| .. |
|
models
|
a5251ca11d
Optimization: Qwen3 next autoregressive pass (#17996)
|
1 ماه پیش |
|
CMakeLists.txt
|
63908b631a
cmake: fix Mach-O current version number (#17877)
|
1 ماه پیش |
|
llama-adapter.cpp
|
fd621880f3
aLoRA Support (#15327)
|
4 ماه پیش |
|
llama-adapter.h
|
fd621880f3
aLoRA Support (#15327)
|
4 ماه پیش |
|
llama-arch.cpp
|
982060fadc
model: fix LFM2_MOE missing tensors (#18132)
|
1 ماه پیش |
|
llama-arch.h
|
7f2b2f3c77
arch: refactor LLM_TENSOR_NAMES (#18051)
|
1 ماه پیش |
|
llama-batch.cpp
|
d9f8f60618
batch : fix sequence id ownership (#17915)
|
1 ماه پیش |
|
llama-batch.h
|
d9f8f60618
batch : fix sequence id ownership (#17915)
|
1 ماه پیش |
|
llama-chat.cpp
|
9f052478c2
model : add openPangu-Embedded (#16941)
|
2 ماه پیش |
|
llama-chat.h
|
9f052478c2
model : add openPangu-Embedded (#16941)
|
2 ماه پیش |
|
llama-context.cpp
|
b1f3a6e5db
llama: automatically set parameters not set by the user in such a way that maximizes GPU utilization (#16653)
|
1 ماه پیش |
|
llama-context.h
|
b1f3a6e5db
llama: automatically set parameters not set by the user in such a way that maximizes GPU utilization (#16653)
|
1 ماه پیش |
|
llama-cparams.cpp
|
c311ac664d
cparams : rename LLAMA_MAX_PARALLEL_SEQUENCES to LLAMA_MAX_SEQ (#14188)
|
7 ماه پیش |
|
llama-cparams.h
|
cd5e3b5754
server : support unified cache across slots (#16736)
|
2 ماه پیش |
|
llama-grammar.cpp
|
e39502e74b
llama : add token matching support to llama-grammar (#17816)
|
1 ماه پیش |
|
llama-grammar.h
|
e39502e74b
llama : add token matching support to llama-grammar (#17816)
|
1 ماه پیش |
|
llama-graph.cpp
|
c560316440
graph : reuse SSM graphs (#16490)
|
1 ماه پیش |
|
llama-graph.h
|
c560316440
graph : reuse SSM graphs (#16490)
|
1 ماه پیش |
|
llama-hparams.cpp
|
3d86c6c2b5
model: support GLM4V vision encoder (#18042)
|
1 ماه پیش |
|
llama-hparams.h
|
3d86c6c2b5
model: support GLM4V vision encoder (#18042)
|
1 ماه پیش |
|
llama-impl.cpp
|
b1f3a6e5db
llama: automatically set parameters not set by the user in such a way that maximizes GPU utilization (#16653)
|
1 ماه پیش |
|
llama-impl.h
|
37adc9c6ba
ggml, llama : use defaulted constructors/destructors (#17649)
|
1 ماه پیش |
|
llama-io.cpp
|
e0dbec0bc6
llama : refactor llama_context, llama_kv_cache, llm_build_context (#12181)
|
10 ماه پیش |
|
llama-io.h
|
e0dbec0bc6
llama : refactor llama_context, llama_kv_cache, llm_build_context (#12181)
|
10 ماه پیش |
|
llama-kv-cache-iswa.cpp
|
16bcc1259d
kv-cache : pad the cache size to 256 for performance (#17046)
|
2 ماه پیش |
|
llama-kv-cache-iswa.h
|
e789095502
llama: print memory breakdown on exit (#15860)
|
3 ماه پیش |
|
llama-kv-cache.cpp
|
4529c660c8
kv-cache: Fix state restore fragmented cache (#17982)
|
1 ماه پیش |
|
llama-kv-cache.h
|
4529c660c8
kv-cache: Fix state restore fragmented cache (#17982)
|
1 ماه پیش |
|
llama-kv-cells.h
|
e3af5563bd
llama: store mrope data in KV cell (#16825)
|
2 ماه پیش |
|
llama-memory-hybrid.cpp
|
c560316440
graph : reuse SSM graphs (#16490)
|
1 ماه پیش |
|
llama-memory-hybrid.h
|
e789095502
llama: print memory breakdown on exit (#15860)
|
3 ماه پیش |
|
llama-memory-recurrent.cpp
|
0c74f32632
memory: Hybrid context shift (#17009)
|
2 ماه پیش |
|
llama-memory-recurrent.h
|
7a0e900e36
llama: consistent ctx <-> buf order for KV cache (#16746)
|
2 ماه پیش |
|
llama-memory.cpp
|
745f11fed0
memory : correctly handle failure in apply() (#14438)
|
6 ماه پیش |
|
llama-memory.h
|
e789095502
llama: print memory breakdown on exit (#15860)
|
3 ماه پیش |
|
llama-mmap.cpp
|
4d4f4cacd1
llama : Async DirectIO model loading on Linux (#18012)
|
1 ماه پیش |
|
llama-mmap.h
|
4d4f4cacd1
llama : Async DirectIO model loading on Linux (#18012)
|
1 ماه پیش |
|
llama-model-loader.cpp
|
4d4f4cacd1
llama : Async DirectIO model loading on Linux (#18012)
|
1 ماه پیش |
|
llama-model-loader.h
|
b1f3a6e5db
llama: automatically set parameters not set by the user in such a way that maximizes GPU utilization (#16653)
|
1 ماه پیش |
|
llama-model-saver.cpp
|
88fc854b4b
llama : improve sep token handling (#14272)
|
7 ماه پیش |
|
llama-model-saver.h
|
10d2af0eaa
llama/ggml: add LLM training support (#10544)
|
8 ماه پیش |
|
llama-model.cpp
|
57c1e05643
llama: offload output layer to GPU first (#18148)
|
1 ماه پیش |
|
llama-model.h
|
2995341730
llama : add support for NVIDIA Nemotron 3 Nano (#18058)
|
1 ماه پیش |
|
llama-quant.cpp
|
b1f3a6e5db
llama: automatically set parameters not set by the user in such a way that maximizes GPU utilization (#16653)
|
1 ماه پیش |
|
llama-quant.h
|
f66f582927
llama : refactor `src/llama.cpp` (#10902)
|
1 سال پیش |
|
llama-sampling.cpp
|
4301e27319
common : restore grammar-based rejection sampling (#18137)
|
1 ماه پیش |
|
llama-sampling.h
|
afa8a9ec9b
llama : add `llama_vocab`, functions -> methods, naming (#11110)
|
1 سال پیش |
|
llama-vocab.cpp
|
9d52f17ae3
model : add KORMo model (#18032)
|
1 ماه پیش |
|
llama-vocab.h
|
e1fcf8b09b
model : add AfmoeForCausalLM support (#16477)
|
2 ماه پیش |
|
llama.cpp
|
57c1e05643
llama: offload output layer to GPU first (#18148)
|
1 ماه پیش |
|
unicode-data.cpp
|
458367a906
server : better security control for public deployments (#9776)
|
1 سال پیش |
|
unicode-data.h
|
a39ab216aa
llama : reduce compile time and binary size (#9712)
|
1 سال پیش |
|
unicode.cpp
|
1be97831e4
fix: prevent segfault in tokenizer on highly repetitive input (#17786)
|
1 ماه پیش |
|
unicode.h
|
624207e676
devops: add s390x & ppc64le CI (#15925)
|
3 ماه پیش |
Johannes Gäßler