Alisson AnjosandClaude Opus 5.5 04d8012020 Panel buttons run models under inference_mode like a node; bump to 1.118.1 ...

Describe refs / Analyse / Auto prompt / target / mask preview ran the VLM and SAM in a server thread
outside torch.inference_mode(). After a workflow run those models hold inference tensors, so the
buttons failed with "Inference tensors do not track version counter". Fixes #14.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

2026-11-07 18:27:25 +00:00

..

fixtures

fix: recognize ComfyUI file-loaded sol-attn module names

2026-10-23 05:30:34 +00:00

test_bfs_h3_history.py

Shot Planner: continuity 'history' - the last 17/34/51 frames of the previous result placed just before frame 0 on H3's timeline (sliding history, after Akatz's H3 Relay / Ethanfel's Context Loop), with every mask mode, crop and the duet canvas; cuts reset it

2026-10-09 01:37:36 +00:00

test_bfs_h3_native_shot_loop.py

Add seconds or frames selection to native H3 shot loop

2026-10-24 14:39:40 +00:00

test_bfs_h3_native_workflow.py

Add head swap mode to native H3 prompt builder

2026-10-24 14:56:37 +00:00

test_bfs_h3_side_panel.py

Duet prompts: TSC's newer strategy - the finished split screen as plain footage (no kept footage / generated / in sync), the pinned region named by place + one clause of what is seen ({panel_region}/{video_region}: top quarter, bottom three quarters...), motion beat by beat with times on the video clock, sounds of the new scene

2026-10-08 21:18:07 +00:00

test_bfs_segment_picker.py

BFS Segment Picker: brush, remove points and eraser; bump to 1.118.0

2026-11-04 20:26:38 +00:00

test_bfs_shot_loop.py

Panel buttons run models under inference_mode like a node; bump to 1.118.1

2026-11-07 18:27:25 +00:00

test_bfs_source_blur.py

Source blur: choose what stays sharp, masked area, Sapiens2 segmenter, invert colours (1.110.0)

2026-10-17 12:02:24 +00:00

test_bfs_video_compare.py

BFS Video Compare: download the frame as PNG; single images save as PNG; bump to 1.117.0

2026-11-03 12:05:58 +00:00

test_h3_cast_and_token_controls.py

Add clickable SAM 3.1 cast bindings and experimental H3 token weight editor

2026-10-24 11:21:49 +00:00

test_h3_inline_prompt_weights.py

Support inline phrase weights and make token inspection stops explicit

2026-10-24 12:42:37 +00:00

test_headswap_auto_config.py

Shape the crop to the latent's aspect, not the latent to the crop (1.41.1)

2026-09-03 17:08:17 +00:00

test_minimax_h3_downscaled_guide.py

Add aligned downscaled MiniMax H3 guides and image/video targets

2026-09-30 17:26:56 +00:00

test_minimax_h3_native_temporal.py

Add optional structure guidance to TARA

2026-10-25 13:29:44 +00:00

test_minimax_h3_reference_binding.py

Add clickable SAM 3.1 cast bindings and experimental H3 token weight editor

2026-10-24 11:21:49 +00:00

test_minimax_h3_reference_rope.py

Add H3 reference RoPE options and caption-only identity conditioning

2026-09-30 20:30:01 +00:00

test_minimax_h3_temporal_bootstrap.py

Select verified temporal kernel prebuilts by Windows/Linux and GPU architecture

2026-10-23 22:15:46 +00:00

test_minimax_h3_temporal_fast.py

Add compile-free FlashAttention temporal backend and guide scaling (1.115.11)

2026-10-23 00:02:49 +00:00

test_minimax_h3_temporal_kernel.py

feat: automatically prepare verified temporal kernels on first use

2026-10-23 07:36:01 +00:00

test_minimax_h3_temporal_kitchen.py

feat: detect Model Attention Backend and add experimental Kitchen temporal bias

2026-10-23 04:44:40 +00:00

test_minimax_h3_temporal_lock.py

Select verified temporal kernel prebuilts by Windows/Linux and GPU architecture

2026-10-23 22:15:46 +00:00

test_minimax_h3_temporal_sol.py

fix: recognize ComfyUI file-loaded sol-attn module names

2026-10-23 05:30:34 +00:00

test_reference_temporal_offset.py

Use standard zero reference offset by default

2026-08-14 10:48:37 +00:00

test_tass_layout.py

Do the layout maths in float, not in the coordinate dtype (1.29.3)

2026-08-23 17:01:25 +00:00