为了安全，强烈建议开启2FA双因子认证：User Settings -> Account -> Enable two-factor authentication！！！

Tags

Tags give the ability to mark specific points in history as being important

b4785

581650b7 · vulkan: improve im2col (#11826) · Feb 28, 2025
b4784

b95c8af3 · cmake: Fix ggml backend dependencies and installation (#11818) · Feb 27, 2025
b4783

a800ae46 · llava : add struct for FFI bindgen (#12079) · Feb 26, 2025
b4778

a82c9e7c · vulkan: fix assertion when qy_needs_dequant (#12068) · Feb 25, 2025
b4777

401af80b · server: handle echo=false on /v1/completions (#12060) · Feb 25, 2025
b4776

c132239b · add OP sigmoid (#12056) · Feb 25, 2025
b4775

393fca62 · ggml-cpu: Fix build with sve (#12059) · Feb 25, 2025
b4774

61d4f39d · vulkan: implement more backpropagation operators (#11914) · Feb 25, 2025
b4773

0b527456 · server: support add_generation_prompt query param (#12062) · Feb 25, 2025
b4771

3e9a2860 · llama : expose llama_model_n_head_kv in the API (#11997) · Feb 25, 2025
b4770

58d07a80 · metal : copy kernels for quant to F32/F16 conversions (#12017) · Feb 25, 2025
b4769

34a846b5 · opencl: fix for small models (#11950) · Feb 24, 2025
b4768

7a2c913e · llava : Add Granite Vision Support (#11794) · Feb 24, 2025
b4767

08d59862 · [SYCL] Optimize mul_mat for Q4_0 on Intel GPU (#12035) · Feb 24, 2025
b4765

8303e8b0 · SYCL: Fix GGML_SYCL_DEBUG macro (#11995) · Feb 24, 2025
b4764

7ad0779f · run: allow to customize prompt by env var LLAMA_PROMPT_PREFIX (#12041) · Feb 23, 2025
b4763

f777a73e · Some llama-run cleanups (#11973) · Feb 23, 2025
b4762

af7747c9 · ggml-cpu: Support s390x SIMD Instruction Set (#12019) · Feb 22, 2025
b4761

a28e0d5e · CUDA: app option to compile without FlashAttention (#12025) · Feb 22, 2025
b4760

36c258ee · llava: build clip image from pixels (#11999) · Feb 22, 2025

1
…
18
19
20
21
22
23
24
25
26
…
178