Commit History

Add isolated four-view v7 cache worker
5af941a
verified

patdev commited on

Run v7 cache builder in isolated child process
2b275d8
verified

patdev commited on

Add v7 staged distillation L4 worker
3b8a1a7
verified

patdev commited on

Add v7 teacher-cache L4 worker
5190bf4
verified

patdev commited on

Skip 809MB SLat DAE checkpoint on full TensorRT path
f6bb94e
verified

patdev commited on

v6.4: use SLat shell plus TRT core; remove full flow checkpoints from nominal Job
645e7c4
verified

patdev commited on

Companion Forge v6.3: skip native SS checkpoints when TensorRT is available
477935c
verified

patdev commited on

Install pinned TensorRT runtime for FLUX Jobs
bf5483d
verified

patdev commited on

Restrict AniGen Job downloads to DINO and DSINE TensorRT plans
4577d9a
verified

patdev commited on

Add FLUX.2 TensorRT 512 reference worker
a45ef81
verified

patdev commited on

Measure true L4 VRAM across PyTorch ORT and TensorRT
988c702
verified

patdev commited on

Companion Forge v6.1 staged TensorRT conditioning pipeline
86d6a06
verified

patdev commited on

Pin TensorRT 11.2.1.2 to match L4 engines
8a1621b
verified

patdev commited on

Pin ONNX Runtime GPU 1.23.2 for AniGen Python 3.10
c5a0cd5
verified

patdev commited on

Install TensorRT 11 and ONNX Runtime GPU for hybrid worker
8a48ef1
verified

patdev commited on

Use TRT then ORT CUDA in hybrid AniGen worker
886afa6
verified

patdev commited on

Companion Forge v6: hybrid ONNX TensorRT L4 backend with native fallback
401cc86
verified

patdev commited on

Trim unused AniGen texture imports and profile fields on L4
c5ee28b
verified

patdev commited on

Use spconv Native and skip 1024 models in TRELLIS Fast L4
ec16ffd
verified

patdev commited on

Install L4-compatible spconv for TRELLIS Jobs
209a19c
verified

patdev commited on

Remove gated BRIA dependency from TRELLIS L4 Job
43d5625
verified

patdev commited on

Optimize rigged AniGen L4 export with vertex colors and no CUDA JIT
15558df
verified

patdev commited on

Use native spconv with FlashAttention for AniGen L4 stability
b3a36c6
verified

patdev commited on

Use xFormers sparse attention for stable AniGen SLat on L4
c341961
verified

patdev commited on

Skip unused AniGen DAE encoders on L4 cold start
1ee082e
verified

patdev commited on

Repair AniGen scientific Python ABI before L4 worker start
533197c
verified

patdev commited on

Use fixed NumPy/SciPy runtime instead of AniGen shims
da91a78
verified

patdev commited on

Patch AniGen PCA and Halton imports for stable L4 Jobs
0713a13
verified

patdev commited on

Make NumPy PCA shim compatible with torch dynamo import inspection
1c0971f
verified

patdev commited on

Bypass broken SciPy dependency with NumPy PCA shim in AniGen Jobs
620f201
verified

patdev commited on

Fix AniGen Space image path for HF Jobs
2869cfb
verified

patdev commited on

Add FLUX.2 Klein 4B L4 reference worker
7a34ac9
verified

patdev commited on

Load Z-Image transformer directly in BF16 on L4
28f42aa
verified

patdev commited on

Fallback to remote TRELLIS snapshot when Job model mount is truncated
fadd818
verified

patdev commited on

Use selective AniGen checkpoint download on L4 Jobs
dd76510
verified

patdev commited on

Fallback to remote Z-Image snapshot when Job model mount is truncated
25d76d9
verified

patdev commited on

Add private repo artifact transport for HF Jobs
ce3f6d7
verified

patdev commited on

Fix Z-Image transformer import for worker image
f8f93ba
verified

patdev commited on

Add optimized TRELLIS.2 L4 Jobs worker
ae245a7
verified

patdev commited on

Add optimized AniGen L4 rigging worker
33f10b4
verified

patdev commited on

Add optimized Z-Image L4 Jobs worker
330e0a4
verified

patdev commited on