 Georgi Gerganov
|
745aa5319b
llama : deprecate llama_kv_self_ API (#14030)
|
преди 7 месеца |
 Sigbjørn Skjæret
|
d17a809ef0
llama : support multiple classifier outputs and labels (#13940)
|
преди 7 месеца |
|
Georgi Gerganov
|
7f37b6cf1e
memory : migrate from llama_kv_cache to more generic llama_memory (#14006)
|
преди 7 месеца |
|
Georgi Gerganov
|
803f8baf4f
llama : deprecate explicit kv_self defrag/update calls (#13921)
|
преди 7 месеца |
|
Georgi Gerganov
|
3600cc2886
llama : use n_swa + n_ubatch cells for SWA cache (#13833)
|
преди 7 месеца |
|
Georgi Gerganov
|
12d0188c0d
kv-cache : refactor + add llama_memory_state_i (#13746)
|
преди 7 месеца |
|
Georgi Gerganov
|
22229314fc
llama : clarify deprecation message (#13794)
|
преди 8 месеца |
|
Georgi Gerganov
|
de2ef53a4b
kv-cache : rework kv_cell (#13706)
|
преди 8 месеца |
|
Georgi Gerganov
|
797f2ac062
kv-cache : simplify the interface (#13660)
|
преди 8 месеца |
|
Georgi Gerganov
|
a4090d1174
llama : remove llama_kv_cache_view API + remove deprecated (#13653)
|
преди 8 месеца |
|
Georgi Gerganov
|
e298d2fbd0
kv-cache : add SWA support (#13194)
|
преди 8 месеца |
 Diego Devesa
|
cf0a43bb64
llama-bench : add defrag-thold, check for invalid ranges (#13487)
|
преди 8 месеца |
 Johannes Gäßler
|
10d2af0eaa
llama/ggml: add LLM training support (#10544)
|
преди 8 месеца |
 David Huang
|
7f323a589f
Add `--no-op-offload` to improve `-ot` pp perf in MoE models like llama4 400B (#13386)
|
преди 8 месеца |
|
Sigbjørn Skjæret
|
d2a4ef05c6
vocab : add ByteDance-Seed/Seed-Coder (#13423)
|
преди 8 месеца |
|
Georgi Gerganov
|
6562e5a4d6
context : allow cache-less context for embeddings (#13108)
|
преди 8 месеца |
|
Georgi Gerganov
|
51fb96b1ff
context : remove logits_all flag (#13284)
|
преди 8 месеца |
|
Georgi Gerganov
|
d9d398f84f
sampling : when top-k <= 0 -> noop (#13173)
|
преди 8 месеца |
 Xuan-Son Nguyen
|
ecda2ec4b3
mtmd : Support Pixtral 12B (#13065)
|
преди 9 месеца |
 Ed Addario
|
71e90e8813
quantize: Handle user-defined quantization levels for additional tensors (#12511)
|
преди 9 месеца |
|
Xuan-Son Nguyen
|
1466621e73
llama : Support llama 4 text-only (#12791)
|
преди 9 месеца |
|
Diego Devesa
|
e0e912f49b
llama : add option to override model tensor buffers (#11397)
|
преди 9 месеца |
|
Sigbjørn Skjæret
|
2c3f8b850a
llama : support BailingMoE (Ling) (#12634)
|
преди 9 месеца |
 Juyoung Suk
|
b3de7cac73
llama : add Trillion 7B model support (#12556)
|
преди 9 месеца |
|
Johannes Gäßler
|
dd373dd3bf
llama: fix error on bad grammar (#12628)
|
преди 9 месеца |
 compilade
|
00d53800e0
llama-vocab : add SuperBPE pre-tokenizer (#12532)
|
преди 10 месеца |
 fairydreaming
|
8fcb563613
Load all MoE experts during warmup (#11571)
|
преди 10 месеца |
|
Georgi Gerganov
|
e0dbec0bc6
llama : refactor llama_context, llama_kv_cache, llm_build_context (#12181)
|
преди 10 месеца |
 Olivier Chafik
|
669912d9a5
`tool-call`: fix Qwen 2.5 Coder support, add micro benchmarks, support trigger patterns for lazy grammars (#12034)
|
преди 10 месеца |
 Xuan-Son Nguyen
|
c43a3e7996
llama : add Phi-4-mini support (supersede #12099) (#12108)
|
преди 10 месеца |
