Supervisor audit finding — 2026-07-06
Org: ByteDance (HuggingFace slug: ByteDance — note case)
Models: 53 models on HF. Recent: Bernini-R (video generation, ~Jun 11), EvoQuality.
Activity: Active, distinct from the already-tracked bytedance-seed and bytedance-research orgs.
Context
The supervisor scanned the Jun 29–Jul 6 window and found ByteDance as a potential gap. ByteDance now operates three separate HF orgs:
| Org |
Status |
bytedance-seed |
✅ In KNOWN_ORGS + MAJOR_HF_ORGS |
bytedance-research |
✅ In KNOWN_ORGS + MAJOR_HF_ORGS |
ByteDance (main) |
❌ Not tracked |
Why this is a manual call
The "sustained real model releases" bar is ambiguous here:
- For: 53 models is a large catalog; the main org exists and is active; ByteDance is one of the world's largest AI companies.
- Against: The confirmed recent release (Bernini-R) is a video generation model — which would be filtered by the
text-to-video task-type gate in is_noise_model before reaching the significance or categorization logic. It's unclear whether the main ByteDance org publishes LLM-type (text-generation) models distinct from what bytedance-seed and bytedance-research already cover.
Adding ByteDance to MAJOR_HF_ORGS with a primarily video/image catalog would increase fetch volume with low signal yield.
Recommendation
Verify whether ByteDance on HF has LLM-type (text-generation) models not already published under the seed/research orgs. If yes and engagement is solid, add then. If the catalog is primarily video/image gen, skip.
What to add if approved:
# PROVIDER_NAMES
"bytedance": "ByteDance",
# KNOWN_ORGS
"bytedance",
# MAJOR_HF_ORGS
"ByteDance",
Note: KNOWN_ORGS and PROVIDER_NAMES use lowercase; MAJOR_HF_ORGS must use the exact HF case (ByteDance) to avoid the casing bug documented in issues #24 and #28.
Supervisor audit finding — 2026-07-06
Org:
ByteDance(HuggingFace slug:ByteDance— note case)Models: 53 models on HF. Recent: Bernini-R (video generation, ~Jun 11), EvoQuality.
Activity: Active, distinct from the already-tracked
bytedance-seedandbytedance-researchorgs.Context
The supervisor scanned the Jun 29–Jul 6 window and found
ByteDanceas a potential gap. ByteDance now operates three separate HF orgs:bytedance-seedbytedance-researchByteDance(main)Why this is a manual call
The "sustained real model releases" bar is ambiguous here:
text-to-videotask-type gate inis_noise_modelbefore reaching the significance or categorization logic. It's unclear whether the mainByteDanceorg publishes LLM-type (text-generation) models distinct from whatbytedance-seedandbytedance-researchalready cover.Adding
ByteDancetoMAJOR_HF_ORGSwith a primarily video/image catalog would increase fetch volume with low signal yield.Recommendation
Verify whether
ByteDanceon HF has LLM-type (text-generation) models not already published under the seed/research orgs. If yes and engagement is solid, add then. If the catalog is primarily video/image gen, skip.What to add if approved:
Note:
KNOWN_ORGSandPROVIDER_NAMESuse lowercase;MAJOR_HF_ORGSmust use the exact HF case (ByteDance) to avoid the casing bug documented in issues #24 and #28.