Repository navigation

fix(python): complete model execution profiles - #856

Closed

chaofengw-nv wants to merge 42 commits into

NVIDIA:mainfrom

chaofengw-nv:fix/python-profiles-self-contained

Closed

fix(python): complete model execution profiles#856
chaofengw-nv wants to merge 42 commits into
NVIDIA:mainfrom
chaofengw-nv:fix/python-profiles-self-contained

Commits on Aug 17, 2026
- fix(minimax-h3): isolate reference hub dependency
  chaofengw-nv
  committed
  3c9b359
- fix(validation): pin compatible Hugging Face hub
  chaofengw-nv
  committed
  2f8fad6
- fix(deepseek-v2): pin tokenizer profile
  chaofengw-nv
  committed
  8a24ade
- fix(validation): pin SacreBLEU dependencies
  chaofengw-nv
  committed
  7aefa40
- fix(elf): close build profile dependencies
  chaofengw-nv
  committed
  1e52ffd
- fix(validation): pin CLIP scoring profile
  chaofengw-nv
  committed
  6add525
- fix(personaplex): close evaluator profile
  chaofengw-nv
  committed
  40e4f01
- fix(olmo): align tokenizer environments
  chaofengw-nv
  committed
  e94f5ca
- fix(phi4): pin chat template dependency
  chaofengw-nv
  committed
  34f8203
- test(ci): derive fake image profile inventory
  chaofengw-nv
  committed
  c5646da
- fix(reference): add chat template dependencies
  chaofengw-nv
  committed
  7e139a9
- fix(internlm): pin tokenizer conversion dependency
  chaofengw-nv
  committed
  b473460
- fix(xglm): align tokenizer execution profiles
  chaofengw-nv
  committed
  6ddbe27
- fix(sana): align build and reference environments
  chaofengw-nv
  committed
  5bb7f0c
- fix(xglm): preserve qualified token framing
  chaofengw-nv
  committed
  2dc09ad
- fix(profiles): complete L4T reference environments
  chaofengw-nv
  committed
  8f572dc
- fix(profiles): complete speech reference environments
  chaofengw-nv
  committed
  c1a9e1c
- fix(profiles): preserve nested environment overlays
  chaofengw-nv
  committed
  d1b8b7e
- fix(profiles): complete speech model environments
  chaofengw-nv
  committed
  f3211b0
- fix(profiles): correct speech parso pins
  chaofengw-nv
  committed
  68f87cf
- fix(profiles): complete speech runtime dependencies
  chaofengw-nv
  committed
  184c4ea
- fix(profiles): verify Nemotron hybrid loader
  chaofengw-nv
  committed
  210a324
- fix(reference): complete Magpie G2P profile
  chaofengw-nv
  committed
  6dbe897
- fix(profiles): bootstrap source dependencies
  chaofengw-nv
  committed
  00b104d
- fix(magpie): use TTS profile for builds
  chaofengw-nv
  committed
  0bf5dd8
- fix(wan): complete official reference profile
  chaofengw-nv
  committed
  bf2d498
- fix(profiles): skip Lance FlashAttention on L4T
  chaofengw-nv
  committed
  c5e4697
- fix(lance): pin official reference dependency
  chaofengw-nv
  committed
  1ba918d
- fix(qwen-image): own reference environment
  chaofengw-nv
  committed
  8f8728f
- test(ci): isolate profile resolution in unit tests
  chaofengw-nv
  committed
  62870dc
- fix(python): pin open-clip text dependency
  chaofengw-nv
  committed
  0afe13f
- fix(python): align DeepSeek MoE profile
  chaofengw-nv
  committed
  144b6b9
- fix(python): isolate timm reference dependency
  chaofengw-nv
  committed
  ed953f6
- fix(python): complete Chronos profile dependencies
  chaofengw-nv
  committed
  810873e
- fix(perf): load Nemotron 3.5 archive compatibly
  chaofengw-nv
  committed
  be7bdb2
- fix(perf): preflight ModelOpt calibration dependency
  chaofengw-nv
  committed
  0fdf622
- fix(perf): load SAM3 processor config
  chaofengw-nv
  committed
  b909a27
- fix(validation): bound TTS transcription memory
  chaofengw-nv
  committed
  5016c13
- fix(perf): disable NeMo RNNT CUDA graphs
  chaofengw-nv
  committed
  b7b2590
- fix(python): target CUDA profile builds
  chaofengw-nv
  committed
  f225c16
- fix(profiles): complete CI profile builds
  chaofengw-nv
  committed
  7e3ea18
- fix(profiles): use build profile for bundles
  chaofengw-nv
  committed
  9159452

Commits

Commits on Aug 17, 2026