Hugging Face
Models
Datasets
Spaces
Buckets
new
Docs
Enterprise
Pricing
Website
Tasks
HuggingChat
Collections
Languages
Organizations
Community
Blog
Posts
Daily Papers
Hardware
Learn
Discord
Forum
GitHub
Solutions
Team & Enterprise
Hugging Face PRO
Enterprise Support
Inference Providers
Inference Endpoints
Storage Buckets
Log In
Sign Up
resoajoe
/
loglens-longvideo-drift
Like
0
English
video-generation
long-horizon
temporal-consistency
drift
edge-inference
reproducibility
autoencoder
vae
evaluation-metrics
License:
mit
Model card
Files
Files and versions
xet
Community
Copy to bucket
new
main
loglens-longvideo-drift
602 kB
Ctrl+K
Ctrl+K
1 contributor
History:
58 commits
resoajoe
Card standard line 5: domain measured / deployment domain tested
c3f3d73
verified
about 1 month ago
articles
CORRECTION: interior optimum does not transfer to a second scene
about 2 months ago
chroma
Isoluminant control extended to SDXL + SD-1.5 on both chromatic axes. All three published periods confirmed unchanged (8/16/24px). Provisional caveat removed.
2 months ago
code
Add measured auto-exposure transfer function (paired locked/auto cameras, 30h): 83% of diurnal luminance range and 86% of colour temperature removed, response inverted. Plus methods article on instrument failure rates.
2 months ago
data
Upload data/chainlaw_drift_by_depth.json with huggingface_hub
about 2 months ago
exposure
Add measured auto-exposure transfer function (paired locked/auto cameras, 30h): 83% of diurnal luminance range and 86% of colour temperature removed, response inverted. Plus methods article on instrument failure rates.
2 months ago
nullcheck
Upload nullcheck/nullcheck.py with huggingface_hub
about 2 months ago
results
Upload folder using huggingface_hub
2 months ago
tripod
Upload tripod/pyproject.toml with huggingface_hub
about 2 months ago
vae-mtf
Upload vae-mtf/measured_LTX-Video_512.json with huggingface_hub
about 2 months ago
warpcost
warpcost: measure what a stabilisation warp costs (GAP 4)
about 2 months ago
.gitattributes
Safe
1.64 kB
Add measured auto-exposure transfer function (paired locked/auto cameras, 30h): 83% of diurnal luminance range and 86% of colour temperature removed, response inverted. Plus methods article on instrument failure rates.
2 months ago
BENCHMARK_stillness_v0.md
Safe
10.9 kB
Three-tier result; censoring methodology; sample duration scales with model quality
about 2 months ago
FINDINGS.md
Safe
8.24 kB
Correction: anchor is a generated image, not a photograph
2 months ago
NEGATIVE_chroma_detector.md
Safe
5.25 kB
Upload NEGATIVE_chroma_detector.md with huggingface_hub
about 2 months ago
PREREGISTRATION_chaining_law.md
Safe
9.54 kB
Upload PREREGISTRATION_chaining_law.md with huggingface_hub
about 2 months ago
PREREGISTRATION_eventrate.md
Safe
9.78 kB
Add pre-registered event-rate result: generated video has no stillness (real 28.8min vs generated 4s max). Instrument v1 void and rebuilt; 2 of 4 sealed predictions wrong, reported as such.
2 months ago
PREREGISTRATION_stillness_ood.md
Safe
4.22 kB
Upload PREREGISTRATION_stillness_ood.md with huggingface_hub
about 2 months ago
README.md
Safe
20.8 kB
Card standard line 5: domain measured / deployment domain tested
about 1 month ago
REGISTRY.md
Safe
9.69 kB
warpcost: measure what a stabilisation warp costs (GAP 4)
about 2 months ago
RESULTS_chaining_depth.md
Safe
7.23 kB
Upload RESULTS_chaining_depth.md with huggingface_hub
about 2 months ago
RESULTS_eventrate.md
Safe
4.87 kB
Add pre-registered event-rate result: generated video has no stillness (real 28.8min vs generated 4s max). Instrument v1 void and rebuilt; 2 of 4 sealed predictions wrong, reported as such.
2 months ago
RESULTS_exposure.md
Safe
5.45 kB
Add measured auto-exposure transfer function (paired locked/auto cameras, 30h): 83% of diurnal luminance range and 86% of colour temperature removed, response inverted. Plus methods article on instrument failure rates.
2 months ago
RESULTS_methods.md
Safe
8.78 kB
Add measured auto-exposure transfer function (paired locked/auto cameras, 30h): 83% of diurnal luminance range and 86% of colour temperature removed, response inverted. Plus methods article on instrument failure rates.
2 months ago
RESULTS_practice.md
Safe
10.5 kB
Add practice guide: how not to believe your own instruments. Five practices with measured cost/saving, the literature-check prompt skeleton, grounding-as-constraints pattern, and the prediction scorecard.
2 months ago
RESULTS_substab.md
Safe
7.9 kB
warpcost: measure what a stabilisation warp costs (GAP 4)
about 2 months ago
scenes_v0.json
Safe
4.19 kB
Add v0 scene set: 20 scenes, 5 seeds, occupancy split, duration guidance, reporting requirements
about 2 months ago
tripod_test.py
Safe
13.2 kB
First cross-configuration results; document reset-gaming and post-processing caveats
about 2 months ago