48108a356923ba4795b37c35cd4906a46abb8745
34
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
  bvandeusenandClaude Opus 5
|
693759f2bb |
fix: an idle GPU agent could not check in, so the roster called it stopped
CI and images / lint (push) Successful in 3s
CI and images / extension-version (push) Successful in 2s
CI and images / frontend-build (push) Successful in 21s
CI and images / backend-lint-and-test (push) Successful in 31s
CI and images / integration (push) Failing after 2m10s
CI and images / sign-extension (push) Skipped
CI and images / build-web (push) Skipped
CI and images / smoke-web (push) Skipped
CI and images / promote (push) Skipped
CI and images / build-agent (push) Skipped
Operator, 2026-09-23: *"I'm running the gpu agent on my device and it
currently reads as 'offline' but it's running and has checked in recently."*
It had checked in — twelve minutes ago. Two cadences that never agreed:
idle lease poll ceiling 900s agent/fc_agent/worker.py (sleep mode)
heartbeat while idle never gated on holding leases
roster "stopped" after 300s api/system_health.py
The roster records an agent check-in on `lease` and `heartbeat`. The heartbeat
loop was gated on `if ids:`, so an agent holding no leases sent nothing at
all — leaving the lease poll as the only check-in, and sleep mode backs that
off exponentially to a 900s ceiling. 900 against 300: an IDLE agent was
structurally guaranteed to read as stopped. Nothing was broken; nothing was
misconfigured; the two halves simply disagreed.
Not a recent regression. Sleep mode landed 2026-07-02; the roster adopted the
lease as its check-in on 2026-09-02 — *"A lease IS the check-in … Recorded on
the call that was already happening"* — without noticing that the call it was
piggybacking on had been deliberately slowed ten weeks earlier.
The heartbeat now sends whether or not it holds leases. An empty one extends
nothing (`id.in_([])` matches no rows) and costs one small POST every 45s —
against the 6/min lease poll sleep mode exists to avoid, that is not a cadence
worth protecting, and it is what makes "is the agent alive" answerable at all.
Still gated on `self._running`: a worker that has been stopped is not checking
in for work, and reporting it as present would be a different lie.
Two things I could NOT determine from the code, both needing the live table:
whether a stale `agent:agent` row exists from an older build that omitted
`agent_id` (the server defaults it), and whether changing `AGENT_ID` has ever
stranded an abandoned row — nothing prunes `service_seen`, so either would sit
there reading "stopped" forever.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LVjrnpQjRgHdvq95rASoiR
|
||
|
bvandeusenandClaude Opus 4.8
|
57e52433d0 |
feat(agent): idle-unload GPU models to free VRAM when the queue is idle
The SigLIP embedder + YOLO proposers load lazily then stay resident for the container's whole lifetime — a 24/7 agent with an empty queue squats on ~5GB of VRAM doing nothing (operator-observed: 4900MiB held at GPU-util 8% / P8). Sleep mode only sheds downloaders + poll cadence; even a UI Stop left the models loaded. Add a monitor thread that unloads the torch-owned models after cfg.idle_unload_seconds (env IDLE_UNLOAD_SECONDS, default 300; 0 disables) with the GPU genuinely idle (active==0, buffer drained, no job completed in the window), then torch.cuda.empty_cache() to hand the blocks back to the driver. They reload lazily on the next job via the existing _ensure_embedder / _proposers_for. Covers both sleep-mode idle and a full Stop. Surfaced in /status (models_loaded) and the agent UI pipe line; the VRAM meter drops too. Residual: imgutils CCIP/person ONNX sessions + the CUDA context stay resident (no clean unload API) — idle VRAM drops substantially, not to zero. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01TbrA36zNczjVhrM6cWThQa |
||
|
bvandeusenandClaude Opus 4.8
|
a4df279343 |
feat(ml): lease announces detector config; agent builds proposers from it live (#134 step 2)
The GPU lease now carries the crop-proposer config from MLSettings in a per-job 'detectors' block (same pattern as embed_model_name). The agent's worker builds its Proposers from the announced config via _effective_cfg (lease block overlaid on env) + _proposers_for (rebuilds only when a config signature changes) — so an operator's UI edit takes effect on the next lease with NO restart, and env is now just the bootstrap fallback until the server announces. enabled-off maps to empty weights (proposer skipped); dedupe_iou + max_regions also come from the effective cfg. Test: lease announces the detectors block with the seeded default weights. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CDgx8bQS5YrGRK76v8HUnM |
||
|
bvandeusenandClaude Fable 5
|
b1cfbcc06a |
fix(agent): sleep mode — an empty queue sheds to one downloader and backs the lease poll off to 15 min
Operator-flagged on the deployed .5 build: the autoscaler grew the pool 1→8 against an EMPTY queue (an empty buffer read as 'GPU starving' regardless of WHY), and every downloader kept polling lease every 10s all night. - New idle signal straight from the lease results: an empty lease sets _idle, any jobs clear it. The occupancy-low branch now distinguishes three cases: queue empty → shed to ONE polling downloader; pinned at the bandwidth cap → shed toward 3; cap headroom + work flowing → grow. - Idle lease polls back off exponentially per downloader to IDLE_POLL_MAX_SECONDS (15 min) and reset the moment work appears — so an idle night costs one HTTP call per 15 min, and new work is noticed within at most ~15 min (operator-accepted trade-off). - UI hint: 'idle — queue empty, lease poll backed off'; /status gains idle. Agent build 2026-07-02.6. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CDgx8bQS5YrGRK76v8HUnM |
||
|
bvandeusenandClaude Fable 5
|
7c19ad91ed |
feat: cap-aware autoscaler + token-gated whole-instance tag reset (operator feedback)
Autoscaler (agent 2026-07-02.5): the buffer-occupancy signal alone would peg downloaders at DL_MAX while the bandwidth CAP — not concurrency — is the real constraint (8 streams sharing 8 MB/s move no more data than 4). Growth is now gated on the pipe having headroom (net < 85% of cap) and a pipe pinned at the cap (>= 95%) sheds streams down to 3; dead band prevents flapping. The UI hint says 'holding at the bandwidth cap' and /status reports bw_capped, so the behavior is legible without tests that need the ML stack. Reset content tagging: stays a FULL-instance reset (operator's call), but now lives in a fenced 'Danger zone' section on Cleanup and the apply is gated by a preview-derived confirm token (mirrors the Tier-C bulk-delete pattern — stale counts are rejected server-side). Copy no longer claims suggestions repopulate: it says plainly the heads' training examples are deleted and re-tagging starts fresh. Moved out of TagMaintenanceCard into DangerZoneCard. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CDgx8bQS5YrGRK76v8HUnM |
||
|
bvandeusenandClaude Fable 5
|
95d2ae1d58 |
feat(agent): global bandwidth cap — the agent can't saturate the desktop's network
One shared TokenBucket (default 8 MB/s; BANDWIDTH_LIMIT_MB_S, 0 = unlimited; live MB/s dial + net readout in the control UI) is charged by every still download (streamed chunk reads) and every ffmpeg video stream (metered from outside via /proc/<pid>/io and SIGSTOP/SIGCONTed into budget). Why: D1 re-measurement 2026-07-02 — the idle link moves ~38 MB/s, but 8 unthrottled downloaders bufferbloated it to ~1-1.5 MB/s PER STREAM (operator's browser included). Capping the aggregate keeps the desktop usable and still beats the collapsed sweep throughput it replaces. Agent build 2026-07-02.4. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01CDgx8bQS5YrGRK76v8HUnM |
||
|
bvandeusenandClaude Fable 5
|
09e2772628 |
fix(gpu-jobs): end the error-tombstone loop — deliberate retry semantics + poison-job guards
The hourly ccip backfill's skip-list lacked 'error' (and the daily
siglip/embed variants re-gated failures on their missing results), so every
permanently-bad file got a fresh doomed job each run — ~24 duplicate error
rows/day per file, the perpetual 'unprocessable' flood. An errored job is now
a TOMBSTONE: no backfill re-enqueues it; retry is deliberate-only via
/retry_errors (an errored back-catalogue needs one button press after a
model swap).
One shared set of dedupe DELETEs (services/ml/gpu_jobs.error_dedupe_statements)
runs before every backfill and inside /retry_errors: error rows made moot by a
later pending/leased/done row go first, then older duplicates (newest reason
survives) — so the error count reads as distinct failing files and a retry
can't fan one file out into duplicate pending jobs. /retry_errors now returns
{requeued, pruned} and the toast shows both.
Poison-loop guards (release and lease-expiry burn no attempts, so a job that
stalls its transfer or crashes the agent every time cycled forever —
operator-observed jobs 99044/125288/131594/143131):
- agent: 3 in-session transient bounces (fetch or submit) → fail with the real
reason instead of another release; strikes never count while stopping, and
clear on submit success. Agent build 2026-07-02.3.
- server: the 60s orphan sweep (statements shared between the beat task and
GpuJobService so they can't drift) converts expired leases with >=5 lease
grants and pending jobs with >=10 to 'error', preserving the last stored
failure reason. Backstops old agent builds.
Tests: tombstone rule across all three backfill variants, moot-row pruning,
poison conversions, and the extended /retry_errors dedupe contract.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CDgx8bQS5YrGRK76v8HUnM
|
||
|
bvandeusenandClaude Fable 5
|
1b1d3732dc |
feat(agent): store ffmpeg's actual failure reason in the job's error field
sample_frames_from_url now returns (frames, reason) — reason carries the SPECIFIC cause on failure (ffmpeg's stderr tail, e.g. "moov atom not found", or the timeout) instead of only logging it agent-side. The worker folds it into the failure it reports, so curator's GpuJob.error reads e.g. no frames sampled from video — ffmpeg exit 183: moov atom not found ... instead of the bare "(unprocessable)". The errored-jobs list becomes self-describing: after a retry sweep, surviving errors name their real defect without needing the agent log. Return-value plumbing (not shared state) so concurrent downloaders stay isolated. Agent VERSION → 2026-07-02.2. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> |
||
|
bvandeusenandClaude Opus 4.8
|
713a11e394 |
fix(agent): server-side rate metrics + killable-on-stop ffmpeg
Two follow-ups from live debugging of "work/min never populates" and "stopped never reached". 1) jobs/min + downloads/min are now computed in the BACKEND on a fixed cadence (_rate_loop, EWMA) and reported ready-to-show. The rates were derived client-side from poll deltas with a dt<30s guard — but a backgrounded/unfocused browser tab throttles its timers to ~1/min, so every delta exceeded 30s and the guard blanked the rates forever. A server-side rate is independent of how often the tab polls. Frontend just displays s.jobs_per_min / s.downloads_per_min. VERSION → .9. 2) ffmpeg video sampling is now killable on Stop. A downloader stuck in a slow/reconnecting decode (observed: 47s, 230s for one video) couldn't see the stop signal until ffmpeg returned, so Stop detached still-running threads and work kept flowing long after — "stopped" that wasn't really stopped. sample_frames_from_url now runs ffmpeg via Popen and polls a `should_stop` callback every 0.5s, terminating (then killing) the process at once on Stop or the per-video timeout. A stop-killed job is handed back (transient), not failed. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
bvandeusenandClaude Opus 4.8
|
6282e753a9 |
fix(agent): real start/stop state machine — kill the stuck "stopping" pill
The Status pill hung on "stopping" forever (operator-flagged 2026-07-01). Root cause: the backend had no lifecycle state — status() only returned running/stopped — so the UI FABRICATED "stopping" in JS as `!running && active>0`. That pill only cleared when the backend's `active` counter hit 0, but stop() (a) blocked the HTTP handler on lease-release calls to curator and (b) left `active>0` whenever a consumer wedged mid-submit/release to an overloaded curator → "stopping" that never resolved. Give the backend a real, truthful state it drives itself: stopped → starting → running → stopping → stopped - start(): → starting; a downloader flips it to running on its FIRST successful lease (so "running" means curator is actually answering, not just "Start was clicked"). If curator's down it honestly stays "starting". - stop(): → stopping; returns immediately (no handler block). A background monitor waits for the worker threads to actually exit, releases leases, then → stopped — bounded by STOPPING_TIMEOUT (20s) so a wedged submit can NEVER hold the UI in "stopping" again. In-flight work is handed back safely. - Buttons follow the real state (Start only from stopped; both disabled through the transition), so you can't fight a transition. - Log every Start/Stop button press (routes) and every transition (worker), so the Logs panel shows exactly what each button did. Frontend now trusts s.state (drops the active>0 hack); VERSION → .8. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
bvandeusenandClaude Opus 4.8
|
91ea06be79 |
feat(agent): Status shows smoothed jobs/min + downloads/min (replace jumpy gauges)
The buffer / on-GPU / downloader counts flip many times a second, so a 3s
status poll only ever samples noise — the tiles looked frozen (same value
twice) or random (wildly different), reading as "the Status section doesn't
update" when the backend was in fact live (operator-flagged 2026-07-01).
Replace the three instantaneous gauge tiles with two derived RATE tiles:
- jobs / min — GPU throughput, from the monotonic `processed` counter
- downloads / min — fetch throughput, from a new monotonic `downloaded`
counter (bumped when a job is decoded into the buffer)
Together they also show pipeline balance (dl/min > j/min ⇒ GPU-bound; the
reverse ⇒ GPU starved). Both are EWMA-smoothed over the poll deltas, clamped
at 0 (agent restart resets the counters), and skip a backgrounded-tab gap.
The still-useful instantaneous state is demoted, not lost: buffer stays as
the occupancy bar; downloaders/consumers/on-GPU move to the sub-line. `waited
out` (transient) gets promoted to a tile.
backend: worker.status() gains `downloaded`; `_bump(downloaded=)`.
frontend: retiled Status + rate math in applyStatus; VERSION → .7.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
||
|
bvandeusenandClaude Opus 4.8
|
98b2ac90dd |
refactor(agent): DRY pass on the GPU agent worker package
Consolidate genuine duplication in agent/fc_agent into single-source helpers (behavior-preserving; DRY Pass process #594): worker.py - _fail(jid, image_id, exc, verb) — 4 terminal "fail this job" blocks (downloader HTTP-fault + decode, consumer non-transient + generic). - _release(job_ids) (was _release_owned) — the one lease hand-back path; 6 inline release([jid])+unhold sites now route through it. - _stopped(stop_evt) + _abort_if_stopped(jid, stop_evt) — 4 stop-check -and-release blocks and every bare stop-check. - _timed(stage) contextmanager — ~8 monotonic()/_record() timing pairs; records only on clean exit, matching the old skip-on-raise behavior. - _ewma(prev, x, alpha) module fn — 3 EWMA updates in the autoscaler. client.py - _submit(path, payload) — submit / submit_embedding (retrying session). - _post_quiet(path, payload) — heartbeat / fail / release fire-and-forget. detectors.py - Proposers._top(detector, image, cap) — merges components() and panels(). config.py - _bool_env(name, default) — auto_start / auto_scale env parsing. Left alone (recorded): the xyxy→norm-xywh conversion duplicated across models.py/detectors.py (2 copies, independent wrapper modules — sharing would couple them), and the _ensure_embedder/_ensure_proposers pair (same lock shape, different concepts). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> |
||
|
bvandeusenandClaude Opus 4.8
|
8a0237eeea |
fix(agent): stream videos via ffmpeg-from-URL instead of downloading the whole file
The failing "poison" jobs were 800MB+ 4K VR videos: the agent pulled the ENTIRE file into memory (r.content) just to sample a few frames, which buffered ~1GB in RAM and — on any slow/contended media store — got cut off mid-download (ChunkedEncodingError), failed, and re-leased forever. Measured the media read at ~4–6 MB/s (raw off the share, curator out of the path), so no serving-layer tweak helps; the file simply shouldn't be fully downloaded. Environment-agnostic fix (works for any deployment, completes even when slow): - media.sample_frames_from_url(): point ffmpeg straight at curator's /images URL. It Range-reads only the video index + up to max_frames of content — never the whole file — and reconnect flags resume a dropped transfer instead of failing. Generous, env-tunable timeout (FFMPEG_TIMEOUT, default 1200s) = completion over speed. Removes the bytes-based sample_frames (dead once videos stream). - worker._download_decode: videos now stream (no fetch_image, no RAM blowup); stills still download+decode. On an ffmpeg miss, probe curator liveness (client.is_reachable) → fail the job if curator is up (unprocessable file, stops the infinite re-lease) vs release if curator is down (transient, survives a redeploy). Auth header passed so it works whether or not /images is gated. Build marker 2026-07-01.6. Refs issue #1225. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
aa0605585b |
fix(agent): log the REAL fetch/submit failure reason, not "curator unreachable"
"curator unreachable" was printed for every transient error, hiding whether a single file's transfer stalled (ReadTimeout — curator is up, that stream is slow) or curator itself is down (ConnectTimeout/ConnectionError) or errored (HTTP 5xx). Those need completely different fixes, and we've been diagnosing the download slowness blind. Add _transient_reason(exc) → a specific label (HTTP <code>, else the exception class: ReadTimeout / ConnectTimeout / ConnectionError / …) and use it in both transient paths: - downloader: "fetch failed job <id> (image <id>, ReadTimeout) — released, backing off" - consumer: "submit failed job <id> (<reason>) — released, re-lease later" Now the logs say which failure it actually is (and which image), so we can tell a slow/stalled transfer apart from an unreachable curator. Build marker 2026-07-01.5. Refs issue #1225. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
ccbb5cbc9e |
fix(agent): stop the downloader pool stampeding a slow curator (congestion collapse)
Operator hit an outage after the machine slept overnight: the agent showed "curator unreachable" in a loop while curator's API (lease) was actually fine and the browser could still load images — just slowly. Root cause is a feedback loop in the new pipeline: every download streams a full original through curator's single Python file-serving path, and the autoscaler grows DOWNLOADERS whenever the buffer is empty. When downloads are merely SLOW/failing, the buffer is empty for that reason — so the agent piled on more concurrent large-file GETs, saturating curator's web workers + NFS, which slowed curator (and its browser) further and produced more failures → more downloaders. Classic congestion collapse. - Failure-aware autoscaling: if transient download failures rose since the last decision, SHRINK the downloader pool toward the floor instead of growing — the empty buffer is caused by failures, not the GPU starving. It ramps back up only once downloads succeed again. - DL_MAX 24 → 8: 24 concurrent large-file downloads through one Python serving path is too many; 8 keeps a fast GPU fed without stampeding curator. - fetch_image timeout 180 → (10, 60): the read timeout is between-bytes, so a large-but-flowing download still completes, but a stuck/dead connection fails in 60s instead of hanging a downloader for 3 min and piling up stuck requests. Build marker 2026-07-01.4. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
7cdce0c474 |
feat(agent): temporal video dedup — drop near-duplicate frames before the GPU
Near-static videos are the dominant GPU load: sampled into up to 64 frames, each re-runs the whole detect→CCIP→SigLIP chain on ~identical content. Add a CPU perceptual-hash frame dedup upstream of the GPU so the redundant frames are never processed at all (not just their embeds). - media.dedupe_frames() + _dhash(): 8×8 difference-hash (64-bit) per frame; greedy keep — a frame survives only if its hash differs from every kept frame by >= min_distance bits (Hamming). A static run collapses to one frame; genuinely distinct scenes all survive. Order + frame_time preserved. - Called in worker._download_decode right after sample_frames, so it runs in the decode stage on the downloader thread (CPU) — the GPU consumers only ever see deduped frames, and buffered video items shrink (less RAM too). - Env-tunable FRAME_DEDUPE_DISTANCE (default 8; higher keeps more frames for brief localized changes an 8×8 hash can miss; 0 disables). Logs `video frames N→M` when it drops any, so video load reduction is visible. Complements the spatial per-frame crop dedup (2026-07-01.2); this is the temporal axis. Build marker 2026-07-01.3. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
eaae896858 |
feat(agent): dedupe near-duplicate crops before the SigLIP embed
Figure boxes are already NMS-merged (iou 0.6) and each YOLO detector self-NMSes, but the combined per-frame crop pile (figure→concept ∪ anatomy component→concept ∪ panel) was embedded with no cross-proposer dedup — so genuine near-duplicates slipped through (a figure box ≈ an anatomy component on a solo bust; overlapping booru head classes on one head), embedding the same region twice and burning a slot against max_regions. Add detectors.dedupe_crops(): a greedy, high-IoU (default 0.85), kind-aware pass over the pending (crop, template) list right before embed_batch — drop boxes that overlap ≥ iou within the same kind, keep the highest score. The high threshold is deliberate: it collapses only true near-identical boxes while preserving intentional nested crops across scopes (a whole figure vs a small head component sit well below it) and distinct kinds (concept vs panel). Env-tunable DEDUPE_IOU (≥1.0 disables). Runs on CPU before the GPU work, so it cuts both embed cost and region count. Temporal (cross-frame) dedup deferred. Build marker 2026-07-01.2. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
afef95a87d |
feat(agent): download/GPU producer-consumer pipeline + fix detector fuse crash
The agent workload is download-bound (download 400–5462ms vs GPU ~300–600ms), so the old N-slot serial chain (each slot: lease→download→decode→GPU→submit) left the fast GPU idle during every download. Rearchitect worker.py into a producer/consumer pipeline: downloader pool (autoscaled by BUFFER OCCUPANCY) → bounded queue → 1–2 GPU consumers (detect+embed→submit) - Downloaders are I/O-bound → many overlap; the autoscaler now tunes DOWNLOADER count by buffer fill (empty = GPU starving → add; full = outpacing GPU → add a 2nd consumer if it has util/VRAM headroom and lifts throughput, else trim). - Bounded buffer (12) = backpressure: a full buffer blocks downloaders, capping RAM + lease look-ahead. VRAM pressure sheds a consumer immediately. - Heartbeat thread keeps every held lease alive (buffered jobs wait on the GPU; curator's 180s TTL would otherwise reclaim them mid-buffer). - Preserves all resilience: lease exp-backoff, submit-path retry (#169), release-on-stop, region caps + video early-exit (#171). Stop drains BOTH pools and releases every held lease at once (single held-set as source of truth). - Consumers SHARE one embedder + proposers instance (a 2nd consumer adds concurrent inference, not N× VRAM — bounds the VRAM creep seen with N slots). - UI reworked for the pipeline: tiles show downloaders · buffer · on-GPU · processed · errors, a buffer-occupancy meter, and a consumers/waited-out line; the dial now tunes downloaders. Build marker 2026-07-01.1. Also fix the operator-flagged detector warning: yolo11n + the comic-panel model threw "'Conv' object has no attribute 'bn'" on every image (ultralytics' load- time Conv+BN fusion on a version-mismatched graph), silently disabling 2 of 3 crop proposers and spamming the log per image. Disable that fusion (unfused inference is correct, marginally slower) and permanently self-disable a proposer on the first inference failure instead of re-throwing forever. Refs milestone 122. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
83f1070a11 |
fix(agent): bound video GPU work — early-exit the frame loop at max_regions
Image 81602 turned out to be a 156 MB mp4, not a huge still: the agent samples up to 64 frames × ~32 regions/frame → ~2000 regions (the 413) and 64 frames of detect+CCIP+embed (the 38s). The MAX_REGIONS backstop (#171) only truncated the SUBMIT — the GPU work was already spent. Break out of the frame loop once accumulated regions reach max_regions, so a long video costs ~a few frames of GPU (~2-3s), not all 64 (~38s). The whole-image 'embed' task is unaffected (it mean-pools all frames and returns before this loop). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
c587ac667c |
fix(agent): cap figures + global region cap + reset active on stop
Three safety/robustness fixes from the operator's run logs: - Cap figures per frame (MAX_FIGURES, default 8) like components/panels already are. Uncapped, a huge/busy image yielded hundreds of figure boxes → hundreds of per-figure CCIP calls + crops → a 38s job AND a submit too big to accept (image 81602 looped on 413). This is the acute fix. - Global per-JOB backstop (MAX_REGIONS, default 128): if total regions still exceed the cap (long video), keep the highest-scoring and log the drop, so a submit body can never blow past curator's limit. - Stale "active" meter: stop() now resets _active to 0 (no slots remain, so the meter must read 0 at once), and _bump clamps at 0 so a slot finishing after the reset can't drive it negative. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
79269da802 |
fix(agent): prompt stop + lazy curator polling + build marker; add agent to CI
Addresses operator reports: Stop never finishes, the agent polls curator constantly, and stale-cached pages get mistaken for a failed deploy. - Stop is prompt: flip _running BEFORE any lock so /status + worker loops see "stopped" immediately, and add a stop/shrink checkpoint in _process (after decode, before the expensive detect+embed) that releases the job and bails — so a Stop doesn't wait out heavy GPU work. - Lazy curator polling: the queue snapshot is fetched only while a browser is actually watching (a /status hit within UI_IDLE_GRACE) and on a 5s cadence, not a constant background loop. The work loop's own lease/submit is curator's only visitor otherwise — nothing polls just to poll. - Build marker: VERSION is embedded in the page and reported on /status; the UI shows a "reload" banner when they differ, so a browser-cached page can't be mistaken for "the new image didn't deploy" (complements the no-store header). CI: the lint lane now also `ruff check`s agent/ and compileall-parses it, so the GPU agent is linted + syntax-checked before its image builds (build.yml only `docker build`s it). Fixed the agent's pre-existing UP037/B905 so it passes. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
e6a7fe7d03 |
feat(agent): per-stage timing breakdown (lease/download/decode/gpu/submit)
Instrument the job pipeline so we can see where wall-clock actually goes and decide — on data, not theory — whether a download/compute split is worth building. Each stage is timed per job and a rolling breakdown is logged every 30s to the agent console, e.g.: timing/30s — lease 8ms · download 310ms · decode 40ms · gpu 165ms · submit 70ms | wall/job 585ms (214 jobs) - lease timed around client.lease() in the slot loop (per batch). - download = fetch_image; decode = image/frame decode; gpu = detect + CCIP + batched embed; submit = the results POST. One-time model load is excluded from the gpu figure. - Thread-safe accumulator (stage -> [sum, count]) summarised + reset by a small daemon reporter thread; logs only when there was work. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
f0f031782d |
fix(agent): unfreeze status view + smoothed throughput-aware autoscaler + log pane
Operator: the status tiles (state/active/processed) and the Start/Stop buttons freeze while the GPU meters stay live. Root cause: /status made an INLINE blocking curator call (queue_status) on every poll, and with curator buried under a 112k-job backlog that call stalled — freezing the whole status refresh (the GPU bars survived because /gpu is a lock-free local read). Made worse by the old util-band autoscaler, which grew workers toward the 32 cap forever because util plateaus ~50% on this IO-bound load and never hit the 70 grow threshold — piling load onto curator and the agent process. - /status is now a pure in-memory read: worker.status() is lock-free, and the curator queue snapshot is refreshed by a background poller (never inline). - Autoscaler replaced with a smoothed, throughput-aware climb that SETTLES: samples util every 2s and EWMA-smooths it (raw util swings 0↔99), then every ~24s grows by one only while each grow keeps lifting smoothed jobs/s; when a grow stops helping it backs off one and holds, re-probing occasionally. No runaway, no flopping. - GPU util bar now shows a smoothed value: the agent's own EWMA (util_smooth, exposed on /gpu) when running, else smoothed client-side — so it glides instead of bouncing 0↔99. - act() aborts a slow Start/Stop POST after 8s so the buttons can't stick; the now-always-fast /status refresh recovers state regardless. - Log pane: bound the page to the viewport (height:100vh) so the Logs card scrolls INTERNALLY instead of overflowing off-screen; cap the ring buffer at 400 lines. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
3b34230fbd |
fix(agent): stable util-band autoscaler + live GPU meters
Two operator-reported issues with the GPU agent: 1. Worker count flopped almost every cycle, spiking the GPU. The hill-climb probed +1, judged it over a too-short noisy throughput window, saw no clear gain and reverted -1 — every tick. Replace it with a GPU-utilization-band controller: HOLD while smoothed util sits in a healthy band, grow only on clear spare capacity (util below the low mark + VRAM headroom), shrink under saturation or memory pressure. Util is EWMA-smoothed and decisions are spaced (DECIDE_EVERY samples), so a noisy nvidia-smi reading can't move the pool. Load stays consistent instead of probe/reverting. 2. GPU util/VRAM bars only updated on manual refresh. They rode the /status poll, which blocks on the curator queue call (slow when curator is busy), so the meters froze between refreshes. Give them a dedicated /gpu endpoint (local nvidia-smi only, no curator round-trip) polled every 1.5s, and drop the curator queue-status timeout 15s -> 5s so /status itself stays snappy. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
2713c3f773 |
perf(agent): batch SigLIP crop embeds per image + load truncated images
Two issues surfaced by the live logs (GPU pegged at ~0% util, 0.5 jobs/s,
truncated-image failures):
- BATCH the SigLIP embeds: collect all of an image's crops (figure + booru_yolo
components + panels) and embed them in ONE forward pass instead of one
forward+lock per crop. The per-crop path serialised every crop through the
inference lock and starved the GPU (≈0% util, autoscaler stuck oscillating);
batching gives a real GPU-bound workload + far higher throughput. CCIP still
runs per figure inline.
- LOAD_TRUNCATED_IMAGES in the agent (matches the server embedder): slightly-
truncated scraped images now load instead of failing the job 3× then erroring
("image file is truncated (N bytes not processed)").
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa
|
||
|
bvandeusenandClaude Opus 4.8
|
c1b099e5a3 |
feat(agent): in-UI log console + a real styling pass on the control page
- logbuf.py: bounded in-memory log ring buffer + a logging.Handler on the root logger; GET /logs serves it; the control page polls it into a console pane — so runs are monitorable without `docker logs`. worker now logs autoscale moves (one line per change, with jobs/s + util + VRAM) and job failures (job + image + reason); detectors already log load/disable. - Restyled the whole control page: a proper dark layout with a header + live connection pill, cards (Control / Status / Logs), a styled Auto switch + worker stepper, status tiles, separate GPU-util and VRAM meters, and the log console. No longer feels like an afterthought; all the existing control hooks are preserved. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
6d7b17b0b5 |
feat(agent): autoscale the worker count (throughput hill-climb), Auto default-on
The new per-job workload (3 detectors + several SigLIP embeds) is far more GPU-bound than the old I/O-bound CCIP pass, so the right worker count shifted and is hard to guess. Add an Auto mode (default ON) that finds it: - _control_loop samples jobs/sec + GPU util/VRAM every ~6s and hill-climbs the target: grow while throughput keeps improving and VRAM stays under budget, revert a step that doesn't help, back off under memory pressure (VRAM >= 90%), then settle and periodically re-probe (the GPU/IO balance shifts over a run). - A manual concurrency set is an override → leaves Auto; an "Auto" toggle in the control UI re-enables it. status() reports `auto`; the dial reflects the auto-chosen count (read-only) while Auto is on. - AUTO_SCALE env (default on) + compose doc. Agent py-compiled (outside CI). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
d5f29f7056 |
feat(agent): crop proposers — booru_yolo anatomy + COCO person + comic panels (#1202)
Better region PROPOSERS feeding the existing crop→SigLIP→max-over-bag heads (no change to the learned-tagging approach; no per-tag cost — propose once, embed each region, all heads in one matmul). - detectors.py: lazy ultralytics YOLO wrapper, each proposer independently optional + guarded (a bad weight spec / inference error self-disables that one, logged, never breaks the worker). Weights resolve from an ultralytics name | http(s) URL | "hf_repo::file", cached under HF_HOME. NMS merge so a figure two detectors both find collapses to one crop. - worker: figure boxes = imgutils detect_person ∪ general COCO person (merged) → CCIP + concept (anime + Western/realistic coverage); booru_yolo anatomy components (head/cat-head/anatomy/…) → concept crops; comic panels → kind= 'panel' concept crops. Capped per frame (MAX_COMPONENTS/MAX_PANELS). - config + compose: PERSON_WEIGHTS (default yolo11n.pt, works OOB), ANATOMY_WEIGHTS + PANEL_WEIGHTS (operator sets booru_yolo URL + mosesb panel hf::file; empty = off). ultralytics added to requirements. - backend: image_region 'kind' doc notes 'panel'; no migration (free String, and the bag scorer keys on a non-null siglip_embedding, not the kind, so any SigLIP region joins the bag automatically). Agent is outside CI — py-compiled here; operator tests on the GPU and checks Western-vs-anime crop quality via /api/ccip observability. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
4daa3f2790 |
feat(ml): operator model swap — GPU re-embed + embedder as a setting (#1190)
Make the SigLIP embedder an operator choice (drop-in to SigLIP 2:
google/siglip2-so400m-patch16-512 is a verified 1152-d model at 512px → no
schema change, better small-cue fidelity). A swap = set model + re-embed +
retrain, all operator-driven; the GPU agent does the re-embed so it's fast.
- settings: embedder_model_name is now a setting (migration 0065) alongside the
existing embedder_model_version; both editable + validated (non-empty) in the
ml admin API. The server embedder loads by HF name (AutoImageProcessor/Model,
model-agnostic), preferring the pre-downloaded local dir for the default so
existing deploys don't re-download; rebuilds on a name change.
- agent: new 'embed' job = whole-image SigLIP embedding (mean-pool video frames)
under the lease-announced model → POST /jobs/submit_embedding writes
image_record.siglip_embedding + siglip_model_version. The lease now announces
the model FROM THE SETTING (not a constant).
- re-embed routing: enqueue_gpu_backfill('embed') selects unembedded + stale-
version images; 'siglip' now re-embeds concept crops whose version != current
(so a swap re-triggers crops, not just the never-embedded back-catalogue). The
CPU ml-worker backfill no longer re-embeds on a version mismatch (it can't
churn the library at 512px) — the GPU agent owns version re-embeds. Daily
'embed' + 'siglip' beats self-heal.
- scoring: score_image only bags embeddings in the CURRENT model's space (whole-
image gated by siglip_model_version, concept regions by embedding_version) so a
mid-swap stale vector isn't scored by new-space heads; legacy NULL = current.
- UI: GpuAgentCard "Embedding model (advanced)" — edit name/version, Save, and
"Re-embed library (GPU)" (queues embed + siglip); points at SigLIP 2.
Tests: lease announces model + submit_embedding round-trip; enqueue 'embed'
selects stale/unembedded; stale-version excluded from scoring; embedder model
settable + empty rejected; siglip gate updated to current-version concept.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa
|
||
|
bvandeusenandClaude Opus 4.8
|
55fa4656ff |
feat(agent): survive + auto-recover when curator is unreachable
For redeploying curator while away with nobody to restart the agent: - _process now distinguishes a TRANSPORT error (curator down/redeploying, 5xx, 401/403/408/409/429, or our lease reclaimed mid-flight) from a genuine job fault. On a transport error it hands the job back (best effort) and signals the loop to back off — instead of calling fail(), which would burn the job's server-side attempt budget (MAX_ATTEMPTS=3) and permanently error good jobs across a redeploy. Job-specific 4xx (404 image gone) still fail so they don't re-lease forever. - lease loop retries with capped exponential backoff (poll_idle → 60s) and resets on the first successful lease, so a long outage is gentle and recovery is automatic within ≤60s of curator returning. Sleeps are interruptible so Stop / pool-shrink stays responsive. - AUTO_START env (default on in compose) resumes the worker on container start, so a host reboot / crash-restart (restart: unless-stopped) self-heals with nobody at the desktop. - control UI shows a "waited out" counter + an "curator unreachable, holding work" banner so the recovering state reads as recovery, not failure. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
c6f38b0dac |
feat(tagging): SigLIP concept crops + max-over-bag scoring (#114)
Lift recall on small/local concepts (glasses, cum, stomach-bulge, xray, lactation) that the whole-image SigLIP vector washes out: the GPU agent now embeds figure crops with SigLIP too, stored as kind='concept' regions, and the suggestion rail scores each image as a BAG (whole-image + every concept crop), taking each head's MAX over the bag. The whole-image vector is always in the bag, so this can never score lower than before. Model-agnostic by construction: the server ANNOUNCES the embedding model (HF name + version) in the lease, so the agent loads whatever the heads were trained in and stays in lock-step — a model swap is a server setting + a re-embed migration, never an agent change. - agent: model-agnostic CropEmbedder (torch/transformers get_image_features, fp16 on CUDA, inference-locked); worker branches on job.task — 'ccip' emits figure(CCIP)+concept(SigLIP) in one pass, 'siglip' emits concept-only so the back-catalogue backfill never churns figure/CCIP regions; torch cu124 + transformers in the image. - server: lease announces embed_model_name/embed_version; score_image is max-over-bag (version-filtered region embeddings); enqueue_gpu_backfill 'siglip' gates on a missing concept region (drains the back-catalogue, retries failures, no double-enqueue); daily siglip-backfill beat; UI button; /api/ccip/overview reports images_with_concept_siglip. - v1 scope: suggestion rail only — auto-apply stays whole-image (conservative; heads' thresholds were calibrated on whole-image). Bulk-apply bag = follow-up. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
b7fd69815e |
feat(agent): raise worker cap to 32 + size the HTTP pool for it (#114)
At 8 workers the GPU sat at ~5% util / <5GB VRAM — the pipeline is I/O-bound (downloading + decoding images over HTTP), so the GPU starves until many workers overlap that I/O. Raise MAX_CONCURRENCY 8→32 and make the UI worker control a number input (reaching 32 by ±1 was tedious); the cap is reported via /status so the UI clamps to it. Also size the shared requests pool (pool_maxsize=64) — the default 10 would have throttled 32 workers + spammed "connection pool is full". Verified by running; watch GPU util/VRAM climb as you dial up. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
4a1a9ec5a7 |
feat(agent): GPU load readout + live worker-count tuning (#114)
Control UI gains what the operator asked for: - GPU load (nvidia-smi): util %, VRAM used/total + bar, temp — so you can see how hard the card is working while you're at the desktop. - Worker count is now a live − / + control (POST /concurrency), not just an env: the worker is a pool of independent slots (shared model, so slots add concurrent inference, not N× VRAM). Dial up for speed, down to free the card. Replaces pause/resume with Start/Stop + the worker dial. - Graceful release on stop / pool-shrink: a slot hands its still-leased jobs back via client.release() so they're re-picked immediately (pairs with the server recovery sweep). Not CI-tested (agent/ outside CI) — verified by running. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |
||
|
bvandeusenandClaude Opus 4.8
|
8419ebd761 |
feat(agent): desktop GPU agent container — CCIP + figure crops over HTTP (#114)
The last piece: a Dockerised desktop-GPU worker that talks to FC ONLY over HTTP (lease → fetch pixels → detect figures + CCIP-embed → submit), so Redis/Postgres stay private. New top-level agent/ (outside CI scope — verified by running it): - fc_agent/worker.py: the lease/compute/submit loop, concurrency 1, start/pause/ stop (stop frees the card; unprocessed leases expire + re-queue). - fc_agent/models.py: imgutils wrappers — detect_person (figures) + CCIP embed. The two API seams to verify against the installed dghs-imgutils (flagged). - fc_agent/media.py: stills + video frame sampling (ffmpeg) at FC's cadence → per-frame instances (the bag). - fc_agent/crops.py: vendored crop primitive. client.py: the FC HTTP client. - fc_agent/app.py: FastAPI localhost control UI (start/pause/stop + progress + queue depth). Dockerfile (CUDA + onnxruntime-gpu + ffmpeg) + requirements + README (token → build → run --gpus all → Start; CPU-fallback path). This completes the CCIP pipeline end to end: agent produces region CCIP vectors → RegionService stores → matcher suggests characters → rail. Verified by running on the desktop (not CI). README calls out the imgutils API + model-string checks. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Ttrj5P7upUTueSfoJcxEqa |