Fetching metadata from the HF Docker repository...
release : v1.5.3
1f8a047
unverified
-
.devops
docker : fix the publishing of the CUDA Docker image (#1704)
-
.github
ci : build with CLBlast + ggml-opencl use GGML_API (#1576)
-
bindings
release : v1.5.3
-
cmake
cmake : update to 3.19 (#351)
-
coreml
coreml : use the correct `n_mel` value (#1458)
-
examples
ci : build with CLBlast + ggml-opencl use GGML_API (#1576)
-
extra
cuda : simplify expression
-
grammars
whisper : add grammar-based sampling (#1229)
-
models
download : fix large q5 model name (#1695)
-
openvino
whisper : replace `tensor->n_dims` with `ggml_n_dims(tensor)` (#1694)
-
samples
Create README.md
-
spm-headers
ios : add support for Swift Package Manager (#1370)
-
tests
whisper : make large version explicit + fix data size units (#1493)
-
804 Bytes
Initial release
-
803 Bytes
server : add a REST Whisper server example with OAI-like API (#1380)
-
96 Bytes
cmake : add submodule whisper.spm
-
19 kB
release : v1.5.3
-
1.07 kB
license : update year (#739)
-
14.7 kB
sync : ggml (VMM, sync-ggml-am, dotprod ARM fixes, CUDA fixes) (#1691)
-
1.78 kB
swift : update Package.swift to use ggml as package dependency (#1701)
-
37 kB
release : v1.5.3
-
28.6 kB
sync : ggml (ggml_scale, ggml_row_size, etc.) (#1677)
-
3.85 kB
sync : ggml (Metal fixes, new ops, tests) (#1633)
-
4.9 kB
ggml : add error handling to graph_compute (#1714)
-
52.5 kB
ggml : add error handling to graph_compute (#1714)
-
8.99 kB
ggml : add error handling to graph_compute (#1714)
-
379 kB
ggml : add error handling to graph_compute (#1714)
-
2.51 kB
sync : ggml (new ops, new backend, etc) (#1602)
-
7.38 kB
sync : ggml (new ops, new backend, etc) (#1602)
-
4.33 kB
ggml : add error handling to graph_compute (#1714)
-
146 kB
ggml : add error handling to graph_compute (#1714)
-
180 kB
metal : add kernel_get_rows_i32
-
71 kB
sync : ggml (new ops, new backend, etc) (#1602)
-
926 Bytes
ci : build with CLBlast + ggml-opencl use GGML_API (#1576)
-
272 kB
ggml : add ggml_vdotq_s32 alias (llama/4715)
-
10.2 kB
sync : ggml (backend v2, k-quants, CUDA opts, Metal opts, etc.) (#1422)
-
655 kB
ggml : add ggml_cpu_has_avx_vnni() (llama/4589)
-
82.8 kB
ggml : add ggml_cpu_has_avx_vnni() (llama/4589)
-
231 kB
ggml : add error handling to graph_compute (#1714)
-
30.2 kB
whisper : remove trailing whitespaces
