Skip to content

Releases: ggml-org/llama.cpp

b7388

13 Dec 23:46
@github-actions github-actions
b7388
4ed2bae
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Choose a tag to compare

View all tags
b7388 Latest
Latest

Warning

Release Format Update: Linux releases will soon use .tar.gz archives instead of .zip. Please make the necessary changes to your deployment scripts.

server-models.cpp: add missing (#18000)

Fixes: #17999

macOS/iOS:

Linux:

Windows:

openEuler:

Assets 28

b7387

13 Dec 21:33
@github-actions github-actions
b7387
5266379
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Choose a tag to compare

View all tags
b7387

Warning

Release Format Update: Linux releases will soon use .tar.gz archives instead of .zip. Please make the necessary changes to your deployment scripts.

llama_context: synchronize before reallocating output buffer (#17974)

macOS/iOS:

Linux:

Windows:

openEuler:

Loading

b7386

13 Dec 21:25
@github-actions github-actions
b7386
4d5ae24
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Choose a tag to compare

View all tags
b7386

Warning

Release Format Update: Linux releases will soon use .tar.gz archives instead of .zip. Please make the necessary changes to your deployment scripts.

arg: fix common_params_parse not accepting negated arg (#17991)

macOS/iOS:

Linux:

Windows:

openEuler:

Loading

b7385

13 Dec 20:54
@github-actions github-actions
b7385
66ba512
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Choose a tag to compare

View all tags
b7385

Warning

Release Format Update: Linux releases will soon use .tar.gz archives instead of .zip. Please make the necessary changes to your deployment scripts.

cmake: correct scope - link ws2_32 for MinGW/w64devkit builds in cpp-httplib (#17972)

  • fix - w64devkit build

  • fix - w64devkit build private scope

macOS/iOS:

Linux:

Windows:

openEuler:

Loading

b7384

13 Dec 19:59
@github-actions github-actions
b7384
36255a2
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Choose a tag to compare

View all tags
b7384

Warning

Release Format Update: Linux releases will soon use .tar.gz archives instead of .zip. Please make the necessary changes to your deployment scripts.

vulkan: support get_rows for i32 (#17941)

macOS/iOS:

Linux:

Windows:

openEuler:

Loading

b7383

13 Dec 18:56
@github-actions github-actions
b7383
3229a23
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Choose a tag to compare

View all tags
b7383

Warning

Release Format Update: Linux releases will soon use .tar.gz archives instead of .zip. Please make the necessary changes to your deployment scripts.

vulkan: support GGML_OP_DIAG (#17893)

macOS/iOS:

Linux:

Windows:

openEuler:

Loading

b7382

13 Dec 18:37
@github-actions github-actions
b7382
303f861
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Choose a tag to compare

View all tags
b7382

Warning

Release Format Update: Linux releases will soon use .tar.gz archives instead of .zip. Please make the necessary changes to your deployment scripts.

vulkan: Multi-pass softmax for large number of cols (#17892)

When the number of cols is large, split each row across multiple workgroups.
There are three phases that communicate partial results through temp buffers:
(1) compute max partials
(2) take max of partials, compute sum(exp(x-max)) partials
(3) sum partials, compute scaled result

macOS/iOS:

Linux:

Windows:

openEuler:

Loading

b7381

13 Dec 17:15
@github-actions github-actions
b7381
3c6391e
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Choose a tag to compare

View all tags
b7381

Warning

Release Format Update: Linux releases will soon use .tar.gz archives instead of .zip. Please make the necessary changes to your deployment scripts.

speculative-simple : free batch on exit (#17985)

macOS/iOS:

Linux:

Windows:

openEuler:

Loading

b7380

13 Dec 16:40
@github-actions github-actions
b7380
8e4d678
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Choose a tag to compare

View all tags
b7380

Warning

Release Format Update: Linux releases will soon use .tar.gz archives instead of .zip. Please make the necessary changes to your deployment scripts.

common : skip model validation when --completion-bash is requested (#17975)

macOS/iOS:

Linux:

Windows:

openEuler:

Loading

b7379

13 Dec 15:38
@github-actions github-actions
b7379
07a10c1
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Choose a tag to compare

View all tags
b7379

Warning

Release Format Update: Linux releases will soon use .tar.gz archives instead of .zip. Please make the necessary changes to your deployment scripts.

vulkan: Allow non-pow2 n_experts in topk_moe (#17872)

macOS/iOS:

Linux:

Windows:

openEuler:

Loading