COMMITS
June 12, 2026
M
Merge pull request #416 from AI-Hypercomputer:plumb_vmem_limit
maxdiffusion authors committed
June 9, 2026
E
Fix: Plumb vmem limit and bkv_compute_in into the custom splash attention kernel
Elisa Tsai committed
June 2, 2026
M
Merge pull request #415 from AI-Hypercomputer:ltx2_tpu_encoder_disable
maxdiffusion authors committed
May 30, 2026
R
feat(ltx2): make run_text_encoder_on_tpu default False and dynamically load torchax
Rishabh Manoj committed
M
Merge pull request #414 from AI-Hypercomputer:ltx2_sharding
maxdiffusion authors committed
May 22, 2026
R
feat(ltx2): implement logical sharding
Rishabh Manoj committed
M
Merge pull request #394 from AI-Hypercomputer:tests_fix
maxdiffusion authors committed
F
Fix NaN in LTX2 scheduler for long sequences
Flying Blackshark committed
May 29, 2026
M
Merge pull request #398 from AI-Hypercomputer:text_encoder_tpu3
maxdiffusion authors committed
April 30, 2026
M
Offload LTX-2 text encoder to TorchAX and resolve lint issues
mbohlool committed
May 27, 2026
M
Merge pull request #406 from AI-Hypercomputer:wan-orbax-checkpoint-cache
maxdiffusion authors committed
May 26, 2026
M
Merge pull request #411 from AI-Hypercomputer:wan_vae_opt
maxdiffusion authors committed
S
feat: add pretrained orbax cache for fast WAN inference loads
sagarchapara committed
R
Optimize and scale WAN VAE spatial sharding
Rishabh Manoj committed
May 21, 2026
M
Merge pull request #409 from AI-Hypercomputer:xprof_upload_fix
maxdiffusion authors committed
M
Merge pull request #413 from AI-Hypercomputer:prisha/img_fix
maxdiffusion authors committed
May 20, 2026
P
feat: add g++ compiler to base image and implement hardware-aware FeedForward sharding
Prisha Jain committed
M
Merge pull request #407 from AI-Hypercomputer:ninatu/wan_vae_sharding_bug
maxdiffusion authors committed
M
Merge pull request #408 from AI-Hypercomputer:ninatu/fix_text_encoder
maxdiffusion authors committed
May 18, 2026
R
Fix tests for Flux, WAN, SDXL, and LTX-Video
Rishabh Manoj committed
May 15, 2026
N
Fix VAE spatial sharding dynamic calculation bug in Wan pipeline.
Nina Shvetsova committed
E
Fix xprof gs bucket upload error
Elisa Tsai committed
N
Add `text_encoder_dtype` and `compile_text_encoder` config parameters for Wan text encoders.
Nina Shvetsova committed
May 13, 2026
M
Merge pull request #400 from AI-Hypercomputer:wan_kv_cache
maxdiffusion authors committed
M
Minor bug fix and changes to tpu_utils.py
maxdiffusion authors committed
May 12, 2026
M
Merge pull request #367 from AI-Hypercomputer:sagarchapara/wananimate-pipeline
maxdiffusion authors committed
C
Cancel in-progress workflow runs on new push (#403)
csgoogle committed
M
Merge pull request #402 from AI-Hypercomputer:ltx23
maxdiffusion authors committed
P
LTX2.3 model support
Prisha Jain committed
May 11, 2026
S
Add WAN Animate pipeline with face motion pre-computation
Sagar Chapara committed