Skip to content

Releases: ggerganov/llama.cpp

b3039

29 May 19:51
975ec63
Compare
Choose a tag to compare
metal : add missing asserts (#7617)

b3038

29 May 18:50
fb76ec3
Compare
Choose a tag to compare
ggml : fix YARN + add tests + add asserts (#7617)

* tests : add rope tests

ggml-ci

* ggml : fixes (hopefully)

ggml-ci

* tests : add non-cont tests

ggml-ci

* cuda : add asserts for rope/norm + fix DS2

ggml-ci

* ggml : assert contiguousness

* tests : reduce RoPE tests

ggml-ci

b3037

29 May 15:47
cce3dcf
Compare
Choose a tag to compare
cuda : non-cont concat support (#7610)

* tests : add non-cont concat tests

* cuda : non-cont concat support

ggml-ci

b3036

29 May 15:24
210d991
Compare
Choose a tag to compare
llama-bench : add support for the RPC backend (#7435)

b3035

29 May 15:23
87bdf2a
Compare
Choose a tag to compare
ggml : use atomic_flag for critical section (#7598)

* ggml : use atomic_flag for critical section

* add windows shims

b3033

29 May 14:18
2ab9772
Compare
Choose a tag to compare
sync : ggml

b3030

29 May 03:17
504f0c3
Compare
Choose a tag to compare
ggml : fix typo in ggml.c (#7603)

b3029

28 May 23:45
b864b50
Compare
Choose a tag to compare
[SYCL] Align GEMM dispatch (#7566)

* align GEMM dispatch

b3028

28 May 21:58
02c1eca
Compare
Choose a tag to compare
Tokenizer WPM fixes (#7500)

* Update random test: add_bos_token.
* Update random test: add WPM models for testing.
* Build vocab.special_tokens_cache using vocab token types.
* Fix and improve WPM preprocessing.
  - Fix unicode edge case combinations.
  - Split by whitspace in the same pass.
* Discard all tokens when no matching found.

b3027

28 May 21:47
6bd12ce
Compare
Choose a tag to compare
sycl : fix assert (#7563)