Tags

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

b3072

549279d8 · llama : avoid double token-to-piece cache (#7654) · Jun 03, 2024
b3071

9e405b6e · kompute : implement op_getrows_f32 (#6403) · Jun 03, 2024
b3070

3413ae21 · fix bug introduced in using calloc (#7701) · Jun 02, 2024
b3067

9422c5e3 · [SYCL] Update rpc-server.cpp to include SYCL backend (#7682) · Jun 02, 2024
b3066

e141ce62 · Fix FlashAttention debug test, FP32 assert (#7684) · Jun 01, 2024
b3065

2e666832 · server : new UI (#7633) · Jun 01, 2024
b3063

750f60c0 · CUDA: fix Pascal FA, deq. KV to FP16 for batch > 8 (#7681) · Jun 01, 2024
b3058

30e238b2 · Improve HIP compatibility (#7672) · May 31, 2024
b3056

0c27e6f6 · ggml : fix loongson compile warnings (#7537) · May 31, 2024
b3051

5921b8f0 · llama : cache llama_token_to_piece (#7587) · May 31, 2024
b3046

9c4c9cc8 · Move convert.py to examples/convert-legacy-llama.py (#7430) · May 30, 2024
b3045

59b0d077 · faster avx512 exp implementation (#7551) · May 30, 2024
b3044

d5c05821 · ggml : fix loongarch build (O2 issue) (#7636) · May 30, 2024
b3042

3854c9d0 · [SYCL] fix intel docker (#7630) · May 30, 2024
b3040

55d62262 · metal : remove invalid asserts (#7617) · May 29, 2024
b3039

975ec63f · metal : add missing asserts (#7617) · May 29, 2024
b3038

fb76ec31 · ggml : fix YARN + add tests + add asserts (#7617) · May 29, 2024
b3037

cce3dcff · cuda : non-cont concat support (#7610) · May 29, 2024
b3036

210d9917 · llama-bench : add support for the RPC backend (#7435) · May 29, 2024
b3035

87bdf2a1 · ggml : use atomic_flag for critical section (#7598) · May 29, 2024

Prev
1
2
3
4
5
…
98
Next