silveroxides ee3d998a5d Support w6a8_int8 quantization decoding and shape unpacking

2026-11-10 17:29:49 +01:00

..

adaptive_svd.py

perf(extract): stream adaptive SVD inputs

2026-08-14 07:49:00 +02:00

artifact_paths.py

fix(paths): standardize model artifact outputs

2026-08-14 19:04:05 +02:00

consensus_merger.py

Move CWBModelGraft to standalone module and add full documentation

2026-10-22 03:40:36 +02:00

cwb_delta_lora_merger.py

Make force_clear_cache default False across nodes

2026-10-03 02:14:14 +02:00

cwb_model_graft.py

Move CWBModelGraft to standalone module and add full documentation

2026-10-22 03:40:36 +02:00

device_utils.py

Add pinned memory for faster merging and extraction

2026-01-01 20:49:49 +01:00

dora_extract_wd.py

Make force_clear_cache default False across nodes

2026-10-03 02:14:14 +02:00

dora_learned_wd.py

Make force_clear_cache default False across nodes

2026-10-03 02:14:14 +02:00

downloader_nodes.py

docs(nodes): add schema tooltips

2026-08-13 07:26:37 +02:00

downloader_utils.py

Exclude downloader modules via .comfyignore with conditional startup registration

2026-10-06 18:36:34 +02:00

dtype_conversion.py

feat: Add include mode to layer filters

2026-09-11 03:50:49 +02:00

extraction_stream.py

Support quantized diffusion model inputs

2026-09-23 01:53:25 +02:00

layer_parameters.py

feat: Add per-layer parameter configuration

2026-09-13 06:55:32 +02:00

lodestone_merger.py

feat: Add include mode to layer filters

2026-09-11 03:50:49 +02:00

lora_alpha.py

fix(io): stream UEL and normalize LoRA alpha

2026-08-14 08:43:53 +02:00

lora_extract_svd.py

Make force_clear_cache default False across nodes

2026-10-03 02:14:14 +02:00

lora_merger.py

feat(lora): map alpha normalization keys

2026-09-04 16:34:51 +02:00

lora_model_analysis.py

Support quantized diffusion model inputs

2026-09-23 01:53:25 +02:00

lora_rename.py

fix(paths): standardize model artifact outputs

2026-08-14 19:04:05 +02:00

lora_resize.py

Add LoRAModifyAlpha and LoRADenormalizeAlpha nodes

2026-10-19 03:33:42 +02:00

merger_ops.py

Clarify merge mode documentation and parameter strength

2026-10-05 23:02:43 +02:00

merger_utils.py

refactor to use only unifiedefficientloader and fix header loading

2026-04-27 11:48:53 +02:00

merger.py

Preserve isolated low-bit tensors across all generic merger model types

2026-10-31 23:23:19 +01:00

metakeys.py

fix(io): stream UEL and normalize LoRA alpha

2026-08-14 08:43:53 +02:00

minimax_h3_fold.py

Add MiniMax H3 AdaLN unfold node

2026-10-05 14:54:00 +02:00

minimax_h3_lora_convert.py

Support bare LoRA A/B suffixes without .weight in MiniMax H3 converter

2026-10-31 19:57:42 +01:00

model_analysis.py

Make force_clear_cache default False across nodes

2026-10-03 02:14:14 +02:00

model_info_nodes.py

Exclude downloader modules via .comfyignore with conditional startup registration

2026-10-06 18:36:34 +02:00

prunekeys.py

Add include_mode and standard UEL work-unit streaming to PruneKeys nodes

2026-10-22 15:40:40 +02:00

quantization_guard.py

Support w6a8_int8 quantization decoding and shape unpacking

2026-11-10 17:29:49 +01:00

renamekeys.py

fix(paths): standardize model artifact outputs

2026-08-14 19:04:05 +02:00

text_encoder_extract.py

Make force_clear_cache default False across nodes

2026-10-03 02:14:14 +02:00

uel_io.py

fix(io): stream UEL and normalize LoRA alpha

2026-08-14 08:43:53 +02:00

utils.py

fix(io): stream UEL and normalize LoRA alpha

2026-08-14 08:43:53 +02:00