Aman Gupta 3e12fbdea5 llama: avoid copying logits during prompt decode in MTP (#23198) ...

* llama: avoid copying logits during prompt decode in MTP

* review: update comment

* llama-graph: call set_output for t_h_pre_norm

2026-05-17 23:30:25 +08:00

..

afmoe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

apertus.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

arcee.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

arctic.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

arwkv7.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

baichuan.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

bailingmoe2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

bailingmoe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

bert.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

bitnet.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

bloom.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

chameleon.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

chatglm.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

codeshell.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

cogvlm.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

cohere2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

command-r.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

dbrx.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

deci.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

deepseek2.cpp

model : fix model type check for granite/llama3 and deepseek2/glm4.7 lite (#22870)

2026-05-10 08:44:29 +02:00

deepseek2ocr.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

deepseek.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

delta-net-base.cpp

llama + spec: MTP Support (#22673)

2026-05-16 20:06:23 +08:00

dots1.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

dream.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

ernie4-5-moe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

ernie4-5.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

eurobert.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

exaone4.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

exaone-moe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

exaone.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

falcon-h1.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

falcon.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

gemma2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

gemma3.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

gemma3n.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

gemma4.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

gemma-embedding.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

gemma.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

glm4-moe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

glm4.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

glm-dsa.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

gpt2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

gptneox.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

granite-hybrid.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

granite-moe.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

granite.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

grok.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

grovemoe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

hunyuan-dense.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

hunyuan-moe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

hunyuan-vl.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

internlm2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

jais2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

jais.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

jamba.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

jina-bert-v2.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

jina-bert-v3.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

kimi-linear.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

lfm2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

lfm2moe.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

llada-moe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

llada.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

llama4.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

llama-embed.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

llama.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

maincoder.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

mamba2.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

mamba-base.cpp

model : wire up Nemotron-H tensors for NVFP4 support (#20561)

2026-03-16 09:19:16 +01:00

mamba.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

mimo2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

minicpm3.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

minicpm.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

minimax-m2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

mistral3.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

mistral4.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

models.h

llama + spec: MTP Support (#22673)

2026-05-16 20:06:23 +08:00

modern-bert.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

mpt.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

nemotron-h-moe.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

nemotron-h.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

nemotron.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

neo-bert.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

nomic-bert-moe.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

nomic-bert.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

olmo2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

olmo.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

olmoe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

openai-moe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

openelm.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

orion.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

paddleocr.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

pangu-embed.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

phi2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

phi3.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

phimoe.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

plamo2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

plamo3.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

plamo.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

plm.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

qwen2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

qwen2moe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

qwen2vl.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

qwen3.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

qwen3moe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

qwen3next.cpp

llama + spec: MTP Support (#22673)

2026-05-16 20:06:23 +08:00

qwen3vl.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

qwen3vlmoe.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

qwen35.cpp

llama: avoid copying logits during prompt decode in MTP (#23198)

2026-05-17 23:30:25 +08:00

qwen35moe.cpp

llama: avoid copying logits during prompt decode in MTP (#23198)

2026-05-17 23:30:25 +08:00

qwen.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

refact.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

rnd1.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

rwkv6-base.cpp

models : deduplicate delta-net graphs for Qwen family (#19597)

2026-02-16 14:35:04 +02:00

rwkv6.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

rwkv6qwen2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

rwkv7-base.cpp

models : deduplicate delta-net graphs for Qwen family (#19597)

2026-02-16 14:35:04 +02:00

rwkv7.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

seed-oss.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

smallthinker.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

smollm3.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

stablelm.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

starcoder2.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

starcoder.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

step35.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

t5.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

t5encoder.cpp

model: move load_hparams and load_tensors to per-model definition (#22004)

2026-05-04 12:36:59 +02:00

wavtokenizer-dec.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00

xverse.cpp

model : NvFP4 quantized LM head support (#23046)

2026-05-16 11:09:27 +02:00