issues Search Results · language:Dune language:Python language:Python language:JavaScript language:CSS language:JavaScript
Filter by
42.4M results
Capability — what the system needs from this
The realtime 0.6B TTS must meet its latency budget for live meetings.
Root cause (first principles)
TTFB/chunk latency claimed but not measured by a repeatable ...
area:meeting
area:testing
area:voice
fsv
ChrisRoyse
Capability — what the system needs from this
Ingest should extract prosody (F0, energy, rate, pitch contour) from separated vocal stems so voice cloning has grounded
prosodic references.
Root cause (first ...
area:pipeline
area:voice
fsv
Capability — what the system needs from this
speak_optimized must actually pick and persist the highest-SECS candidate.
Root cause (first principles)
Best-of-N ranking exists but isn t FSV-verified ...
area:testing
area:voice
fsv
Capability — what the system needs from this
Cloned voice must match the real person at SECS 0.95 with full prosody; this is a hard quality floor, no trade-offs for
speed.
Root cause (first principles) ...
area:voice
fsv
Capability — what the system needs from this
The offline LatentSync path must produce the known-best quality and never re-introduce the blur-amplifying CodeFormer
step.
Root cause (first principles) ...
area:avatar
fsv
Capability — what the system needs from this
Generated clone frames must be the right person. Per prior finding, SigLIP global embeddings fail identity (an
interviewer scored 0.985 vs Santa); ArcFace ...
area:avatar
fsv
Capability — what the system needs from this
Clone video generation must run on any install, not only this machine.
Root cause (first principles)
src/clipcannon/tools/constellation.py:~140 hardcodes ...
area:avatar
fsv
root-cause
Capability — what the system needs from this
Either make EchoMimic build for Blackwell or formally migrate off it — but stop carrying a blocked training path.
Root cause (first principles)
mmcv (EchoMimic ...
area:avatar
area:gpu
blocker
nvidia-upgrade
root-cause
Capability — what the system needs from this
The meeting avatar s realtime lip-sync must either work or not exist — a method that lazy-loads then raises
MeetingLipSyncError is the worst state.
Root cause ...
area:avatar
area:meeting
blocker
fsv
root-cause
Capability — what the system needs from this
NVIDIA s diffusion-based Audio2Face-3D improves lip/tongue movement and is Blackwell-native (no mmcv). The repo already
has src/phoenix/adapters/audio2face_adapter.py ...
area:avatar
blocker
fsv
nvidia-upgrade
Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip! Restrict your search to the title by using the in:title qualifier.
Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip! Restrict your search to the title by using the in:title qualifier.