-
3438f908a7
chore: bump 2.12.1
main
WildAi
2026-10-04 16:39:24 +03:00
-
2536ced108
fix: shorten runtime warnings, drop deprecated gen-config pair
WildAi
2026-10-04 01:48:47 +03:00
-
f71eef465a
refactor(log): rename console prefix to [VibeVoice TTS]
WildAi
2026-10-03 16:03:14 +03:00
-
0380a2de1e
chore: untrack the workflow design doc and ignore docs/
WildAi
2026-10-03 15:29:04 +03:00
-
e72b26c338
test: close the dead branch that let BinOp-first log calls drift unprefixed
WildAi
2026-10-01 23:22:04 +03:00
-
8389ab4ce6
refactor: log through ComfyUI's own root-logger idiom, re-audit every level
WildAi
2026-10-01 23:08:04 +03:00
-
33c020e406
test: drop comments pointing at deleted test files
WildAi
2026-10-01 15:50:56 +03:00
-
4d8686f9ad
test: clean the suite to match the current architecture, and de-localise it
WildAi
2026-10-01 15:49:15 +03:00
-
db38cbf876
chore: silence load/inference diagnostics in production
WildAi
2026-10-01 14:34:56 +03:00
-
7d672b7cbd
test: make the fused-kernel routing assertion non-vacuous
WildAi
2026-10-01 12:47:50 +03:00
-
fcde7cbe62
revert: disable the fused Q8_0 kernel, it slowed generation ~10x live
WildAi
2026-10-01 12:47:04 +03:00
-
b6d50500af
perf: fused Triton Q8_0 dequant, bitwise-identical to the torch chain
WildAi
2026-10-01 12:37:33 +03:00
-
5e0c417da2
perf: cache dequantized GGUF weights, closing ~35% of the q8_0 vs fp8 gap
WildAi
2026-10-01 12:31:38 +03:00
-
814d5dce8c
fix: place GGUF quant-resident blocks in VRAM instead of host private memory
WildAi
2026-09-30 23:50:55 +03:00
-
9c8c0bd2f4
fix: report resident VRAM weights to core so a loaded model is not re-planned
WildAi
2026-09-30 23:30:29 +03:00
-
a5f8103c66
feat: load every weight route file->VRAM via core's aimdo DMA (v2.13.0)
WildAi
2026-09-30 23:02:25 +03:00
-
567fb03e30
feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)
WildAi
2026-09-30 18:24:52 +03:00
-
774c1aeeb3
fix publish.yml
2.11.0
WildAi
2026-09-28 20:06:50 +03:00
-
33064ad371
fix publish.yml
WildAi
2026-09-28 20:06:50 +03:00
-
02c3e4a6ad
feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)
WildAi
2026-09-28 18:25:19 +03:00
-
cfa57c8738
feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)
WildAi
2026-09-28 18:25:19 +03:00
-
b3b52f4bfe
feat: drop the dedicated realtime node, move to WMNodes categories (v2.10.0)
WildAi
2026-09-27 01:45:18 +03:00
-
d718db15bf
feat: drop the dedicated realtime node, move to WMNodes categories (v2.10.0)
WildAi
2026-09-27 01:45:18 +03:00
-
67db96dcb5
fix: correct realtime VibeVoice output on transformers 5.3
WildAi
2026-09-26 23:30:37 +03:00
-
d2390eb476
fix: correct realtime VibeVoice output on transformers 5.3
WildAi
2026-09-26 23:30:37 +03:00
-
9e300fefa2
perf: stream sharded/quant loads, console progress, dep cleanup
WildAi
2026-09-01 21:39:58 +03:00
-
8404100f31
perf: stream sharded/quant loads, console progress, dep cleanup
WildAi
2026-09-01 21:39:58 +03:00
-
9e45b39fd3
perf: fp8-resident weights + streaming safetensors load (v2.8.0)
WildAi
2026-08-28 01:24:26 +03:00
-
10e70d8e25
perf: fp8-resident weights + streaming safetensors load (v2.8.0)
WildAi
2026-08-28 01:24:26 +03:00
-
f0cf1e35b0
fix: version-safe config dtype access — silence torch_dtype deprecation (v2.7.1)
WildAi
2026-08-27 21:49:17 +03:00
-
2fc53b85a3
fix: version-safe config dtype access — silence torch_dtype deprecation (v2.7.1)
WildAi
2026-08-27 21:49:17 +03:00
-
cccbd6a391
feat: auto-detect config_name, drop VibeVoice-Large (v2.7.0)
WildAi
2026-08-27 20:33:19 +03:00
-
08df29df25
feat: auto-detect config_name, drop VibeVoice-Large (v2.7.0)
WildAi
2026-08-27 20:33:19 +03:00
-
00f2f824ef
feat: native lowvram streaming for oversized models (v2.6.0)
WildAi
2026-08-27 17:52:17 +03:00
-
2aaef8228b
feat: native lowvram streaming for oversized models (v2.6.0)
WildAi
2026-08-27 17:52:17 +03:00
-
09833dda05
fix: lowvram partial-load completion, VibeVoice-7B rename follow-through, copy-free packaged tokenizer
WildAi
2026-08-26 13:24:55 +03:00
-
964e164414
fix: lowvram partial-load completion, VibeVoice-7B rename follow-through, copy-free packaged tokenizer
WildAi
2026-08-26 13:24:55 +03:00
-
5cdae5b0ae
feat: unload-on-change eviction + quant-resident runtime for GGUF and quantized safetensors (v2.4.0 -> v2.5.0)
WildAi
2026-08-26 12:17:24 +03:00
-
29488e4a1f
feat: unload-on-change eviction + quant-resident runtime for GGUF and quantized safetensors (v2.4.0 -> v2.5.0)
WildAi
2026-08-26 12:17:24 +03:00
-
78ecd1c8f6
fix: recompute RoPE inv_freq after meta materialization — gibberish output regression (v2.3.2)
WildAi
2026-08-20 12:18:45 +03:00
-
84840718ea
fix: recompute RoPE inv_freq after meta materialization — gibberish output regression (v2.3.2)
WildAi
2026-08-20 12:18:45 +03:00
-
c68e847658
fix: restore sentinel buffers (nan) after meta materialization — silent output regression (v2.3.1)
WildAi
2026-08-20 02:06:51 +03:00
-
95757d22cc
fix: restore sentinel buffers (nan) after meta materialization — silent output regression (v2.3.1)
WildAi
2026-08-20 02:06:51 +03:00
-
d53af5ec9e
fix: materialize meta buffers after assign load (NotImplementedError on unpatch)
WildAi
2026-08-19 23:28:23 +03:00
-
5c03866a17
fix: materialize meta buffers after assign load (NotImplementedError on unpatch)
WildAi
2026-08-19 23:28:23 +03:00
-
788f0bf705
perf: meta-init + assign loading, non-destructive offload contract (v2.3.0)
WildAi
2026-08-19 18:56:35 +03:00
-
7139ca22c6
perf: meta-init + assign loading, non-destructive offload contract (v2.3.0)
WildAi
2026-08-19 18:56:35 +03:00
-
002a7910fd
fix: guard empty generation output + warn on low-bit/naive quantization (v2.2.3)
WildAi
2026-08-18 16:15:39 +03:00
-
5cafb2333f
fix: guard empty generation output + warn on low-bit/naive quantization (v2.2.3)
WildAi
2026-08-18 16:15:39 +03:00
-
cdaa476670
fix: detect linked external_model by presence during validation (v2.2.2)
WildAi
2026-08-17 02:22:44 +03:00
-
86be533986
fix: detect linked external_model by presence during validation (v2.2.2)
WildAi
2026-08-17 02:22:44 +03:00
-
7dd6662038
fix: list and load .gguf files in Load VibeVoice Model node (v2.2.1)
WildAi
2026-08-17 01:29:12 +03:00
-
a6fb400162
fix: list and load .gguf files in Load VibeVoice Model node (v2.2.1)
WildAi
2026-08-17 01:29:12 +03:00
-
57d82969bc
feat: external model input via VibeVoiceLoadExternalModel node (v2.2.0)
WildAi
2026-08-16 23:59:04 +03:00
-
e20b4fd9d8
feat: external model input via VibeVoiceLoadExternalModel node (v2.2.0)
WildAi
2026-08-16 23:59:04 +03:00
-
1e9ac4eb42
fix: reject streaming models in standard TTS node (TypeError on realtime model)
WildAi
2026-08-16 19:44:57 +03:00
-
1aba31b0a1
fix: reject streaming models in standard TTS node (TypeError on realtime model)
WildAi
2026-08-16 19:44:57 +03:00
-
b4fce48cff
feat: report standard ComfyUI progress during TTS/ASR inference (v2.1.1)
WildAi
2026-08-16 02:24:52 +03:00
-
da936e6719
feat: report standard ComfyUI progress during TTS/ASR inference (v2.1.1)
WildAi
2026-08-16 02:24:52 +03:00
-
b2feff00d4
feat: torchaudio-primary audio backend, librosa now optional (v2.1.0)
WildAi
2026-08-15 21:54:53 +03:00
-
43fb085228
feat: torchaudio-primary audio backend, librosa now optional (v2.1.0)
WildAi
2026-08-15 21:54:53 +03:00
-
ee92424b44
fix: cpu-first loading + neg-branch rope fix
WildAi
2026-08-15 16:46:56 +03:00
-
dc64c17c00
fix: cpu-first loading + neg-branch rope fix
WildAi
2026-08-15 16:46:56 +03:00
-
1cc3f347a1
sound fix
WildAi
2026-08-14 20:49:06 +03:00
-
37710ce609
sound fix
WildAi
2026-08-14 20:49:06 +03:00
-
19d03648e5
refactor init
WildAi
2026-08-14 20:26:31 +03:00
-
851647983b
refactor init
WildAi
2026-08-14 20:26:31 +03:00
-
99a98031d7
v1.5.1
1.5.1
WildAi
2025-09-25 13:16:21 +03:00
-
649b5cf341
v1.5.1
WildAi
2025-09-25 13:16:21 +03:00
-
1ee7d7cbad
fix tokenizer.json issue, fix num_hidden_layers
WildAi
2025-09-25 13:15:07 +03:00
-
68ccb138b1
fix tokenizer.json issue, fix num_hidden_layers
WildAi
2025-09-25 13:15:07 +03:00
-
9c750be007
v1.5.0
1.5.0
WildAi
2025-09-24 18:34:42 +03:00
-
ece3d645c7
v1.5.0
WildAi
2025-09-24 18:34:42 +03:00
-
e6d8dce54a
Update README.md
WildAi
2025-09-24 18:31:25 +03:00
-
f8a2c6af79
Update README.md
WildAi
2025-09-24 18:31:25 +03:00
-
7ec3a8a378
Update README.md
WildAi
2025-09-24 17:51:40 +03:00
-
c0ce4d7360
Update README.md
WildAi
2025-09-24 17:51:40 +03:00
-
696ef69152
voice bleeding fix, audio quality, input speakers tags, zero-shot voices
WildAi
2025-09-24 17:42:30 +03:00
-
62912003f4
voice bleeding fix, audio quality, input speakers tags, zero-shot voices
WildAi
2025-09-24 17:42:30 +03:00
-
d04665d073
change the node default params
WildAi
2025-09-10 13:50:47 +03:00
-
0403bc6b31
change the node default params
WildAi
2025-09-10 13:50:47 +03:00
-
85edc3bf81
v1.4.1
WildAi
2025-09-10 13:00:32 +03:00
-
260f80ad9c
v1.4.1
WildAi
2025-09-10 13:00:32 +03:00
-
51c1ccff75
Update requirements.txt
WildAi
2025-09-10 12:59:34 +03:00
-
6579f2abfd
Update requirements.txt
WildAi
2025-09-10 12:59:34 +03:00
-
5dd167a1b5
ver 1.4.0
1.4.0
WildAi
2025-09-10 12:41:01 +03:00
-
9be666e5a0
ver 1.4.0
WildAi
2025-09-10 12:41:01 +03:00
-
4296f75cc4
fix model link
WildAi
2025-09-10 12:33:40 +03:00
-
4747b9434c
fix model link
WildAi
2025-09-10 12:33:40 +03:00
-
4c9785da8b
major refactoring
WildAi
2025-09-10 12:06:26 +03:00
-
ef1b65b9fa
major refactoring
WildAi
2025-09-10 12:06:26 +03:00
-
00e3476f4d
Merge pull request #43 from RodriMora/patch-1
WildAi
2025-09-10 11:19:52 +03:00
-
-
712315d963
Merge pull request #43 from RodriMora/patch-1
WildAi
2025-09-10 11:19:52 +03:00
-
-
5f7eb9b57e
Update vibevoice_nodes.py
RodriMora
2025-09-08 17:07:32 +02:00
-
-
2034c04bb2
Update vibevoice_nodes.py
RodriMora
2025-09-08 17:07:32 +02:00
-
-
06f72b88a2
Update pyproject.toml
WildAi
2025-09-07 19:42:46 +03:00
-
2bdcf9679b
Update pyproject.toml
WildAi
2025-09-07 19:42:46 +03:00
-
3e2a5b272c
Merge pull request #35 from Jarvik7/patch-1
WildAi
2025-09-07 19:28:25 +03:00
-
-
879293fc30
Merge pull request #35 from Jarvik7/patch-1
WildAi
2025-09-07 19:28:25 +03:00
-
-
dce37a234c
Update sage_attention_patch.py
Jarvik7
2025-09-05 10:19:13 -03:00
-