JamePeng cc231cb0da dflash: pass missing NVFP4 scales to attention operations (#28000) ...

- DFlash2 NVFP4 draft models produced almost no accepted speculative
tokens because the Q, K, V, and output projection scales were not
passed to the corresponding graph operations.

2026-08-30 11:34:39 +03:00

..

models

dflash: pass missing NVFP4 scales to attention operations (#28000)

2026-08-30 11:34:39 +03:00

CMakeLists.txt

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-adapter.cpp

llama : check LoRA tensor data is within file bounds (#27056)

2026-08-16 09:38:01 +03:00

llama-adapter.h

…

llama-arch.cpp

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-arch.h

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-batch.cpp

llama-batch: fix allowed decreasing pos in a seq (#25449)

2026-07-08 19:24:34 +03:00

llama-batch.h

llama-batch: add n_keep_tail in split_equal for recurrent models (#25278)

2026-07-08 15:55:19 +08:00

llama-chat.cpp

…

llama-chat.h

…

llama-context.cpp

context : disable non-fused GDN and LID ops (#27877)

2026-08-28 16:34:26 +03:00

llama-context.h

llama: refactor fused ops (#24646)

2026-07-08 18:18:09 +08:00

llama-cparams.cpp

…

llama-cparams.h

llama : support multi-output backend sampling (#25532)

2026-08-10 16:58:56 +03:00

llama-ext.h

spec : add DFlash2 support (local convolution + candidate selector) (#27342) (#27816)

2026-08-27 19:17:07 +02:00

llama-grammar.cpp

grammar : parse \- in char classes as literal hyphen (#27591)

2026-08-25 09:05:24 +03:00

llama-grammar.h

…

llama-graph.cpp

model: add dots3-note (#27060)

2026-08-21 19:52:34 +02:00

llama-graph.h

model: add dots3-note (#27060)

2026-08-21 19:52:34 +02:00

llama-hparams.cpp

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-hparams.h

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-impl.cpp

…

llama-impl.h

llama: refactor fused ops (#24646)

2026-07-08 18:18:09 +08:00

llama-io.cpp

…

llama-io.h

…

llama-kv-cache-dsa-iswa.cpp

model: add dots3-note (#27060)

2026-08-21 19:52:34 +02:00

llama-kv-cache-dsa-iswa.h

model: add dots3-note (#27060)

2026-08-21 19:52:34 +02:00

llama-kv-cache-dsa.cpp

llama : allocate indexer cache only in "full" indexer layers (#26474)

2026-08-03 14:56:30 +02:00

llama-kv-cache-dsa.h

llama : allocate indexer cache only in "full" indexer layers (#26474)

2026-08-03 14:56:30 +02:00

llama-kv-cache-dsv4.cpp

Deepseek 4: -sm tensor (#26490)

2026-08-24 09:20:25 +03:00

llama-kv-cache-dsv4.h

DeepseekV4 MTP + DSpark (#25784)

2026-08-02 20:55:34 +08:00

llama-kv-cache-iswa.cpp

llama-batch: add n_keep_tail in split_equal for recurrent models (#25278)

2026-07-08 15:55:19 +08:00

llama-kv-cache-iswa.h

DeepSeek V4 (#24162)

2026-06-29 16:58:51 +08:00

llama-kv-cache-msa.cpp

model: M3: Move MSA into a new memory implementation (#26338)

2026-08-03 16:30:08 +03:00

llama-kv-cache-msa.h

model: M3: Move MSA into a new memory implementation (#26338)

2026-08-03 16:30:08 +03:00

llama-kv-cache.cpp

memory : copy Hadamard matrix to k_rot tensor only if it has buffer assigned to prevent crashes during context shift of unquantized K cache (#27967)

2026-08-30 07:47:15 +02:00

llama-kv-cache.h

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-kv-cells.h

llama: add token ID tracking to KV cell (#27762)

2026-08-26 23:34:28 +02:00

llama-memory-hybrid-idx.cpp

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-memory-hybrid-idx.h

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-memory-hybrid-iswa.cpp

llama-batch: add n_keep_tail in split_equal for recurrent models (#25278)

2026-07-08 15:55:19 +08:00

llama-memory-hybrid-iswa.h

…

llama-memory-hybrid.cpp

llama-batch: add n_keep_tail in split_equal for recurrent models (#25278)

2026-07-08 15:55:19 +08:00

llama-memory-hybrid.h

…

llama-memory-recurrent.cpp

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-memory-recurrent.h

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-memory.cpp

…

llama-memory.h

…

llama-mmap.cpp

llama: model_loader: add TENSOR_READ_LAZY (#27794)

2026-08-27 15:14:34 +02:00

llama-mmap.h

llama: model_loader: add TENSOR_READ_LAZY (#27794)

2026-08-27 15:14:34 +02:00

llama-model-loader.cpp

bench: add --tensor-read-lazy (#27881)

2026-08-28 20:51:05 +02:00

llama-model-loader.h

bench: add --tensor-read-lazy (#27881)

2026-08-28 20:51:05 +02:00

llama-model-saver.cpp

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-model-saver.h

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-model.cpp

bench: add --tensor-read-lazy (#27881)

2026-08-28 20:51:05 +02:00

llama-model.h

model: add DSpark support for Nemotron3.5 (#27804)

2026-08-28 01:49:27 +02:00

llama-quant.cpp

model: add Qwen3.8-Flash-Next (qwen4exp) (#27742)

2026-08-27 21:32:31 +02:00

llama-quant.h

…

llama-sampler.cpp

llama : support multi-output backend sampling (#25532)

2026-08-10 16:58:56 +03:00

llama-sampler.h

llama : support multi-output backend sampling (#25532)

2026-08-10 16:58:56 +03:00

llama-vocab.cpp

vocab : support integer tokenizer scores (#27260)

2026-08-17 20:13:15 +02:00

llama-vocab.h

Add support for Laguna XS.2 & M.1 (#25165)

2026-07-22 09:54:08 +08:00

llama.cpp

bench: add --tensor-read-lazy (#27881)

2026-08-28 20:51:05 +02:00

unicode-data.cpp

…

unicode-data.h

…

unicode.cpp

unicode : include '~' in collapsed symbol class (#26972)

2026-08-18 15:15:22 +02:00

unicode.h

…