Skip to content

Releases: ggml-org/llama.cpp

b7642

06 Jan 08:43
e75ee11

Choose a tag to compare

b7640

06 Jan 00:16
e443fbc

Choose a tag to compare

ggml webgpu: add CEIL operation support (#18605)

  • ggml-webgpu: add CEIL operation support

    Add support for the CEIL unary operation in the WebGPU backend:
    - Add CEIL_FUNC shader template in unary_op.wgsl
    - Add 4 shader variants (f32, f16, inplace versions)
    - Initialize CEIL pipelines in ggml-webgpu.cpp
    - Register CEIL in supports_op function
    
  • docs: update WebGPU ops support for CEIL

macOS/iOS:

Linux:

Windows:

openEuler:

b7639

05 Jan 23:26
73d284a

Choose a tag to compare

b7638

05 Jan 23:00
df17a4c

Choose a tag to compare

b7636

05 Jan 21:16
f47edb8

Choose a tag to compare

b7635

05 Jan 19:38
da143b9

Choose a tag to compare

b7634

05 Jan 16:33
f1768d8

Choose a tag to compare

b7633

05 Jan 15:56
2da64a2

Choose a tag to compare

b7632

05 Jan 13:55
b37124d

Choose a tag to compare

vulkan: handle quantize_q8_1 overflowing the max workgroup count (#18515)

  • vulkan: handle quantize_q8_1 overflowing the max workgroup count

  • vulkan: Fix small tile size matmul on lavapipe

  • fix mul_mat_id failures

macOS/iOS:

Linux:

Windows:

openEuler:

b7631

05 Jan 09:47
eadc418

Choose a tag to compare

llama : refactor rope_freq_base/scale_swa conversion and init (#18553)

  • refactor rope_freq_base/scale_swa conversion and init

  • safe defaults for unknowns

  • update relevant models

  • grammar

  • add get_rope_freq_scale to modern-bert

  • const

  • const

  • log swa info

macOS/iOS:

Linux:

Windows:

openEuler: