[Cherry-pick] PRs #1648 #1650 #1594 #1269 #1326 #1652 #1651 #1601 #1653 #1558 #1670 #1662 #1677 #1327 #1673 #1676 #1687 #1678 #1691 #1697 #1702 #1704 #1726 #1729#1734

Merged

kevalmorabia97 merged 24 commits into

release/0.45.0from

cherry-picks/release-0.45.0

Jun 15, 2026

Commits on Jun 15, 2026

Fix CodeCov upload issues (#1648)
kevalmorabia97
committed
fix(llm_eval): repair test_qwen3_eval_fp8 end-to-end (#1650)

kevalmorabia97
and
claude
committed
Add Alpamayo-1 example (#1594)

rohansjoshi
authored and
kevalmorabia97
committed
Skip Softmax diffusion export (#1269)

committed
Add DMD2 distillation for Qwen-Image (fastgen) (#1326)

authored and
kevalmorabia97
committed
feat(recipes): add kv_fp8_cast variants for partial-NVFP4 and weight-only PTQ recipes (#1652)

authored and
kevalmorabia97
committed
ci: cache JIT-compiled CUDA torch extensions in GPU/example tests (#1651)

kevalmorabia97
and
claude
committed
Migrate Nemotron-3-Nano tutorial PTQ to MBridge scripts and move under examples/megatron_bridge (#1601)

kevalmorabia97
and
claude
committed
feat(deepseek): add --cast_mxfp4_to_nvfp4 to deepseek_v4 quantize step (#1653)

authored and
kevalmorabia97
committed
[5924759] Fix fp16 ONNX INT8 entropy calibration on numpy >= 2.0 (#1558)

authored and
kevalmorabia97
committed
Fix chat_template handling in get_dataset_dataloader (#1670)
kevalmorabia97
committed
docs: add modelopt_recipes README and PTQ recipe/scheme guide (#1662)

authored and
kevalmorabia97
committed
specdec_bench: keep method=mtp when adding model=<assistant> for Gemma 4 MTP (#1677)

ChenhanYu
authored and
kevalmorabia97
committed
Add Nemotron-H mixed-precision PTQ recipe (GGUF Q4_K_M-mirrored) (#1327)

authored and
kevalmorabia97
committed
Fix garbage generation preview in hf_ptq.py when pad_token == eos_token (#1673)

Fridah-nv
authored and
kevalmorabia97
committed
[6294905] Fix --quant_cfg CLI parsing type in transformer trainer (#1676)

kinjalpatel27
authored and
kevalmorabia97
committed
[nvbug 6294744] Exclude Qwen visual modules from NVFP4 quantization (#1687)

meenchen
authored and
kevalmorabia97
committed
Fix GPT-OSS MXFP4->NVFP4 PTQ load, export, and cast (nvbug 6295279, 6295242) (#1678)

cjluo-nv
authored and
kevalmorabia97
committed
Exclude multimodal vision branch from quantization by default (NVBug 6293731, 6293762) (#1691)

authored and
kevalmorabia97
committed
[6287717][ONNX][Quantization] Preserve trt.plugins custom-op value_info in clear_stale_value_info (#1697)

gcunhase
authored and
kevalmorabia97
committed
Fix DeepSeek V3 ptq.py inference-repo path resolution (nvbug 6311147) (#1702)

authored and
kevalmorabia97
committed
Fix nested use_cache disabling for calibration (#1704)

meenchen
authored and
kevalmorabia97
committed
fix(puzzletron): use prebuilt KD dataset to avoid 136GB download (#1726)

TheSabari07
authored and
kevalmorabia97
committed
fix memory leak issue during puzzletron scoring, #1681 (#1729)

Separius
authored and
kevalmorabia97
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[Cherry-pick] PRs #1648 #1650 #1594 #1269 #1326 #1652 #1651 #1601 #1653 #1558 #1670 #1662 #1677 #1327 #1673 #1676 #1687 #1678 #1691 #1697 #1702 #1704 #1726 #1729#1734

[Cherry-pick] PRs #1648 #1650 #1594 #1269 #1326 #1652 #1651 #1601 #1653 #1558 #1670 #1662 #1677 #1327 #1673 #1676 #1687 #1678 #1691 #1697 #1702 #1704 #1726 #1729#1734
kevalmorabia97 merged 24 commits into
release/0.45.0from
cherry-picks/release-0.45.0

Commits on Jun 15, 2026

Fix CodeCov upload issues (#1648)

fix(llm_eval): repair test_qwen3_eval_fp8 end-to-end (#1650)

Add Alpamayo-1 example (#1594)

Skip Softmax diffusion export (#1269)

Add DMD2 distillation for Qwen-Image (fastgen) (#1326)

feat(recipes): add kv_fp8_cast variants for partial-NVFP4 and weight-only PTQ recipes (#1652)

ci: cache JIT-compiled CUDA torch extensions in GPU/example tests (#1651)

Migrate Nemotron-3-Nano tutorial PTQ to MBridge scripts and move under examples/megatron_bridge (#1601)

feat(deepseek): add --cast_mxfp4_to_nvfp4 to deepseek_v4 quantize step (#1653)

[5924759] Fix fp16 ONNX INT8 entropy calibration on numpy >= 2.0 (#1558)

Fix chat_template handling in get_dataset_dataloader (#1670)

docs: add modelopt_recipes README and PTQ recipe/scheme guide (#1662)

specdec_bench: keep method=mtp when adding model=<assistant> for Gemma 4 MTP (#1677)

Add Nemotron-H mixed-precision PTQ recipe (GGUF Q4_K_M-mirrored) (#1327)

Fix garbage generation preview in hf_ptq.py when pad_token == eos_token (#1673)

[6294905] Fix --quant_cfg CLI parsing type in transformer trainer (#1676)

[nvbug 6294744] Exclude Qwen visual modules from NVFP4 quantization (#1687)

Fix GPT-OSS MXFP4->NVFP4 PTQ load, export, and cast (nvbug 6295279, 6295242) (#1678)

Exclude multimodal vision branch from quantization by default (NVBug 6293731, 6293762) (#1691)

[6287717][ONNX][Quantization] Preserve trt.plugins custom-op value_info in clear_stale_value_info (#1697)

Fix DeepSeek V3 ptq.py inference-repo path resolution (nvbug 6311147) (#1702)

Fix nested use_cache disabling for calibration (#1704)

fix(puzzletron): use prebuilt KD dataset to avoid 136GB download (#1726)

fix memory leak issue during puzzletron scoring, #1681 (#1729)