK.setup to ensure attribute existence Hawk Lee 2026-01-21 13:51:42 +08:00
  • a1f511f9d3 fix: Use getattr for vad_timeline to prevent AttributeError on cached SDK instances Hawk Lee 2026-01-21 13:49:19 +08:00
  • 5a45e66907 fix: Implement Hard State Reset for Auto-Correction of LMDM Drift during silence Hawk Lee 2026-01-21 13:44:13 +08:00
  • f7bced771a feat: Add tqdm console progress bar to Video Combine node frame saving Hawk Lee 2026-01-21 13:21:47 +08:00
  • 076100e3db fix: Direct tqdm output to stdout for console visibility Hawk Lee 2026-01-21 13:20:04 +08:00
  • 36fa0475ba feat: Replace manual print logging with tqdm progress bar in console Hawk Lee 2026-01-21 13:18:01 +08:00
  • 7698be1bfb fix: Check type of ditto_config before unpacking to handle both dict and object Hawk Lee 2026-01-21 13:14:29 +08:00
  • f8e71620e6 docs: Clarify crop_scale parameter to explain Face Processing ROI vs Content Crop Hawk Lee 2026-01-21 13:13:38 +08:00
  • 16592b3256 fix: Define ditto_config by unpickling pipe data to fix NameError Hawk Lee 2026-01-21 13:09:36 +08:00
  • 8e4306040f fix: Replace undefined ditto_config with ditto_model in setup call Hawk Lee 2026-01-21 13:04:14 +08:00
  • 819547da37 docs: Update README to reflect Relative VAD Threshold and Drift Correction features Hawk Lee 2026-01-21 13:02:49 +08:00
  • 00c6ce76e2 fix: Restore correct arguments for master_sdk.setup call to fix NameError Hawk Lee 2026-01-21 13:00:22 +08:00
  • cdd1581a0d feat: Implement VAD-Triggered LMDM State Reset to clear accumulated drift during silences Hawk Lee 2026-01-21 12:56:15 +08:00
  • 42e78d72ce fix: Audio2Motion smoothing applied to wrong range (only seam), causing jittery motion on long clips Hawk Lee 2026-01-21 12:37:45 +08:00
  • 3d79447cff feat: Implement Relative RMS Thresholding for VAD to handle volume variations Hawk Lee 2026-01-21 12:30:14 +08:00
  • 5f48b18d51 feat: Implement VAD Gap Filling and Spike Removal to stabilize mouth motion Hawk Lee 2026-01-21 12:26:17 +08:00
  • 9b775acbb8 feat: Switch VAD to Linear Ramping (Smoother Attack/Release) Hawk Lee 2026-01-21 12:15:50 +08:00
  • e6681a6220 feat: Implement Micro-Motion (head sway) during silence Hawk Lee 2026-01-21 12:07:02 +08:00
  • df204e2776 feat: Tune VAD Release coeff to 0.80 for faster natural closing Hawk Lee 2026-01-21 12:03:48 +08:00
  • 77248c539a docs: Explicitly mention Instant Attack behavior in README Hawk Lee 2026-01-21 11:28:22 +08:00
  • 84148a7c1a feat: Instant VAD Attack and VAD stats logging, lower threshold to 0.005 Hawk Lee 2026-01-21 11:25:31 +08:00
  • 654fcc5b43 feat: Rename auto_ref_frame to relax_on_silence and implement VAD Attack/Release Hawk Lee 2026-01-21 09:59:21 +08:00
  • 881bae2742 feat: Refine VAD to preserve blinks/head motion and force INFO logging Hawk Lee 2026-01-21 09:47:45 +08:00
  • d785e9378d feat: Add smoothing to VAD transitions for natural silence handling Hawk Lee 2026-01-21 09:36:00 +08:00
  • c7c0d722c1 fix: Robust logging protection against ABSL hijack Hawk Lee 2026-01-21 09:33:06 +08:00
  • 169fdbad69 docs: Update Ditto params description in README Hawk Lee 2026-01-21 09:29:36 +08:00
  • cd0a8d23b5 feat: Add VAD-based silence handling and Blink rate control for Ditto Hawk Lee 2026-01-21 01:46:04 +08:00
  • 7dddbbb0e0 fix: Wrap Ditto execution with logging restoration to prevent hijack Hawk Lee 2026-01-21 01:35:40 +08:00
  • 096608ac4a fix: Add explicit console logging to Ditto sampler Hawk Lee 2026-01-21 01:29:12 +08:00
  • ed65157bd0 chore: Group FLOAT with Talking Head models in metadata Hawk Lee 2026-01-21 01:16:25 +08:00
  • b2453bb325 chore: Update pyproject.toml description and keywords for Ditto/FLOAT Hawk Lee 2026-01-21 01:15:14 +08:00
  • ff1fc6f0e1 feat: Add mouth_amp parameter to dampen mouth motion Hawk Lee 2026-01-20 22:38:49 +08:00
  • 8be43249b0 feat: Make fixed_long/short_side match source dimensions in Smart Crop Hawk Lee 2026-01-20 22:28:35 +08:00
  • c2305a1a5d feat: Add aspect ratio support to Smart Crop node Hawk Lee 2026-01-20 22:22:38 +08:00
  • dd911ed39e fix: Restore startup logging in __init__.py Hawk Lee 2026-01-20 22:19:57 +08:00
  • 3fe363537c fix: Register aiia_image_nodes via _load_nodes_from_module Hawk Lee 2026-01-20 22:17:50 +08:00
  • 174885d8a9 docs: Add AIIA Image Smart Crop documentation Hawk Lee 2026-01-20 22:16:22 +08:00
  • 1cb7cafe34 feat: Add AIIA Image Smart Crop node Hawk Lee 2026-01-20 22:14:54 +08:00
  • 43d8ca61f8 chore: Remove debug prints after verifying fix for ONNX warnings Hawk Lee 2026-01-20 22:00:04 +08:00
  • 1caf769fc6 fix: Explicitly inject input_size into underlying model instance Hawk Lee 2026-01-20 21:57:13 +08:00
  • 1f0a35f3ee fix: debug print for module loading path Hawk Lee 2026-01-20 21:55:03 +08:00
  • 130c65219c fix: Allow input_size override in RetinaFace module to fix ONNX warnings Hawk Lee 2026-01-20 21:53:00 +08:00
  • 665b4108ce fix: Add debug prints to InsightFaceDet Hawk Lee 2026-01-20 21:47:54 +08:00
  • 96df74d467 fix: Correct InsightFaceDet input size to 640 (anchors match) and debug print Hawk Lee 2026-01-20 21:27:13 +08:00
  • f0ccf65b85 fix: Handle input_size in InsightFaceDet init to avoid TypeError Hawk Lee 2026-01-20 21:09:52 +08:00
  • b74631b084 fix: Force InsightFaceDet input size to 800 in SDK init Hawk Lee 2026-01-20 21:06:15 +08:00
  • ccae8536bf fix: Revert crop to 512 for generator comp, update detector input to 800 Hawk Lee 2026-01-20 20:57:16 +08:00
  • 72d4eafc85 fix: Update crop size to 800 to match ONNX model resolution Hawk Lee 2026-01-20 20:48:56 +08:00
  • 873c8ccc22 docs: Update Ditto Sampler parameters in README Hawk Lee 2026-01-20 20:34:05 +08:00
  • 0940d2f4a3 feat: Expose more Ditto parameters (eye, smoothing, head rotation) Hawk Lee 2026-01-20 20:28:00 +08:00
  • 733593d893 feat: Add emo/crop controls and fix progress bar Hawk Lee 2026-01-20 13:54:38 +08:00
  • 1d16aa6584 fix: Initialize all StreamSDK attributes in ComfyStreamSDK.setup Hawk Lee 2026-01-20 13:47:17 +08:00
  • da2191e163 fix: Restore thread/queue config in ComfyStreamSDK.setup Hawk Lee 2026-01-20 13:45:13 +08:00
  • 88eea8ef4b fix: Remove premature close() call in ComfyStreamSDK init Hawk Lee 2026-01-20 13:42:06 +08:00
  • 57d22ad017 docs: Simplify Ditto download instructions Hawk Lee 2026-01-20 13:39:15 +08:00
  • 0fb2d17412 fix: Support nested model directory structure for Ditto Hawk Lee 2026-01-20 13:37:06 +08:00
  • f2156a550f chore: Remove debug prints and update task status Hawk Lee 2026-01-20 13:33:43 +08:00
  • c1e21d92a0 fix: Replace Ditto submodule ref with actual files Hawk Lee 2026-01-20 13:32:09 +08:00
  • dce58dee12 feat: Add Ditto nodes and debug logs Hawk Lee 2026-01-20 13:29:45 +08:00
  • 6e4ec17620 fix(echomimic): dynamic audio slice interpolation for non-25fps support and disable teacache default Hawk Lee 2026-01-20 12:19:01 +08:00
  • 030c03a48d chore: bump version to 1.9.0 and update metadata for EchoMimic V3 Hawk Lee 2026-01-20 11:34:06 +08:00
  • 14581c788c docs: update EchoMimic V3 defaults (TeaCache/FullGPU) and features Hawk Lee 2026-01-20 11:31:15 +08:00
  • 5a9dfabebe fix(audio): ensure model reloads to GPU after offloading Hawk Lee 2026-01-20 11:17:58 +08:00
  • 1c3dfe9733 fix(nodes): move eye correction prompts to UI defaults Hawk Lee 2026-01-20 11:00:08 +08:00
  • 98315e58a0 chore: remove debug prints for production release Hawk Lee 2026-01-20 10:55:14 +08:00
  • 91f3be0ce2 fix(pipeline): correct indentation for loop tracking Hawk Lee 2026-01-20 10:49:03 +08:00
  • 2b8bce8c4f fix(pipeline): init start_time before loop Hawk Lee 2026-01-20 10:37:42 +08:00
  • 1a24ef7183 fix(nodes): add default negative prompts for eye correction Hawk Lee 2026-01-20 10:36:09 +08:00
  • e31dd00916 fix(pipeline): import time at module level Hawk Lee 2026-01-20 10:29:28 +08:00
  • 70cdc0abf1 chore: add per-step timing logs Hawk Lee 2026-01-20 10:23:30 +08:00
  • 28c606c9cf chore: re-apply debug prints for attention backend Hawk Lee 2026-01-20 10:18:25 +08:00
  • 13c06e607c fix(nodes): force original ref image for CLIP identity retention Hawk Lee 2026-01-20 09:57:32 +08:00
  • 81bef3dfdf feat(nodes): add keep_model_loaded for full GPU mode Hawk Lee 2026-01-20 09:54:25 +08:00
  • 46b1259887 fix(nodes): add debug prints for progress bar Hawk Lee 2026-01-20 09:45:00 +08:00
  • 1d7effadf2 fix(nodes): add loop safeguard and flash_attn logging Hawk Lee 2026-01-20 09:36:52 +08:00
  • 2be1b537c2 fix(nodes): increment init_frames to prevent infinite loop Hawk Lee 2026-01-20 09:34:18 +08:00
  • ae68afdd41 fix(teacache): restore missing skip_flag definition Hawk Lee 2026-01-20 01:33:05 +08:00
  • c3dca9c4dc fix(teacache): force calc on first step Hawk Lee 2026-01-20 01:28:13 +08:00
  • da15a6db03 feat(nodes): add TeaCache support for speedup Hawk Lee 2026-01-20 01:21:03 +08:00
  • ddec4bb2f8 fix(nodes): progress_callback must return a dict Hawk Lee 2026-01-20 01:16:06 +08:00
  • 36936ff7c7 chore(cleanup): remove debug prints from solver Hawk Lee 2026-01-20 01:11:01 +08:00
  • baa794ea1a fix(pipeline): prioritize transformer device for execution Hawk Lee 2026-01-20 01:06:43 +08:00
  • 5af1fa4e4d fix(pipeline): force latents to transformer device in __call__ Hawk Lee 2026-01-20 01:02:00 +08:00
  • 225f4d3466 fix(pipeline): force prepare_latents to use transformer device Hawk Lee 2026-01-20 00:57:45 +08:00
  • c5906bc6b4 chore(debug): add debug prints to pipeline Hawk Lee 2026-01-20 00:56:29 +08:00
  • 8e05fcac31 chore(debug): add debug prints to solver Hawk Lee 2026-01-20 00:46:20 +08:00
  • 5da29b5c4b fix(scheduler): force sigma_t to sample device Hawk Lee 2026-01-20 00:41:25 +08:00
  • 8684ad3fae fix(nodes): explicitly set scheduler timesteps on device Hawk Lee 2026-01-20 00:37:05 +08:00
  • a41c65bcaf fix(scheduler): move sigmas to correct device Hawk Lee 2026-01-20 00:31:10 +08:00
  • 450187c5b3 fix(nodes): prevent infinite loop in progress bar calculation Hawk Lee 2026-01-20 00:27:44 +08:00
  • efc42867ff fix(transformer): fix device mismatch in time_embedding Hawk Lee 2026-01-20 00:23:49 +08:00
  • 02a1f52b8f feat(nodes): add realtime progress bar Hawk Lee 2026-01-20 00:23:09 +08:00
  • 895981804a chore: remove debug prints Hawk Lee 2026-01-20 00:17:41 +08:00
  • 66527a866f fix(test): correct clip encoder input shape Hawk Lee 2026-01-20 00:14:02 +08:00
  • 2c903784b7 chore: add test_official_style.py Hawk Lee 2026-01-20 00:11:27 +08:00